Merge pull request #5744 from open-webui/dev

0.3.31

.github/ISSUE_TEMPLATE/bug_report.md

@@ -8,6 +8,10 @@ assignees: ''
 
 # Bug Report
 
+**Important: Before submitting a bug report, please check whether a similar issue or feature request has already been posted in the Issues or Discussions section. It's likely we're already tracking it. In case of uncertainty, initiate a discussion post first. This helps us all to efficiently focus on improving the project.**
+
+**Let's collaborate respectfully. If you bring negativity, please understand our capacity to engage may be limited. If you're open to learning and communicating constructively, we're more than happy to assist you. Remember, Open WebUI is a volunteer-driven project maintained by a single maintainer, supported by our amazing contributors who also manage full-time jobs. We respect your time; please respect ours. If you have an issue, We highly encourage you to submit a pull request or to fork the project. We actively work to prevent contributor burnout to preserve the quality and continuity of Open WebUI.**
+
 ## Installation Method
 
 [Describe the method you used to install the project, e.g., git clone, Docker, pip, etc.]

.github/ISSUE_TEMPLATE/feature_request.md

@@ -6,6 +6,12 @@ labels: ''
 assignees: ''
 ---
 
+# Feature Request
+
+**Important: Before submitting a feature request, please check whether a similar issue or feature request has already been posted in the Issues or Discussions section. It's likely we're already tracking it. In case of uncertainty, initiate a discussion post first. This helps us all to efficiently focus on improving the project.**
+
+**Let's collaborate respectfully. If you bring negativity, please understand our capacity to engage may be limited. If you're open to learning and communicating constructively, we're more than happy to assist you. Remember, Open WebUI is a volunteer-driven project maintained by a single maintainer, supported by our amazing contributors who also manage full-time jobs. We respect your time; please respect ours. If you have an issue, We highly encourage you to submit a pull request or to fork the project. We actively work to prevent contributor burnout to preserve the quality and continuity of Open WebUI.**
+
 **Is your feature request related to a problem? Please describe.**
 A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
 

.github/workflows/docker-build.yaml

@@ -56,25 +56,19 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      - name: Get version number from package.json
-        id: get_version
-        run: |
-          VERSION=$(jq -r '.version' package.json)
-          echo "version=$VERSION" >> $GITHUB_OUTPUT
-
-      - name: Extract metadata for Docker images
+      - name: Extract metadata for Docker images (default latest tag)
         id: meta
         uses: docker/metadata-action@v5
         with:
           images: ${{ env.FULL_IMAGE_NAME }}
           tags: |
-            type=raw,value=latest,enable=${{ github.ref == 'refs/heads/main' }}
-            type=raw,value=${{ steps.get_version.outputs.version }},enable=${{ github.ref == 'refs/heads/main' }}
             type=ref,event=branch
             type=ref,event=tag
             type=sha,prefix=git-
             type=semver,pattern={{version}}
             type=semver,pattern={{major}}.{{minor}}
+          flavor: |
+            latest=${{ github.ref == 'refs/heads/main' }}
 
       - name: Extract metadata for Docker cache
         id: cache-meta
@@ -88,7 +82,7 @@ jobs:
             prefix=cache-${{ matrix.platform }}-
             latest=false
 
-      - name: Build Docker image
+      - name: Build Docker image (latest)
         uses: docker/build-push-action@v5
         id: build
         with:
@@ -96,8 +90,7 @@ jobs:
           push: true
           platforms: ${{ matrix.platform }}
           labels: ${{ steps.meta.outputs.labels }}
-          tags: ${{ steps.meta.outputs.tags }}
-          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push=true
+          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
           cache-from: type=registry,ref=${{ steps.cache-meta.outputs.tags }}
           cache-to: type=registry,ref=${{ steps.cache-meta.outputs.tags }},mode=max
           build-args: |
@@ -160,25 +153,21 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      - name: Get version number from package.json
-        id: get_version
-        run: |
-          VERSION=$(jq -r '.version' package.json)
-          echo "version=$VERSION" >> $GITHUB_OUTPUT
-
-      - name: Extract metadata for Docker images
+      - name: Extract metadata for Docker images (cuda tag)
         id: meta
         uses: docker/metadata-action@v5
         with:
           images: ${{ env.FULL_IMAGE_NAME }}
           tags: |
-            type=raw,value=latest-cuda,enable=${{ github.ref == 'refs/heads/main' }}
-            type=raw,value=${{ steps.get_version.outputs.version }}-cuda,enable=${{ github.ref == 'refs/heads/main' }}
-            type=ref,event=branch,suffix=-cuda
-            type=ref,event=tag,suffix=-cuda
-            type=sha,prefix=git-,suffix=-cuda
-            type=semver,pattern={{version}},suffix=-cuda
-            type=semver,pattern={{major}}.{{minor}},suffix=-cuda
+            type=ref,event=branch
+            type=ref,event=tag
+            type=sha,prefix=git-
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=raw,enable=${{ github.ref == 'refs/heads/main' }},prefix=,suffix=,value=cuda
+          flavor: |
+            latest=${{ github.ref == 'refs/heads/main' }}
+            suffix=-cuda,onlatest=true
 
       - name: Extract metadata for Docker cache
         id: cache-meta
@@ -200,8 +189,7 @@ jobs:
           push: true
           platforms: ${{ matrix.platform }}
           labels: ${{ steps.meta.outputs.labels }}
-          tags: ${{ steps.meta.outputs.tags }}
-          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push=true
+          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
           cache-from: type=registry,ref=${{ steps.cache-meta.outputs.tags }}
           cache-to: type=registry,ref=${{ steps.cache-meta.outputs.tags }},mode=max
           build-args: |
@@ -265,25 +253,21 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      - name: Get version number from package.json
-        id: get_version
-        run: |
-          VERSION=$(jq -r '.version' package.json)
-          echo "version=$VERSION" >> $GITHUB_OUTPUT
-
-      - name: Extract metadata for Docker images
+      - name: Extract metadata for Docker images (ollama tag)
         id: meta
         uses: docker/metadata-action@v5
         with:
           images: ${{ env.FULL_IMAGE_NAME }}
           tags: |
-            type=raw,value=latest-ollama,enable=${{ github.ref == 'refs/heads/main' }}
-            type=raw,value=${{ steps.get_version.outputs.version }}-ollama,enable=${{ github.ref == 'refs/heads/main' }}
-            type=ref,event=branch,suffix=-ollama
-            type=ref,event=tag,suffix=-ollama
-            type=sha,prefix=git-,suffix=-ollama
-            type=semver,pattern={{version}},suffix=-ollama
-            type=semver,pattern={{major}}.{{minor}},suffix=-ollama
+            type=ref,event=branch
+            type=ref,event=tag
+            type=sha,prefix=git-
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=raw,enable=${{ github.ref == 'refs/heads/main' }},prefix=,suffix=,value=ollama
+          flavor: |
+            latest=${{ github.ref == 'refs/heads/main' }}
+            suffix=-ollama,onlatest=true
 
       - name: Extract metadata for Docker cache
         id: cache-meta
@@ -305,8 +289,7 @@ jobs:
           push: true
           platforms: ${{ matrix.platform }}
           labels: ${{ steps.meta.outputs.labels }}
-          tags: ${{ steps.meta.outputs.tags }}
-          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push=true
+          outputs: type=image,name=${{ env.FULL_IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
           cache-from: type=registry,ref=${{ steps.cache-meta.outputs.tags }}
           cache-to: type=registry,ref=${{ steps.cache-meta.outputs.tags }},mode=max
           build-args: |
@@ -326,9 +309,10 @@ jobs:
           path: /tmp/digests/*
           if-no-files-found: error
           retention-days: 1
+
   merge-main-images:
     runs-on: ubuntu-latest
-    needs: [ build-main-image ]
+    needs: [build-main-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking
@@ -380,10 +364,9 @@ jobs:
         run: |
           docker buildx imagetools inspect ${{ env.FULL_IMAGE_NAME }}:${{ steps.meta.outputs.version }}
 
-
   merge-cuda-images:
     runs-on: ubuntu-latest
-    needs: [ build-cuda-image ]
+    needs: [build-cuda-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking
@@ -439,7 +422,7 @@ jobs:
 
   merge-ollama-images:
     runs-on: ubuntu-latest
-    needs: [ build-ollama-image ]
+    needs: [build-ollama-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking

.github/workflows/format-backend.yaml

@@ -23,7 +23,7 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Set up Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
 

.github/workflows/format-build-frontend.yaml

@@ -21,7 +21,7 @@ jobs:
       - name: Setup Node.js
         uses: actions/setup-node@v4
         with:
-          node-version: '20' # Or specify any other version you want to use
+          node-version: '22' # Or specify any other version you want to use
 
       - name: Install Dependencies
         run: npm install
@@ -48,7 +48,7 @@ jobs:
       - name: Setup Node.js
         uses: actions/setup-node@v4
         with:
-          node-version: '20'
+          node-version: '22'
 
       - name: Install Dependencies
         run: npm ci

.github/workflows/integration-test.yml

@@ -85,7 +85,7 @@ jobs:
   #     - uses: actions/checkout@v4
 
   #     - name: Set up Python
-  #       uses: actions/setup-python@v4
+  #       uses: actions/setup-python@v5
   #       with:
   #         python-version: ${{ matrix.python-version }}
 
@@ -157,7 +157,7 @@ jobs:
           GLOBAL_LOG_LEVEL: debug
         run: |
           cd backend
-          uvicorn main:app --port "8080" --forwarded-allow-ips '*' &
+          uvicorn open_webui.main:app --port "8080" --forwarded-allow-ips '*' &
           UVICORN_PID=$!
           # Wait up to 40 seconds for the server to start
           for i in {1..40}; do
@@ -182,9 +182,12 @@ jobs:
           WEBUI_SECRET_KEY: secret-key
           GLOBAL_LOG_LEVEL: debug
           DATABASE_URL: postgresql://postgres:postgres@localhost:5432/postgres
+          DATABASE_POOL_SIZE: 10
+          DATABASE_POOL_MAX_OVERFLOW: 10
+          DATABASE_POOL_TIMEOUT: 30
         run: |
           cd backend
-          uvicorn main:app --port "8081" --forwarded-allow-ips '*' &
+          uvicorn open_webui.main:app --port "8081" --forwarded-allow-ips '*' &
           UVICORN_PID=$!
           # Wait up to 20 seconds for the server to start
           for i in {1..20}; do
@@ -230,7 +233,7 @@ jobs:
 #          DATABASE_URL: mysql://root:mysql@localhost:3306/mysql
 #        run: |
 #          cd backend
-#          uvicorn main:app --port "8083" --forwarded-allow-ips '*' &
+#          uvicorn open_webui.main:app --port "8083" --forwarded-allow-ips '*' &
 #          UVICORN_PID=$!
 #          # Wait up to 20 seconds for the server to start
 #          for i in {1..20}; do

.github/workflows/lint-backend.disabled

@@ -16,7 +16,7 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - name: Use Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
       - name: Use Bun
         uses: oven-sh/setup-bun@v1
       - name: Install dependencies

.github/workflows/release-pypi.yml

@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main # or whatever branch you want to use
+      - pypi-release
 
 jobs:
   release:

CHANGELOG.md

@@ -5,6 +5,252 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.3.31] - 2024-10-06
+
+### Added
+
+- **📚 Knowledge Feature**: Reimagined documents feature, now more performant with a better UI for enhanced organization; includes streamlined API integration for Retrieval-Augmented Generation (RAG). Detailed documentation forthcoming: https://docs.openwebui.com/
+- **🌐 New Landing Page**: Freshly designed landing page; toggle between the new UI and the classic chat UI from Settings > Interface for a personalized experience.
+- **📁 Full Document Retrieval Mode**: Toggle between full document retrieval or traditional snippets by clicking on the file item. This mode enhances document capabilities and supports comprehensive tasks like summarization by utilizing the entire content instead of RAG.
+- **📄 Extracted File Content Display**: View extracted content directly by clicking on the file item, simplifying file analysis.
+- **🎨 Artifacts Feature**: Render web content and SVGs directly in the interface, supporting quick iterations and live changes.
+- **🖊️ Editable Code Blocks**: Supercharged code blocks now allow live editing directly in the LLM response, with live reloads supported by artifacts.
+- **🔧 Code Block Enhancements**: Introduced a floating copy button in code blocks to facilitate easier code copying without scrolling.
+- **🔍 SVG Pan/Zoom**: Enhanced interaction with SVG images, including Mermaid diagrams, via new pan and zoom capabilities.
+- **🔍 Text Select Quick Actions**: New floating buttons appear when text is highlighted in LLM responses, offering deeper interactions like "Ask a Question" or "Explain".
+- **🗃️ Database Pool Configuration**: Enhanced database handling to support scalable user growth.
+- **🔊 Experimental Audio Compression**: Compress audio files to navigate around the 25MB limit for OpenAI's speech-to-text processing.
+- **🔍 Query Embedding**: Adjusted embedding behavior to enhance system performance by not repeating query embedding.
+- **💾 Lazy Load Optimizations**: Implemented lazy loading of large dependencies to minimize initial memory usage, boosting performance.
+- **🍏 Apple Touch Icon Support**: Optimizes the display of icons for web bookmarks on Apple mobile devices.
+- **🔽 Expandable Content Markdown Support**: Introducing 'details', 'summary' tag support for creating expandable content sections in markdown, facilitating cleaner, organized documentation and interactive content display.
+
+### Fixed
+
+- **🔘 Action Button Issue**: Resolved a bug where action buttons were not functioning, enhancing UI reliability.
+- **🔄 Multi-Model Chat Loop**: Fixed an infinite loop issue in multi-model chat environments, ensuring smoother chat operations.
+- **📄 Chat PDF/TXT Export Issue**: Resolved problems with exporting chat logs to PDF and TXT formats.
+- **🔊 Call to Text-to-Speech Issues**: Rectified problems with text-to-speech functions to improve audio interactions.
+
+### Changed
+
+- **⚙️ Endpoint Renaming**: Renamed 'rag' endpoints to 'retrieval' for clearer function description.
+- **🎨 Styling and Interface Updates**: Multiple refinements across the platform to enhance visual appeal and user interaction.
+
+### Removed
+
+- **🗑️ Deprecated 'DOCS_DIR'**: Removed the outdated 'docs_dir' variable in favor of more direct file management solutions, with direct file directory syncing and API uploads for a more integrated experience.
+
+## [0.3.30] - 2024-09-26
+
+### Fixed
+
+- **🍞 Update Available Toast Dismissal**: Enhanced user experience by ensuring that once the update available notification is dismissed, it won't reappear for 24 hours.
+- **📋 Ollama /embed Form Data**: Adjusted the integration inaccuracies in the /embed form data to ensure it perfectly matches with Ollama's specifications.
+- **🔧 O1 Max Completion Tokens Issue**: Resolved compatibility issues with OpenAI's o1 models max_completion_tokens param to ensure smooth operation.
+- **🔄 Pip Install Database Issue**: Fixed a critical issue where database changes during pip installations were reverting and not saving chat logs, now ensuring data persistence and reliability in chat operations.
+- **🏷️ Chat Rename Tab Update**: Fixed the functionality to change the web browser's tab title simultaneously when a chat is renamed, keeping tab titles consistent.
+
+## [0.3.29] - 2023-09-25
+
+### Fixed
+
+- **🔧 KaTeX Rendering Improvement**: Resolved specific corner cases in KaTeX rendering to enhance the display of complex mathematical notation.
+- **📞 'Call' URL Parameter Fix**: Corrected functionality for 'call' URL search parameter ensuring reliable activation of voice calls through URL triggers.
+- **🔄 Configuration Reset Fix**: Fixed the RESET_CONFIG_ON_START to ensure settings revert to default correctly upon each startup, improving reliability in configuration management.
+- **🌍 Filter Outlet Hook Fix**: Addressed issues in the filter outlet hook, ensuring all filter functions operate as intended.
+
+## [0.3.28] - 2024-09-24
+
+### Fixed
+
+- **🔍 Web Search Functionality**: Corrected an issue where the web search option was not functioning properly.
+
+## [0.3.27] - 2024-09-24
+
+### Fixed
+
+- **🔄 Periodic Cleanup Error Resolved**: Fixed a critical RuntimeError related to the 'periodic_usage_pool_cleanup' coroutine, ensuring smooth and efficient performance post-pip install, correcting a persisting issue from version 0.3.26.
+- **📊 Enhanced LaTeX Rendering**: Improved rendering for LaTeX content, enhancing clarity and visual presentation in documents and mathematical models.
+
+## [0.3.26] - 2024-09-24
+
+### Fixed
+
+- **🔄 Event Loop Error Resolution**: Addressed a critical error where a missing running event loop caused 'periodic_usage_pool_cleanup' to fail with pip installs. This fix ensures smoother and more reliable updates and installations, enhancing overall system stability.
+
+## [0.3.25] - 2024-09-24
+
+### Fixed
+
+- **🖼️ Image Generation Functionality**: Resolved an issue where image generation was not functioning, restoring full capability for visual content creation.
+- **⚖️ Rate Response Corrections**: Addressed a problem where rate responses were not working, ensuring reliable feedback mechanisms are operational.
+
+## [0.3.24] - 2024-09-24
+
+### Added
+
+- **🚀 Rendering Optimization**: Significantly improved message rendering performance, enhancing user experience and webui responsiveness.
+- **💖 Favorite Response Feature in Chat Overview**: Users can now mark responses as favorite directly from the chat overview, enhancing ease of retrieval and organization of preferred responses.
+- **💬 Create Message Pairs with Shortcut**: Implemented creation of new message pairs using Cmd/Ctrl+Shift+Enter, making conversation editing faster and more intuitive.
+- **🌍 Expanded User Prompt Variables**: Added weekday, timezone, and language information variables to user prompts to match system prompt variables.
+- **🎵 Enhanced Audio Support**: Now includes support for 'audio/x-m4a' files, broadening compatibility with audio content within the platform.
+- **🔏 Model URL Search Parameter**: Added an ability to select a model directly via URL parameters, streamlining navigation and model access.
+- **📄 Enhanced PDF Citations**: PDF citations now open at the associated page, streamlining reference checks and document handling.
+- **🔧Use of Redis in Sockets**: Enhanced socket implementation to fully support Redis, enabling effective stateless instances suitable for scalable load balancing.
+- **🌍 Stream Individual Model Responses**: Allows specific models to have individualized streaming settings, enhancing performance and customization.
+- **🕒 Display Model Hash and Last Modified Timestamp for Ollama Models**: Provides critical model details directly in the Models workspace for enhanced tracking.
+- **❗ Update Info Notification for Admins**: Ensures administrators receive immediate updates upon login, keeping them informed of the latest changes and system statuses.
+
+### Fixed
+
+- **🗑️ Temporary File Handling On Windows**: Fixed an issue causing errors when accessing a temporary file being used by another process, Tools & Functions should now work as intended.
+- **🔓 Authentication Toggle Issue**: Resolved the malfunction where setting 'WEBUI_AUTH=False' did not appropriately disable authentication, ensuring that user experience and system security settings function as configured.
+- **🔧 Save As Copy Issue for Many Model Chats**: Resolved an error preventing users from save messages as copies in many model chats.
+- **🔒 Sidebar Closure on Mobile**: Resolved an issue where the mobile sidebar remained open after menu engagement, improving user interface responsivity and comfort.
+- **🛡️ Tooltip XSS Vulnerability**: Resolved a cross-site scripting (XSS) issue within tooltips, ensuring enhanced security and data integrity during user interactions.
+
+### Changed
+
+- **↩️ Deprecated Interface Stream Response Settings**: Moved to advanced parameters to streamline interface settings and enhance user clarity.
+- **⚙️ Renamed 'speedRate' to 'playbackRate'**: Standardizes terminology, improving usability and understanding in media settings.
+
+## [0.3.23] - 2024-09-21
+
+### Added
+
+- **🚀 WebSocket Redis Support**: Enhanced load balancing capabilities for multiple instance setups, promoting better performance and reliability in WebUI.
+- **🔧 Adjustable Chat Controls**: Introduced width-adjustable chat controls, enabling a personalized and more comfortable user interface.
+- **🌎 i18n Updates**: Improved and updated the Chinese translations.
+
+### Fixed
+
+- **🌐 Task Model Unloading Issue**: Modified task handling to use the Ollama /api/chat endpoint instead of OpenAI compatible endpoint, ensuring models stay loaded and ready with custom parameters, thus minimizing delays in task execution.
+- **📝 Title Generation Fix for OpenAI Compatible APIs**: Resolved an issue preventing the generation of titles, enhancing consistency and reliability when using multiple API providers.
+- **🗃️ RAG Duplicate Collection Issue**: Fixed a bug causing repeated processing of the same uploaded file. Now utilizes indexed files to prevent unnecessary duplications, optimizing resource usage.
+- **🖼️ Image Generation Enhancement**: Refactored OpenAI image generation endpoint to be asynchronous, preventing the WebUI from becoming unresponsive during processing, thus enhancing user experience.
+- **🔓 Downgrade Authlib**: Reverted Authlib to version 1.3.1 to address and resolve issues concerning OAuth functionality.
+
+### Changed
+
+- **🔍 Improved Message Interaction**: Enhanced the message node interface to allow for easier focus redirection with a simple click, streamlining user interaction.
+- **✨ Styling Refactor**: Updated WebUI styling for a cleaner, more modern look, enhancing user experience across the platform.
+
+## [0.3.22] - 2024-09-19
+
+### Added
+
+- **⭐ Chat Overview**: Introducing a node-based interactive messages diagram for improved visualization of conversation flows.
+- **🔗 Multiple Vector DB Support**: Now supports multiple vector databases, including the newly added Milvus support. Community contributions for additional database support are highly encouraged!
+- **📡 Experimental Non-Stream Chat Completion**: Experimental feature allowing the use of OpenAI o1 models, which do not support streaming, ensuring more versatile model deployment.
+- **🔍 Experimental Colbert-AI Reranker Integration**: Added support for "jinaai/jina-colbert-v2" as a reranker, enhancing search relevance and accuracy. Note: it may not function at all on low-spec computers.
+- **🕸️ ENABLE_WEBSOCKET_SUPPORT**: Added environment variable for instances to ignore websocket upgrades, stabilizing connections on platforms with websocket issues.
+- **🔊 Azure Speech Service Integration**: Added support for Azure Speech services for Text-to-Speech (TTS).
+- **🎚️ Customizable Playback Speed**: Playback speed control is now available in Call mode settings, allowing users to adjust audio playback speed to their preferences.
+- **🧠 Enhanced Error Messaging**: System now displays helpful error messages directly to users during chat completion issues.
+- **📂 Save Model as Transparent PNG**: Model profile images are now saved as PNGs, supporting transparency and improving visual integration.
+- **📱 iPhone Compatibility Adjustments**: Added padding to accommodate the iPhone navigation bar, improving UI display on these devices.
+- **🔗 Secure Response Headers**: Implemented security response headers, bolstering web application security.
+- **🔧 Enhanced AUTOMATIC1111 Settings**: Users can now configure 'CFG Scale', 'Sampler', and 'Scheduler' parameters directly in the admin settings, enhancing workflow flexibility without source code modifications.
+- **🌍 i18n Updates**: Enhanced translations for Chinese, Ukrainian, Russian, and French, fostering a better localized experience.
+
+### Fixed
+
+- **🛠️ Chat Message Deletion**: Resolved issues with chat message deletion, ensuring a smoother user interaction and system stability.
+- **🔢 Ordered List Numbering**: Fixed the incorrect ordering in lists.
+
+### Changed
+
+- **🎨 Transparent Icon Handling**: Allowed model icons to be displayed on transparent backgrounds, improving UI aesthetics.
+- **📝 Improved RAG Template**: Enhanced Retrieval-Augmented Generation template, optimizing context handling and error checking for more precise operation.
+
+## [0.3.21] - 2024-09-08
+
+### Added
+
+- **📊 Document Count Display**: Now displays the total number of documents directly within the dashboard.
+- **🚀 Ollama Embed API Endpoint**: Enabled /api/embed endpoint proxy support.
+
+### Fixed
+
+- **🐳 Docker Launch Issue**: Resolved the problem preventing Open-WebUI from launching correctly when using Docker.
+
+### Changed
+
+- **🔍 Enhanced Search Prompts**: Improved the search query generation prompts for better accuracy and user interaction, enhancing the overall search experience.
+
+## [0.3.20] - 2024-09-07
+
+### Added
+
+- **🌐 Translation Update**: Updated Catalan translations to improve user experience for Catalan speakers.
+
+### Fixed
+
+- **📄 PDF Download**: Resolved a configuration issue with fonts directory, ensuring PDFs are now downloaded with the correct formatting.
+- **🛠️ Installation of Tools & Functions Requirements**: Fixed a bug where necessary requirements for tools and functions were not properly installing.
+- **🔗 Inline Image Link Rendering**: Enabled rendering of images directly from links in chat.
+- **📞 Post-Call User Interface Cleanup**: Adjusted UI behavior to automatically close chat controls after a voice call ends, reducing screen clutter.
+- **🎙️ Microphone Deactivation Post-Call**: Addressed an issue where the microphone remained active after calls.
+- **✍️ Markdown Spacing Correction**: Corrected spacing in Markdown rendering, ensuring text appears neatly and as expected.
+- **🔄 Message Re-rendering**: Fixed an issue causing all response messages to re-render with each new message, now improving chat performance.
+
+### Changed
+
+- **🌐 Refined Web Search Integration**: Deprecated the Search Query Generation Prompt threshold; introduced a toggle button for "Enable Web Search Query Generation" allowing users to opt-in to using web search more judiciously.
+- **📝 Default Prompt Templates Update**: Emptied environment variable templates for search and title generation now default to the Open WebUI default prompt templates, simplifying configuration efforts.
+
+## [0.3.19] - 2024-09-05
+
+### Added
+
+- **🌐 Translation Update**: Improved Chinese translations.
+
+### Fixed
+
+- **📂 DATA_DIR Overriding**: Fixed an issue to avoid overriding DATA_DIR, preventing errors when directories are set identically, ensuring smoother operation and data management.
+- **🛠️ Frontmatter Extraction**: Fixed the extraction process for frontmatter in tools and functions.
+
+### Changed
+
+- **🎨 UI Styling**: Refined the user interface styling for enhanced visual coherence and user experience.
+
+## [0.3.18] - 2024-09-04
+
+### Added
+
+- **🛠️ Direct Database Execution for Tools & Functions**: Enhanced the execution of Python files for tools and functions, now directly loading from the database for a more streamlined backend process.
+
+### Fixed
+
+- **🔄 Automatic Rewrite of Import Statements in Tools & Functions**: Tool and function scripts that import 'utils', 'apps', 'main', 'config' will now automatically rename these with 'open_webui.', ensuring compatibility and consistency across different modules.
+- **🎨 Styling Adjustments**: Minor fixes in the visual styling to improve user experience and interface consistency.
+
+## [0.3.17] - 2024-09-04
+
+### Added
+
+- **🔄 Import/Export Configuration**: Users can now import and export webui configurations from admin settings > Database, simplifying setup replication across systems.
+- **🌍 Web Search via URL Parameter**: Added support for activating web search directly through URL by setting 'web-search=true'.
+- **🌐 SearchApi Integration**: Added support for SearchApi as an alternative web search provider, enhancing search capabilities within the platform.
+- **🔍 Literal Type Support in Tools**: Tools now support the Literal type.
+- **🌍 Updated Translations**: Improved translations for Chinese, Ukrainian, and Catalan.
+
+### Fixed
+
+- **🔧 Pip Install Issue**: Resolved the issue where pip install failed due to missing 'alembic.ini', ensuring smoother installation processes.
+- **🌃 Automatic Theme Update**: Fixed an issue where the color theme did not update dynamically with system changes.
+- **🛠️ User Agent in ComfyUI**: Added default headers in ComfyUI to fix access issues, improving reliability in network communications.
+- **🔄 Missing Chat Completion Response Headers**: Ensured proper return of proxied response headers during chat completion, improving API reliability.
+- **🔗 Websocket Connection Prioritization**: Modified socket.io configuration to prefer websockets and more reliably fallback to polling, enhancing connection stability.
+- **🎭 Accessibility Enhancements**: Added missing ARIA labels for buttons, improving accessibility for visually impaired users.
+- **⚖️ Advanced Parameter**: Fixed an issue ensuring that advanced parameters are correctly applied in all scenarios, ensuring consistent behavior of user-defined settings.
+
+### Changed
+
+- **🔁 Namespace Reorganization**: Reorganized all Python files under the 'open_webui' namespace to streamline the project structure and improve maintainability. Tools and functions importing from 'utils' should now use 'open_webui.utils'.
+- **🚧 Dependency Updates**: Updated several backend dependencies like 'aiohttp', 'authlib', 'duckduckgo-search', 'flask-cors', and 'langchain' to their latest versions, enhancing performance and security.
+
 ## [0.3.16] - 2024-08-27
 
 ### Added

Dockerfile

@@ -17,7 +17,7 @@ ARG UID=0
 ARG GID=0
 
 ######## WebUI frontend ########
-FROM --platform=$BUILDPLATFORM node:21-alpine3.19 as build
+FROM --platform=$BUILDPLATFORM node:22-alpine3.20 AS build
 ARG BUILD_HASH
 
 WORKDIR /app
@@ -30,7 +30,7 @@ ENV APP_BUILD_HASH=${BUILD_HASH}
 RUN npm run build
 
 ######## WebUI backend ########
-FROM python:3.11-slim-bookworm as base
+FROM python:3.11-slim-bookworm AS base
 
 # Use args
 ARG USE_CUDA
@@ -74,11 +74,15 @@ ENV RAG_EMBEDDING_MODEL="$USE_EMBEDDING_MODEL_DOCKER" \
 
 ## Hugging Face download cache ##
 ENV HF_HOME="/app/backend/data/cache/embedding/models"
+
+## Torch Extensions ##
+# ENV TORCH_EXTENSIONS_DIR="/.cache/torch_extensions"
+
 #### Other models ##########################################################
 
 WORKDIR /app/backend
 
-ENV HOME /root
+ENV HOME=/root
 # Create user and group if not root
 RUN if [ $UID -ne 0 ]; then \
     if [ $GID -ne 0 ]; then \
@@ -96,7 +100,7 @@ RUN chown -R $UID:$GID /app $HOME
 RUN if [ "$USE_OLLAMA" = "true" ]; then \
     apt-get update && \
     # Install pandoc and netcat
-    apt-get install -y --no-install-recommends pandoc netcat-openbsd curl && \
+    apt-get install -y --no-install-recommends git build-essential pandoc netcat-openbsd curl && \
     apt-get install -y --no-install-recommends gcc python3-dev && \
     # for RAG OCR
     apt-get install -y --no-install-recommends ffmpeg libsm6 libxext6 && \
@@ -109,7 +113,7 @@ RUN if [ "$USE_OLLAMA" = "true" ]; then \
     else \
     apt-get update && \
     # Install pandoc, netcat and gcc
-    apt-get install -y --no-install-recommends pandoc gcc netcat-openbsd curl jq && \
+    apt-get install -y --no-install-recommends git build-essential pandoc gcc netcat-openbsd curl jq && \
     apt-get install -y --no-install-recommends gcc python3-dev && \
     # for RAG OCR
     apt-get install -y --no-install-recommends ffmpeg libsm6 libxext6 && \
@@ -157,5 +161,6 @@ USER $UID:$GID
 
 ARG BUILD_HASH
 ENV WEBUI_BUILD_VERSION=${BUILD_HASH}
+ENV DOCKER=true
 
 CMD [ "bash", "start.sh"]

README.md

@@ -1,4 +1,4 @@
-# Open WebUI (Formerly Ollama WebUI) 👋
+# Open WebUI 👋
 
 ![GitHub stars](https://img.shields.io/github/stars/open-webui/open-webui?style=social)
 ![GitHub forks](https://img.shields.io/github/forks/open-webui/open-webui?style=social)
@@ -37,7 +37,7 @@ Open WebUI is an [extensible](https://github.com/open-webui/pipelines), feature-
 
 - 📚 **Local RAG Integration**: Dive into the future of chat interactions with groundbreaking Retrieval Augmented Generation (RAG) support. This feature seamlessly integrates document interactions into your chat experience. You can load documents directly into the chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
 
-- 🔍 **Web Search for RAG**: Perform web searches using providers like `SearXNG`, `Google PSE`, `Brave Search`, `serpstack`, `serper`, `Serply`, `DuckDuckGo` and `TavilySearch` and inject the results directly into your chat experience.
+- 🔍 **Web Search for RAG**: Perform web searches using providers like `SearXNG`, `Google PSE`, `Brave Search`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `TavilySearch` and `SearchApi` and inject the results directly into your chat experience.
 
 - 🌐 **Web Browsing Capability**: Seamlessly integrate websites into your chat experience using the `#` command followed by a URL. This feature allows you to incorporate web content directly into your conversations, enhancing the richness and depth of your interactions.
 
@@ -59,11 +59,31 @@ Don't forget to explore our sibling project, [Open WebUI Community](https://open
 
 ## How to Install 🚀
 
-> [!NOTE]  
-> Please note that for certain Docker environments, additional configurations might be needed. If you encounter any connection issues, our detailed guide on [Open WebUI Documentation](https://docs.openwebui.com/) is ready to assist you.
+### Installation via Python pip 🐍
+
+Open WebUI can be installed using pip, the Python package installer. Before proceeding, ensure you're using **Python 3.11** to avoid compatibility issues.
+
+1. **Install Open WebUI**:
+   Open your terminal and run the following command to install Open WebUI:
+
+   ```bash
+   pip install open-webui
+   ```
+
+2. **Running Open WebUI**:
+   After installation, you can start Open WebUI by executing:
+
+   ```bash
+   open-webui serve
+   ```
+
+This will start the Open WebUI server, which you can access at [http://localhost:8080](http://localhost:8080)
 
 ### Quick Start with Docker 🐳
 
+> [!NOTE]  
+> Please note that for certain Docker environments, additional configurations might be needed. If you encounter any connection issues, our detailed guide on [Open WebUI Documentation](https://docs.openwebui.com/) is ready to assist you.
+
 > [!WARNING]
 > When using Docker to install Open WebUI, make sure to include the `-v open-webui:/app/backend/data` in your Docker command. This step is crucial as it ensures your database is properly mounted and prevents any loss of data.
 
@@ -86,7 +106,7 @@ Don't forget to explore our sibling project, [Open WebUI Community](https://open
   docker run -d -p 3000:8080 -e OLLAMA_BASE_URL=https://example.com -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
   ```
 
-  - **To run Open WebUI with Nvidia GPU support**, use this command:
+- **To run Open WebUI with Nvidia GPU support**, use this command:
 
   ```bash
   docker run -d -p 3000:8080 --gpus all --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:cuda
@@ -150,7 +170,7 @@ docker run --rm --volume /var/run/docker.sock:/var/run/docker.sock containrrr/wa
 
 In the last part of the command, replace `open-webui` with your container name if it is different.
 
-Check our Migration Guide available in our [Open WebUI Documentation](https://docs.openwebui.com/migration/).
+Check our Migration Guide available in our [Open WebUI Documentation](https://docs.openwebui.com/tutorials/migration/).
 
 ### Using the Dev Branch 🌙
 
@@ -200,4 +220,4 @@ If you have any questions, suggestions, or need assistance, please open an issue
 
 ---
 
-Created by [Timothy J. Baek](https://github.com/tjbck) - Let's make Open WebUI even more amazing together! 💪
+Created by [Timothy Jaeryang Baek](https://github.com/tjbck) - Let's make Open WebUI even more amazing together! 💪

backend/.gitignore

@@ -8,9 +8,5 @@ _test
 Pipfile
 !/data
 /data/*
-!/data/litellm
-/data/litellm/*
-!data/litellm/config.yaml
-
-!data/config.json
+/open_webui/data/*
 .webui_secret_key

backend/apps/socket/main.py

@@ -1,171 +0,0 @@
-import socketio
-import asyncio
-
-
-from apps.webui.models.users import Users
-from utils.utils import decode_token
-
-sio = socketio.AsyncServer(cors_allowed_origins=[], async_mode="asgi")
-app = socketio.ASGIApp(sio, socketio_path="/ws/socket.io")
-
-# Dictionary to maintain the user pool
-
-SESSION_POOL = {}
-USER_POOL = {}
-USAGE_POOL = {}
-# Timeout duration in seconds
-TIMEOUT_DURATION = 3
-
-
-@sio.event
-async def connect(sid, environ, auth):
-    user = None
-    if auth and "token" in auth:
-        data = decode_token(auth["token"])
-
-        if data is not None and "id" in data:
-            user = Users.get_user_by_id(data["id"])
-
-        if user:
-            SESSION_POOL[sid] = user.id
-            if user.id in USER_POOL:
-                USER_POOL[user.id].append(sid)
-            else:
-                USER_POOL[user.id] = [sid]
-
-            print(f"user {user.name}({user.id}) connected with session ID {sid}")
-
-            await sio.emit("user-count", {"count": len(set(USER_POOL))})
-            await sio.emit("usage", {"models": get_models_in_use()})
-
-
-@sio.on("user-join")
-async def user_join(sid, data):
-    print("user-join", sid, data)
-
-    auth = data["auth"] if "auth" in data else None
-    if not auth or "token" not in auth:
-        return
-
-    data = decode_token(auth["token"])
-    if data is None or "id" not in data:
-        return
-
-    user = Users.get_user_by_id(data["id"])
-    if not user:
-        return
-
-    SESSION_POOL[sid] = user.id
-    if user.id in USER_POOL:
-        USER_POOL[user.id].append(sid)
-    else:
-        USER_POOL[user.id] = [sid]
-
-    print(f"user {user.name}({user.id}) connected with session ID {sid}")
-
-    await sio.emit("user-count", {"count": len(set(USER_POOL))})
-
-
-@sio.on("user-count")
-async def user_count(sid):
-    await sio.emit("user-count", {"count": len(set(USER_POOL))})
-
-
-def get_models_in_use():
-    # Aggregate all models in use
-    models_in_use = []
-    for model_id, data in USAGE_POOL.items():
-        models_in_use.append(model_id)
-
-    return models_in_use
-
-
-@sio.on("usage")
-async def usage(sid, data):
-    model_id = data["model"]
-
-    # Cancel previous callback if there is one
-    if model_id in USAGE_POOL:
-        USAGE_POOL[model_id]["callback"].cancel()
-
-    # Store the new usage data and task
-
-    if model_id in USAGE_POOL:
-        USAGE_POOL[model_id]["sids"].append(sid)
-        USAGE_POOL[model_id]["sids"] = list(set(USAGE_POOL[model_id]["sids"]))
-
-    else:
-        USAGE_POOL[model_id] = {"sids": [sid]}
-
-    # Schedule a task to remove the usage data after TIMEOUT_DURATION
-    USAGE_POOL[model_id]["callback"] = asyncio.create_task(
-        remove_after_timeout(sid, model_id)
-    )
-
-    # Broadcast the usage data to all clients
-    await sio.emit("usage", {"models": get_models_in_use()})
-
-
-async def remove_after_timeout(sid, model_id):
-    try:
-        await asyncio.sleep(TIMEOUT_DURATION)
-        if model_id in USAGE_POOL:
-            print(USAGE_POOL[model_id]["sids"])
-            USAGE_POOL[model_id]["sids"].remove(sid)
-            USAGE_POOL[model_id]["sids"] = list(set(USAGE_POOL[model_id]["sids"]))
-
-            if len(USAGE_POOL[model_id]["sids"]) == 0:
-                del USAGE_POOL[model_id]
-
-            # Broadcast the usage data to all clients
-            await sio.emit("usage", {"models": get_models_in_use()})
-    except asyncio.CancelledError:
-        # Task was cancelled due to new 'usage' event
-        pass
-
-
-@sio.event
-async def disconnect(sid):
-    if sid in SESSION_POOL:
-        user_id = SESSION_POOL[sid]
-        del SESSION_POOL[sid]
-
-        USER_POOL[user_id].remove(sid)
-
-        if len(USER_POOL[user_id]) == 0:
-            del USER_POOL[user_id]
-
-        await sio.emit("user-count", {"count": len(USER_POOL)})
-    else:
-        print(f"Unknown session ID {sid} disconnected")
-
-
-def get_event_emitter(request_info):
-    async def __event_emitter__(event_data):
-        await sio.emit(
-            "chat-events",
-            {
-                "chat_id": request_info["chat_id"],
-                "message_id": request_info["message_id"],
-                "data": event_data,
-            },
-            to=request_info["session_id"],
-        )
-
-    return __event_emitter__
-
-
-def get_event_call(request_info):
-    async def __event_call__(event_data):
-        response = await sio.call(
-            "chat-events",
-            {
-                "chat_id": request_info["chat_id"],
-                "message_id": request_info["message_id"],
-                "data": event_data,
-            },
-            to=request_info["session_id"],
-        )
-        return response
-
-    return __event_call__

backend/apps/webui/utils.py

@@ -1,124 +0,0 @@
-from importlib import util
-import os
-import re
-import sys
-import subprocess
-
-
-from apps.webui.models.tools import Tools
-from apps.webui.models.functions import Functions
-from config import TOOLS_DIR, FUNCTIONS_DIR
-
-
-def extract_frontmatter(file_path):
-    """
-    Extract frontmatter as a dictionary from the specified file path.
-    """
-    frontmatter = {}
-    frontmatter_started = False
-    frontmatter_ended = False
-    frontmatter_pattern = re.compile(r"^\s*([a-z_]+):\s*(.*)\s*$", re.IGNORECASE)
-
-    try:
-        with open(file_path, "r", encoding="utf-8") as file:
-            first_line = file.readline()
-            if first_line.strip() != '"""':
-                # The file doesn't start with triple quotes
-                return {}
-
-            frontmatter_started = True
-
-            for line in file:
-                if '"""' in line:
-                    if frontmatter_started:
-                        frontmatter_ended = True
-                        break
-
-                if frontmatter_started and not frontmatter_ended:
-                    match = frontmatter_pattern.match(line)
-                    if match:
-                        key, value = match.groups()
-                        frontmatter[key.strip()] = value.strip()
-
-    except FileNotFoundError:
-        print(f"Error: The file {file_path} does not exist.")
-        return {}
-    except Exception as e:
-        print(f"An error occurred: {e}")
-        return {}
-
-    return frontmatter
-
-
-def load_toolkit_module_by_id(toolkit_id):
-    toolkit_path = os.path.join(TOOLS_DIR, f"{toolkit_id}.py")
-
-    if not os.path.exists(toolkit_path):
-        tool = Tools.get_tool_by_id(toolkit_id)
-        if tool:
-            with open(toolkit_path, "w") as file:
-                file.write(tool.content)
-        else:
-            raise Exception(f"Toolkit not found: {toolkit_id}")
-
-    spec = util.spec_from_file_location(toolkit_id, toolkit_path)
-    module = util.module_from_spec(spec)
-    frontmatter = extract_frontmatter(toolkit_path)
-
-    try:
-        install_frontmatter_requirements(frontmatter.get("requirements", ""))
-        spec.loader.exec_module(module)
-        print(f"Loaded module: {module.__name__}")
-        if hasattr(module, "Tools"):
-            return module.Tools(), frontmatter
-        else:
-            raise Exception("No Tools class found")
-    except Exception as e:
-        print(f"Error loading module: {toolkit_id}")
-        # Move the file to the error folder
-        os.rename(toolkit_path, f"{toolkit_path}.error")
-        raise e
-
-
-def load_function_module_by_id(function_id):
-    function_path = os.path.join(FUNCTIONS_DIR, f"{function_id}.py")
-
-    if not os.path.exists(function_path):
-        function = Functions.get_function_by_id(function_id)
-        if function:
-            with open(function_path, "w") as file:
-                file.write(function.content)
-        else:
-            raise Exception(f"Function not found: {function_id}")
-
-    spec = util.spec_from_file_location(function_id, function_path)
-    module = util.module_from_spec(spec)
-    frontmatter = extract_frontmatter(function_path)
-
-    try:
-        install_frontmatter_requirements(frontmatter.get("requirements", ""))
-        spec.loader.exec_module(module)
-        print(f"Loaded module: {module.__name__}")
-        if hasattr(module, "Pipe"):
-            return module.Pipe(), "pipe", frontmatter
-        elif hasattr(module, "Filter"):
-            return module.Filter(), "filter", frontmatter
-        elif hasattr(module, "Action"):
-            return module.Action(), "action", frontmatter
-        else:
-            raise Exception("No Function class found")
-    except Exception as e:
-        print(f"Error loading module: {function_id}")
-        # Move the file to the error folder
-        os.rename(function_path, f"{function_path}.error")
-        raise e
-
-
-def install_frontmatter_requirements(requirements):
-    if requirements:
-        req_list = [req.strip() for req in requirements.split(",")]
-        for req in req_list:
-            print(f"Installing requirement: {req}")
-            subprocess.check_call([sys.executable, "-m", "pip", "install", req])
-    else:
-        print("No requirements found in frontmatter.")

backend/data/litellm/config.yaml

@@ -1,4 +0,0 @@
-general_settings: {}
-litellm_settings: {}
-model_list: []
-router_settings: {}

backend/data/readme.txt

@@ -1 +1 @@
-dir for backend files (db, documents, etc.)
+docker dir for backend files (db, documents, etc.)

backend/dev.sh

@@ -1,2 +1,2 @@
 PORT="${PORT:-8080}"
-uvicorn main:app --port $PORT --host 0.0.0.0 --forwarded-allow-ips '*' --reload
+uvicorn open_webui.main:app --port $PORT --host 0.0.0.0 --forwarded-allow-ips '*' --reload

backend/open_webui/__init__.py

@@ -9,8 +9,6 @@ import uvicorn
 app = typer.Typer()
 
 KEY_FILE = Path.cwd() / ".webui_secret_key"
-if (frontend_build_dir := Path(__file__).parent / "frontend").exists():
-    os.environ["FRONTEND_BUILD_DIR"] = str(frontend_build_dir)
 
 
 @app.command()
@@ -18,6 +16,7 @@ def serve(
     host: str = "0.0.0.0",
     port: int = 8080,
 ):
+    os.environ["FROM_INIT_PY"] = "true"
     if os.getenv("WEBUI_SECRET_KEY") is None:
         typer.echo(
             "Loading WEBUI_SECRET_KEY from file, not provided as an environment variable."
@@ -40,9 +39,23 @@ def serve(
                 "/usr/local/lib/python3.11/site-packages/nvidia/cudnn/lib",
             ]
         )
-    import main  # we need set environment variables before importing main
+        try:
+            import torch
 
-    uvicorn.run(main.app, host=host, port=port, forwarded_allow_ips="*")
+            assert torch.cuda.is_available(), "CUDA not available"
+            typer.echo("CUDA seems to be working")
+        except Exception as e:
+            typer.echo(
+                "Error when testing CUDA but USE_CUDA_DOCKER is true. "
+                "Resetting USE_CUDA_DOCKER to false and removing "
+                f"LD_LIBRARY_PATH modifications: {e}"
+            )
+            os.environ["USE_CUDA_DOCKER"] = "false"
+            os.environ["LD_LIBRARY_PATH"] = ":".join(LD_LIBRARY_PATH)
+
+    import open_webui.main  # we need set environment variables before importing main
+
+    uvicorn.run(open_webui.main.app, host=host, port=port, forwarded_allow_ips="*")
 
 
 @app.command()
@@ -52,7 +65,11 @@ def dev(
     reload: bool = True,
 ):
     uvicorn.run(
-        "main:app", host=host, port=port, reload=reload, forwarded_allow_ips="*"
+        "open_webui.main:app",
+        host=host,
+        port=port,
+        reload=reload,
+        forwarded_allow_ips="*",
     )
 
 

backend/open_webui/apps/audio/main.py

@@ -5,48 +5,44 @@ import os
 import uuid
 from functools import lru_cache
 from pathlib import Path
+from pydub import AudioSegment
+from pydub.silence import split_on_silence
 
 import requests
-from fastapi import (
-    FastAPI,
-    Request,
-    Depends,
-    HTTPException,
-    status,
-    UploadFile,
-    File,
+from open_webui.config import (
+    AUDIO_STT_ENGINE,
+    AUDIO_STT_MODEL,
+    AUDIO_STT_OPENAI_API_BASE_URL,
+    AUDIO_STT_OPENAI_API_KEY,
+    AUDIO_TTS_API_KEY,
+    AUDIO_TTS_ENGINE,
+    AUDIO_TTS_MODEL,
+    AUDIO_TTS_OPENAI_API_BASE_URL,
+    AUDIO_TTS_OPENAI_API_KEY,
+    AUDIO_TTS_SPLIT_ON,
+    AUDIO_TTS_VOICE,
+    AUDIO_TTS_AZURE_SPEECH_REGION,
+    AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT,
+    CACHE_DIR,
+    CORS_ALLOW_ORIGIN,
+    WHISPER_MODEL,
+    WHISPER_MODEL_AUTO_UPDATE,
+    WHISPER_MODEL_DIR,
+    AppConfig,
 )
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS, DEVICE_TYPE
+from fastapi import Depends, FastAPI, File, HTTPException, Request, UploadFile, status
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import FileResponse
 from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
+
+# Constants
+MAX_FILE_SIZE_MB = 25
+MAX_FILE_SIZE = MAX_FILE_SIZE_MB * 1024 * 1024  # Convert MB to bytes
 
-from config import (
-    SRC_LOG_LEVELS,
-    CACHE_DIR,
-    WHISPER_MODEL,
-    WHISPER_MODEL_DIR,
-    WHISPER_MODEL_AUTO_UPDATE,
-    DEVICE_TYPE,
-    AUDIO_STT_OPENAI_API_BASE_URL,
-    AUDIO_STT_OPENAI_API_KEY,
-    AUDIO_TTS_OPENAI_API_BASE_URL,
-    AUDIO_TTS_OPENAI_API_KEY,
-    AUDIO_TTS_API_KEY,
-    AUDIO_STT_ENGINE,
-    AUDIO_STT_MODEL,
-    AUDIO_TTS_ENGINE,
-    AUDIO_TTS_MODEL,
-    AUDIO_TTS_VOICE,
-    AUDIO_TTS_SPLIT_ON,
-    AppConfig,
-    CORS_ALLOW_ORIGIN,
-)
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    get_current_user,
-    get_verified_user,
-    get_admin_user,
-)
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["AUDIO"])
@@ -75,6 +71,9 @@ app.state.config.TTS_VOICE = AUDIO_TTS_VOICE
 app.state.config.TTS_API_KEY = AUDIO_TTS_API_KEY
 app.state.config.TTS_SPLIT_ON = AUDIO_TTS_SPLIT_ON
 
+app.state.config.TTS_AZURE_SPEECH_REGION = AUDIO_TTS_AZURE_SPEECH_REGION
+app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT
+
 # setting device type for whisper model
 whisper_device_type = DEVICE_TYPE if DEVICE_TYPE and DEVICE_TYPE == "cuda" else "cpu"
 log.info(f"whisper_device_type: {whisper_device_type}")
@@ -91,6 +90,8 @@ class TTSConfigForm(BaseModel):
     MODEL: str
     VOICE: str
     SPLIT_ON: str
+    AZURE_SPEECH_REGION: str
+    AZURE_SPEECH_OUTPUT_FORMAT: str
 
 
 class STTConfigForm(BaseModel):
@@ -143,6 +144,8 @@ async def get_audio_config(user=Depends(get_admin_user)):
             "MODEL": app.state.config.TTS_MODEL,
             "VOICE": app.state.config.TTS_VOICE,
             "SPLIT_ON": app.state.config.TTS_SPLIT_ON,
+            "AZURE_SPEECH_REGION": app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_OUTPUT_FORMAT": app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
         },
         "stt": {
             "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
@@ -164,6 +167,10 @@ async def update_audio_config(
     app.state.config.TTS_MODEL = form_data.tts.MODEL
     app.state.config.TTS_VOICE = form_data.tts.VOICE
     app.state.config.TTS_SPLIT_ON = form_data.tts.SPLIT_ON
+    app.state.config.TTS_AZURE_SPEECH_REGION = form_data.tts.AZURE_SPEECH_REGION
+    app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = (
+        form_data.tts.AZURE_SPEECH_OUTPUT_FORMAT
+    )
 
     app.state.config.STT_OPENAI_API_BASE_URL = form_data.stt.OPENAI_API_BASE_URL
     app.state.config.STT_OPENAI_API_KEY = form_data.stt.OPENAI_API_KEY
@@ -179,6 +186,8 @@ async def update_audio_config(
             "MODEL": app.state.config.TTS_MODEL,
             "VOICE": app.state.config.TTS_VOICE,
             "SPLIT_ON": app.state.config.TTS_SPLIT_ON,
+            "AZURE_SPEECH_REGION": app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_OUTPUT_FORMAT": app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
         },
         "stt": {
             "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
@@ -211,7 +220,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
             body = json.loads(body)
             body["model"] = app.state.config.TTS_MODEL
             body = json.dumps(body).encode("utf-8")
-        except Exception as e:
+        except Exception:
             pass
 
         r = None
@@ -314,68 +323,114 @@ async def speech(request: Request, user=Depends(get_verified_user)):
                 detail=error_detail,
             )
 
+    elif app.state.config.TTS_ENGINE == "azure":
+        payload = None
+        try:
+            payload = json.loads(body.decode("utf-8"))
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(status_code=400, detail="Invalid JSON payload")
 
-@app.post("/transcriptions")
-def transcribe(
-    file: UploadFile = File(...),
-    user=Depends(get_current_user),
-):
-    log.info(f"file.content_type: {file.content_type}")
+        region = app.state.config.TTS_AZURE_SPEECH_REGION
+        language = app.state.config.TTS_VOICE
+        locale = "-".join(app.state.config.TTS_VOICE.split("-")[:1])
+        output_format = app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT
+        url = f"https://{region}.tts.speech.microsoft.com/cognitiveservices/v1"
 
-    if file.content_type not in ["audio/mpeg", "audio/wav"]:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.FILE_NOT_SUPPORTED,
-        )
+        headers = {
+            "Ocp-Apim-Subscription-Key": app.state.config.TTS_API_KEY,
+            "Content-Type": "application/ssml+xml",
+            "X-Microsoft-OutputFormat": output_format,
+        }
 
-    try:
-        ext = file.filename.split(".")[-1]
+        data = f"""<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="{locale}">
+                <voice name="{language}">{payload["input"]}</voice>
+            </speak>"""
 
-        id = uuid.uuid4()
-        filename = f"{id}.{ext}"
+        response = requests.post(url, headers=headers, data=data)
 
-        file_dir = f"{CACHE_DIR}/audio/transcriptions"
-        os.makedirs(file_dir, exist_ok=True)
-        file_path = f"{file_dir}/{filename}"
-
-        print(filename)
-
-        contents = file.file.read()
-        with open(file_path, "wb") as f:
-            f.write(contents)
-            f.close()
-
-        if app.state.config.STT_ENGINE == "":
-            from faster_whisper import WhisperModel
-
-            whisper_kwargs = {
-                "model_size_or_path": WHISPER_MODEL,
-                "device": whisper_device_type,
-                "compute_type": "int8",
-                "download_root": WHISPER_MODEL_DIR,
-                "local_files_only": not WHISPER_MODEL_AUTO_UPDATE,
-            }
-
-            log.debug(f"whisper_kwargs: {whisper_kwargs}")
-
-            try:
-                model = WhisperModel(**whisper_kwargs)
-            except Exception:
-                log.warning(
-                    "WhisperModel initialization failed, attempting download with local_files_only=False"
-                )
-                whisper_kwargs["local_files_only"] = False
-                model = WhisperModel(**whisper_kwargs)
-
-            segments, info = model.transcribe(file_path, beam_size=5)
-            log.info(
-                "Detected language '%s' with probability %f"
-                % (info.language, info.language_probability)
+        if response.status_code == 200:
+            with open(file_path, "wb") as f:
+                f.write(response.content)
+            return FileResponse(file_path)
+        else:
+            log.error(f"Error synthesizing speech - {response.reason}")
+            raise HTTPException(
+                status_code=500, detail=f"Error synthesizing speech - {response.reason}"
             )
 
-            transcript = "".join([segment.text for segment in list(segments)])
 
-            data = {"text": transcript.strip()}
+def transcribe(file_path):
+    print("transcribe", file_path)
+    filename = os.path.basename(file_path)
+    file_dir = os.path.dirname(file_path)
+    id = filename.split(".")[0]
+
+    if app.state.config.STT_ENGINE == "":
+        from faster_whisper import WhisperModel
+
+        whisper_kwargs = {
+            "model_size_or_path": WHISPER_MODEL,
+            "device": whisper_device_type,
+            "compute_type": "int8",
+            "download_root": WHISPER_MODEL_DIR,
+            "local_files_only": not WHISPER_MODEL_AUTO_UPDATE,
+        }
+
+        log.debug(f"whisper_kwargs: {whisper_kwargs}")
+
+        try:
+            model = WhisperModel(**whisper_kwargs)
+        except Exception:
+            log.warning(
+                "WhisperModel initialization failed, attempting download with local_files_only=False"
+            )
+            whisper_kwargs["local_files_only"] = False
+            model = WhisperModel(**whisper_kwargs)
+
+        segments, info = model.transcribe(file_path, beam_size=5)
+        log.info(
+            "Detected language '%s' with probability %f"
+            % (info.language, info.language_probability)
+        )
+
+        transcript = "".join([segment.text for segment in list(segments)])
+
+        data = {"text": transcript.strip()}
+
+        # save the transcript to a json file
+        transcript_file = f"{file_dir}/{id}.json"
+        with open(transcript_file, "w") as f:
+            json.dump(data, f)
+
+        print(data)
+        return data
+    elif app.state.config.STT_ENGINE == "openai":
+        if is_mp4_audio(file_path):
+            print("is_mp4_audio")
+            os.rename(file_path, file_path.replace(".wav", ".mp4"))
+            # Convert MP4 audio file to WAV format
+            convert_mp4_to_wav(file_path.replace(".wav", ".mp4"), file_path)
+
+        headers = {"Authorization": f"Bearer {app.state.config.STT_OPENAI_API_KEY}"}
+
+        files = {"file": (filename, open(file_path, "rb"))}
+        data = {"model": app.state.config.STT_MODEL}
+
+        print(files, data)
+
+        r = None
+        try:
+            r = requests.post(
+                url=f"{app.state.config.STT_OPENAI_API_BASE_URL}/audio/transcriptions",
+                headers=headers,
+                files=files,
+                data=data,
+            )
+
+            r.raise_for_status()
+
+            data = r.json()
 
             # save the transcript to a json file
             transcript_file = f"{file_dir}/{id}.json"
@@ -383,58 +438,82 @@ def transcribe(
                 json.dump(data, f)
 
             print(data)
+            return data
+        except Exception as e:
+            log.exception(e)
+            error_detail = "Open WebUI: Server Connection Error"
+            if r is not None:
+                try:
+                    res = r.json()
+                    if "error" in res:
+                        error_detail = f"External: {res['error']['message']}"
+                except Exception:
+                    error_detail = f"External: {e}"
+
+            raise error_detail
+
+
+@app.post("/transcriptions")
+def transcription(
+    file: UploadFile = File(...),
+    user=Depends(get_verified_user),
+):
+    log.info(f"file.content_type: {file.content_type}")
+
+    if file.content_type not in ["audio/mpeg", "audio/wav", "audio/ogg", "audio/x-m4a"]:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.FILE_NOT_SUPPORTED,
+        )
+
+    try:
+        ext = file.filename.split(".")[-1]
+        id = uuid.uuid4()
+
+        filename = f"{id}.{ext}"
+        contents = file.file.read()
+
+        file_dir = f"{CACHE_DIR}/audio/transcriptions"
+        os.makedirs(file_dir, exist_ok=True)
+        file_path = f"{file_dir}/{filename}"
+
+        with open(file_path, "wb") as f:
+            f.write(contents)
+
+        try:
+            if os.path.getsize(file_path) > MAX_FILE_SIZE:  # file is bigger than 25MB
+                log.debug(f"File size is larger than {MAX_FILE_SIZE_MB}MB")
+                audio = AudioSegment.from_file(file_path)
+                audio = audio.set_frame_rate(16000).set_channels(1)  # Compress audio
+                compressed_path = f"{file_dir}/{id}_compressed.opus"
+                audio.export(compressed_path, format="opus", bitrate="32k")
+                log.debug(f"Compressed audio to {compressed_path}")
+                file_path = compressed_path
+
+                if (
+                    os.path.getsize(file_path) > MAX_FILE_SIZE
+                ):  # Still larger than 25MB after compression
+                    log.debug(
+                        f"Compressed file size is still larger than {MAX_FILE_SIZE_MB}MB: {os.path.getsize(file_path)}"
+                    )
+                    raise HTTPException(
+                        status_code=status.HTTP_400_BAD_REQUEST,
+                        detail=ERROR_MESSAGES.FILE_TOO_LARGE(
+                            size=f"{MAX_FILE_SIZE_MB}MB"
+                        ),
+                    )
+
+                data = transcribe(file_path)
+            else:
+                data = transcribe(file_path)
 
             return data
-
-        elif app.state.config.STT_ENGINE == "openai":
-            if is_mp4_audio(file_path):
-                print("is_mp4_audio")
-                os.rename(file_path, file_path.replace(".wav", ".mp4"))
-                # Convert MP4 audio file to WAV format
-                convert_mp4_to_wav(file_path.replace(".wav", ".mp4"), file_path)
-
-            headers = {"Authorization": f"Bearer {app.state.config.STT_OPENAI_API_KEY}"}
-
-            files = {"file": (filename, open(file_path, "rb"))}
-            data = {"model": app.state.config.STT_MODEL}
-
-            print(files, data)
-
-            r = None
-            try:
-                r = requests.post(
-                    url=f"{app.state.config.STT_OPENAI_API_BASE_URL}/audio/transcriptions",
-                    headers=headers,
-                    files=files,
-                    data=data,
-                )
-
-                r.raise_for_status()
-
-                data = r.json()
-
-                # save the transcript to a json file
-                transcript_file = f"{file_dir}/{id}.json"
-                with open(transcript_file, "w") as f:
-                    json.dump(data, f)
-
-                print(data)
-                return data
-            except Exception as e:
-                log.exception(e)
-                error_detail = "Open WebUI: Server Connection Error"
-                if r is not None:
-                    try:
-                        res = r.json()
-                        if "error" in res:
-                            error_detail = f"External: {res['error']['message']}"
-                    except Exception:
-                        error_detail = f"External: {e}"
-
-                raise HTTPException(
-                    status_code=r.status_code if r != None else 500,
-                    detail=error_detail,
-                )
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(
+                status_code=status.HTTP_400_BAD_REQUEST,
+                detail=ERROR_MESSAGES.DEFAULT(e),
+            )
 
     except Exception as e:
         log.exception(e)
@@ -456,7 +535,7 @@ def get_available_models() -> list[dict]:
 
         try:
             response = requests.get(
-                "https://api.elevenlabs.io/v1/models", headers=headers
+                "https://api.elevenlabs.io/v1/models", headers=headers, timeout=5
             )
             response.raise_for_status()
             models = response.json()
@@ -488,9 +567,24 @@ def get_available_voices() -> dict:
     elif app.state.config.TTS_ENGINE == "elevenlabs":
         try:
             ret = get_elevenlabs_voices()
-        except Exception as e:
+        except Exception:
             # Avoided @lru_cache with exception
             pass
+    elif app.state.config.TTS_ENGINE == "azure":
+        try:
+            region = app.state.config.TTS_AZURE_SPEECH_REGION
+            url = f"https://{region}.tts.speech.microsoft.com/cognitiveservices/voices/list"
+            headers = {"Ocp-Apim-Subscription-Key": app.state.config.TTS_API_KEY}
+
+            response = requests.get(url, headers=headers)
+            response.raise_for_status()
+            voices = response.json()
+            for voice in voices:
+                ret[voice["ShortName"]] = (
+                    f"{voice['DisplayName']} ({voice['ShortName']})"
+                )
+        except requests.RequestException as e:
+            log.error(f"Error fetching voices: {str(e)}")
 
     return ret
 

backend/open_webui/apps/images/main.py

@@ -1,52 +1,45 @@
-from fastapi import (
-    FastAPI,
-    Request,
-    Depends,
-    HTTPException,
-)
-from fastapi.middleware.cors import CORSMiddleware
-from typing import Optional
-from pydantic import BaseModel
-from pathlib import Path
-import mimetypes
-import uuid
+import asyncio
 import base64
 import json
 import logging
+import mimetypes
 import re
+import uuid
+from pathlib import Path
+from typing import Optional
+
 import requests
-import asyncio
-
-from utils.utils import (
-    get_verified_user,
-    get_admin_user,
-)
-
-from apps.images.utils.comfyui import (
-    ComfyUIWorkflow,
+from open_webui.apps.images.utils.comfyui import (
     ComfyUIGenerateImageForm,
+    ComfyUIWorkflow,
     comfyui_generate_image,
 )
-
-from constants import ERROR_MESSAGES
-from config import (
-    SRC_LOG_LEVELS,
-    CACHE_DIR,
-    IMAGE_GENERATION_ENGINE,
-    ENABLE_IMAGE_GENERATION,
-    AUTOMATIC1111_BASE_URL,
+from open_webui.config import (
     AUTOMATIC1111_API_AUTH,
+    AUTOMATIC1111_BASE_URL,
+    AUTOMATIC1111_CFG_SCALE,
+    AUTOMATIC1111_SAMPLER,
+    AUTOMATIC1111_SCHEDULER,
+    CACHE_DIR,
     COMFYUI_BASE_URL,
     COMFYUI_WORKFLOW,
     COMFYUI_WORKFLOW_NODES,
-    IMAGES_OPENAI_API_BASE_URL,
-    IMAGES_OPENAI_API_KEY,
+    CORS_ALLOW_ORIGIN,
+    ENABLE_IMAGE_GENERATION,
+    IMAGE_GENERATION_ENGINE,
     IMAGE_GENERATION_MODEL,
     IMAGE_SIZE,
     IMAGE_STEPS,
-    CORS_ALLOW_ORIGIN,
+    IMAGES_OPENAI_API_BASE_URL,
+    IMAGES_OPENAI_API_KEY,
     AppConfig,
 )
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+from fastapi import Depends, FastAPI, HTTPException, Request
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["IMAGES"])
@@ -75,6 +68,9 @@ app.state.config.MODEL = IMAGE_GENERATION_MODEL
 
 app.state.config.AUTOMATIC1111_BASE_URL = AUTOMATIC1111_BASE_URL
 app.state.config.AUTOMATIC1111_API_AUTH = AUTOMATIC1111_API_AUTH
+app.state.config.AUTOMATIC1111_CFG_SCALE = AUTOMATIC1111_CFG_SCALE
+app.state.config.AUTOMATIC1111_SAMPLER = AUTOMATIC1111_SAMPLER
+app.state.config.AUTOMATIC1111_SCHEDULER = AUTOMATIC1111_SCHEDULER
 app.state.config.COMFYUI_BASE_URL = COMFYUI_BASE_URL
 app.state.config.COMFYUI_WORKFLOW = COMFYUI_WORKFLOW
 app.state.config.COMFYUI_WORKFLOW_NODES = COMFYUI_WORKFLOW_NODES
@@ -95,6 +91,9 @@ async def get_config(request: Request, user=Depends(get_admin_user)):
         "automatic1111": {
             "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
             "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
+            "AUTOMATIC1111_CFG_SCALE": app.state.config.AUTOMATIC1111_CFG_SCALE,
+            "AUTOMATIC1111_SAMPLER": app.state.config.AUTOMATIC1111_SAMPLER,
+            "AUTOMATIC1111_SCHEDULER": app.state.config.AUTOMATIC1111_SCHEDULER,
         },
         "comfyui": {
             "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
@@ -112,6 +111,9 @@ class OpenAIConfigForm(BaseModel):
 class Automatic1111ConfigForm(BaseModel):
     AUTOMATIC1111_BASE_URL: str
     AUTOMATIC1111_API_AUTH: str
+    AUTOMATIC1111_CFG_SCALE: Optional[str]
+    AUTOMATIC1111_SAMPLER: Optional[str]
+    AUTOMATIC1111_SCHEDULER: Optional[str]
 
 
 class ComfyUIConfigForm(BaseModel):
@@ -143,7 +145,23 @@ async def update_config(form_data: ConfigForm, user=Depends(get_admin_user)):
         form_data.automatic1111.AUTOMATIC1111_API_AUTH
     )
 
-    app.state.config.COMFYUI_BASE_URL = form_data.comfyui.COMFYUI_BASE_URL
+    app.state.config.AUTOMATIC1111_CFG_SCALE = (
+        float(form_data.automatic1111.AUTOMATIC1111_CFG_SCALE)
+        if form_data.automatic1111.AUTOMATIC1111_CFG_SCALE
+        else None
+    )
+    app.state.config.AUTOMATIC1111_SAMPLER = (
+        form_data.automatic1111.AUTOMATIC1111_SAMPLER
+        if form_data.automatic1111.AUTOMATIC1111_SAMPLER
+        else None
+    )
+    app.state.config.AUTOMATIC1111_SCHEDULER = (
+        form_data.automatic1111.AUTOMATIC1111_SCHEDULER
+        if form_data.automatic1111.AUTOMATIC1111_SCHEDULER
+        else None
+    )
+
+    app.state.config.COMFYUI_BASE_URL = form_data.comfyui.COMFYUI_BASE_URL.strip("/")
     app.state.config.COMFYUI_WORKFLOW = form_data.comfyui.COMFYUI_WORKFLOW
     app.state.config.COMFYUI_WORKFLOW_NODES = form_data.comfyui.COMFYUI_WORKFLOW_NODES
 
@@ -157,6 +175,9 @@ async def update_config(form_data: ConfigForm, user=Depends(get_admin_user)):
         "automatic1111": {
             "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
             "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
+            "AUTOMATIC1111_CFG_SCALE": app.state.config.AUTOMATIC1111_CFG_SCALE,
+            "AUTOMATIC1111_SAMPLER": app.state.config.AUTOMATIC1111_SAMPLER,
+            "AUTOMATIC1111_SCHEDULER": app.state.config.AUTOMATIC1111_SCHEDULER,
         },
         "comfyui": {
             "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
@@ -186,7 +207,7 @@ async def verify_url(user=Depends(get_admin_user)):
             )
             r.raise_for_status()
             return True
-        except Exception as e:
+        except Exception:
             app.state.config.ENABLED = False
             raise HTTPException(status_code=400, detail=ERROR_MESSAGES.INVALID_URL)
     elif app.state.config.ENGINE == "comfyui":
@@ -194,7 +215,7 @@ async def verify_url(user=Depends(get_admin_user)):
             r = requests.get(url=f"{app.state.config.COMFYUI_BASE_URL}/object_info")
             r.raise_for_status()
             return True
-        except Exception as e:
+        except Exception:
             app.state.config.ENABLED = False
             raise HTTPException(status_code=400, detail=ERROR_MESSAGES.INVALID_URL)
     else:
@@ -202,6 +223,7 @@ async def verify_url(user=Depends(get_admin_user)):
 
 
 def set_image_model(model: str):
+    log.info(f"Setting image model to {model}")
     app.state.config.MODEL = model
     if app.state.config.ENGINE in ["", "automatic1111"]:
         api_auth = get_automatic1111_api_auth()
@@ -255,7 +277,8 @@ async def get_image_config(user=Depends(get_admin_user)):
 
 @app.post("/image/config/update")
 async def update_image_config(form_data: ImageConfigForm, user=Depends(get_admin_user)):
-    app.state.config.MODEL = form_data.MODEL
+
+    set_image_model(form_data.MODEL)
 
     pattern = r"^\d+x\d+$"
     if re.match(pattern, form_data.IMAGE_SIZE):
@@ -397,7 +420,6 @@ def save_url_image(url):
         r = requests.get(url)
         r.raise_for_status()
         if r.headers["content-type"].split("/")[0] == "image":
-
             mime_type = r.headers["content-type"]
             image_format = mimetypes.guess_extension(mime_type)
 
@@ -412,7 +434,7 @@ def save_url_image(url):
                     image_file.write(chunk)
             return image_filename
         else:
-            log.error(f"Url does not point to an image.")
+            log.error("Url does not point to an image.")
             return None
 
     except Exception as e:
@@ -430,7 +452,6 @@ async def image_generations(
     r = None
     try:
         if app.state.config.ENGINE == "openai":
-
             headers = {}
             headers["Authorization"] = f"Bearer {app.state.config.OPENAI_API_KEY}"
             headers["Content-Type"] = "application/json"
@@ -449,7 +470,9 @@ async def image_generations(
                 "response_format": "b64_json",
             }
 
-            r = requests.post(
+            # Use asyncio.to_thread for the requests.post call
+            r = await asyncio.to_thread(
+                requests.post,
                 url=f"{app.state.config.OPENAI_API_BASE_URL}/images/generations",
                 json=data,
                 headers=headers,
@@ -534,6 +557,15 @@ async def image_generations(
             if form_data.negative_prompt is not None:
                 data["negative_prompt"] = form_data.negative_prompt
 
+            if app.state.config.AUTOMATIC1111_CFG_SCALE:
+                data["cfg_scale"] = app.state.config.AUTOMATIC1111_CFG_SCALE
+
+            if app.state.config.AUTOMATIC1111_SAMPLER:
+                data["sampler_name"] = app.state.config.AUTOMATIC1111_SAMPLER
+
+            if app.state.config.AUTOMATIC1111_SCHEDULER:
+                data["scheduler"] = app.state.config.AUTOMATIC1111_SCHEDULER
+
             # Use asyncio.to_thread for the requests.post call
             r = await asyncio.to_thread(
                 requests.post,

backend/open_webui/apps/images/utils/comfyui.py

@@ -1,34 +1,45 @@
 import asyncio
-import websocket  # NOTE: websocket-client (https://github.com/websocket-client/websocket-client)
 import json
-import urllib.request
-import urllib.parse
-import random
 import logging
+import random
+import urllib.parse
+import urllib.request
+from typing import Optional
 
-from config import SRC_LOG_LEVELS
+import websocket  # NOTE: websocket-client (https://github.com/websocket-client/websocket-client)
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["COMFYUI"])
 
-from pydantic import BaseModel
-
-from typing import Optional
+default_headers = {"User-Agent": "Mozilla/5.0"}
 
 
 def queue_prompt(prompt, client_id, base_url):
     log.info("queue_prompt")
     p = {"prompt": prompt, "client_id": client_id}
     data = json.dumps(p).encode("utf-8")
-    req = urllib.request.Request(f"{base_url}/prompt", data=data)
-    return json.loads(urllib.request.urlopen(req).read())
+    log.debug(f"queue_prompt data: {data}")
+    try:
+        req = urllib.request.Request(
+            f"{base_url}/prompt", data=data, headers=default_headers
+        )
+        response = urllib.request.urlopen(req).read()
+        return json.loads(response)
+    except Exception as e:
+        log.exception(f"Error while queuing prompt: {e}")
+        raise e
 
 
 def get_image(filename, subfolder, folder_type, base_url):
     log.info("get_image")
     data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
     url_values = urllib.parse.urlencode(data)
-    with urllib.request.urlopen(f"{base_url}/view?{url_values}") as response:
+    req = urllib.request.Request(
+        f"{base_url}/view?{url_values}", headers=default_headers
+    )
+    with urllib.request.urlopen(req) as response:
         return response.read()
 
 
@@ -41,7 +52,11 @@ def get_image_url(filename, subfolder, folder_type, base_url):
 
 def get_history(prompt_id, base_url):
     log.info("get_history")
-    with urllib.request.urlopen(f"{base_url}/history/{prompt_id}") as response:
+
+    req = urllib.request.Request(
+        f"{base_url}/history/{prompt_id}", headers=default_headers
+    )
+    with urllib.request.urlopen(req) as response:
         return json.loads(response.read())
 
 

backend/open_webui/apps/ollama/main.py

@@ -1,54 +1,46 @@
-from fastapi import (
-    FastAPI,
-    Request,
-    HTTPException,
-    Depends,
-    UploadFile,
-    File,
-)
-from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import StreamingResponse
-
-from pydantic import BaseModel, ConfigDict
-
-import os
-import re
-import random
-import requests
-import json
-import aiohttp
 import asyncio
+import json
 import logging
+import os
+import random
+import re
 import time
-from urllib.parse import urlparse
 from typing import Optional, Union
+from urllib.parse import urlparse
 
-from starlette.background import BackgroundTask
-
-from apps.webui.models.models import Models
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    get_verified_user,
-    get_admin_user,
-)
-
-from config import (
-    SRC_LOG_LEVELS,
-    OLLAMA_BASE_URLS,
-    ENABLE_OLLAMA_API,
-    AIOHTTP_CLIENT_TIMEOUT,
+import aiohttp
+import requests
+from open_webui.apps.webui.models.models import Models
+from open_webui.config import (
+    CORS_ALLOW_ORIGIN,
     ENABLE_MODEL_FILTER,
+    ENABLE_OLLAMA_API,
     MODEL_FILTER_LIST,
+    OLLAMA_BASE_URLS,
     UPLOAD_DIR,
     AppConfig,
-    CORS_ALLOW_ORIGIN,
 )
-from utils.misc import (
+from open_webui.env import AIOHTTP_CLIENT_TIMEOUT
+
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+from fastapi import Depends, FastAPI, File, HTTPException, Request, UploadFile
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel, ConfigDict
+from starlette.background import BackgroundTask
+
+
+from open_webui.utils.misc import (
     calculate_sha256,
+)
+from open_webui.utils.payload import (
     apply_model_params_to_body_ollama,
     apply_model_params_to_body_openai,
     apply_model_system_prompt_to_body,
 )
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["OLLAMA"])
@@ -127,7 +119,7 @@ async def update_ollama_api_url(form_data: UrlUpdateForm, user=Depends(get_admin
 
 
 async def fetch_url(url):
-    timeout = aiohttp.ClientTimeout(total=5)
+    timeout = aiohttp.ClientTimeout(total=3)
     try:
         async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
             async with session.get(url) as response:
@@ -553,6 +545,65 @@ class GenerateEmbeddingsForm(BaseModel):
     keep_alive: Optional[Union[int, str]] = None
 
 
+class GenerateEmbedForm(BaseModel):
+    model: str
+    input: str
+    truncate: Optional[bool]
+    options: Optional[dict] = None
+    keep_alive: Optional[Union[int, str]] = None
+
+
+@app.post("/api/embed")
+@app.post("/api/embed/{url_idx}")
+async def generate_embeddings(
+    form_data: GenerateEmbedForm,
+    url_idx: Optional[int] = None,
+    user=Depends(get_verified_user),
+):
+    if url_idx is None:
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
+        else:
+            raise HTTPException(
+                status_code=400,
+                detail=ERROR_MESSAGES.MODEL_NOT_FOUND(form_data.model),
+            )
+
+    url = app.state.config.OLLAMA_BASE_URLS[url_idx]
+    log.info(f"url: {url}")
+
+    r = requests.request(
+        method="POST",
+        url=f"{url}/api/embed",
+        headers={"Content-Type": "application/json"},
+        data=form_data.model_dump_json(exclude_none=True).encode(),
+    )
+    try:
+        r.raise_for_status()
+
+        return r.json()
+    except Exception as e:
+        log.exception(e)
+        error_detail = "Open WebUI: Server Connection Error"
+        if r is not None:
+            try:
+                res = r.json()
+                if "error" in res:
+                    error_detail = f"Ollama: {res['error']}"
+            except Exception:
+                error_detail = f"Ollama: {e}"
+
+        raise HTTPException(
+            status_code=r.status_code if r else 500,
+            detail=error_detail,
+        )
+
+
 @app.post("/api/embeddings")
 @app.post("/api/embeddings/{url_idx}")
 async def generate_embeddings(
@@ -738,6 +789,7 @@ async def generate_chat_completion(
 ):
     payload = {**form_data.model_dump(exclude_none=True)}
     log.debug(f"{payload = }")
+
     if "metadata" in payload:
         del payload["metadata"]
 
@@ -775,7 +827,10 @@ async def generate_chat_completion(
     log.debug(payload)
 
     return await post_streaming_url(
-        f"{url}/api/chat", json.dumps(payload), content_type="application/x-ndjson"
+        f"{url}/api/chat",
+        json.dumps(payload),
+        stream=form_data.stream,
+        content_type="application/x-ndjson",
     )
 
 

backend/open_webui/apps/openai/main.py

@@ -1,44 +1,39 @@
-from fastapi import FastAPI, Request, HTTPException, Depends
-from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import StreamingResponse, FileResponse
-
-import requests
-import aiohttp
 import asyncio
+import hashlib
 import json
 import logging
+from pathlib import Path
+from typing import Literal, Optional, overload
 
+import aiohttp
+import requests
+from open_webui.apps.webui.models.models import Models
+from open_webui.config import (
+    CACHE_DIR,
+    CORS_ALLOW_ORIGIN,
+    ENABLE_MODEL_FILTER,
+    ENABLE_OPENAI_API,
+    MODEL_FILTER_LIST,
+    OPENAI_API_BASE_URLS,
+    OPENAI_API_KEYS,
+    AppConfig,
+)
+from open_webui.env import AIOHTTP_CLIENT_TIMEOUT
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+from fastapi import Depends, FastAPI, HTTPException, Request
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import FileResponse, StreamingResponse
 from pydantic import BaseModel
 from starlette.background import BackgroundTask
 
-from apps.webui.models.models import Models
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    get_verified_user,
-    get_admin_user,
-)
-from utils.misc import (
+from open_webui.utils.payload import (
     apply_model_params_to_body_openai,
     apply_model_system_prompt_to_body,
 )
 
-from config import (
-    SRC_LOG_LEVELS,
-    ENABLE_OPENAI_API,
-    AIOHTTP_CLIENT_TIMEOUT,
-    OPENAI_API_BASE_URLS,
-    OPENAI_API_KEYS,
-    CACHE_DIR,
-    ENABLE_MODEL_FILTER,
-    MODEL_FILTER_LIST,
-    AppConfig,
-    CORS_ALLOW_ORIGIN,
-)
-from typing import Optional, Literal, overload
-
-
-import hashlib
-from pathlib import Path
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["OPENAI"])
@@ -52,7 +47,6 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-
 app.state.config = AppConfig()
 
 app.state.config.ENABLE_MODEL_FILTER = ENABLE_MODEL_FILTER
@@ -185,7 +179,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
 
 
 async def fetch_url(url, key):
-    timeout = aiohttp.ClientTimeout(total=5)
+    timeout = aiohttp.ClientTimeout(total=3)
     try:
         headers = {"Authorization": f"Bearer {key}"}
         async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
@@ -225,7 +219,17 @@ def merge_models_lists(model_lists):
                     for model in models
                     if "api.openai.com"
                     not in app.state.config.OPENAI_API_BASE_URLS[idx]
-                    or "gpt" in model["id"]
+                    or not any(
+                        name in model["id"]
+                        for name in [
+                            "babbage",
+                            "dall-e",
+                            "davinci",
+                            "embedding",
+                            "tts",
+                            "whisper",
+                        ]
+                    )
                 ]
             )
 
@@ -328,10 +332,24 @@ async def get_models(url_idx: Optional[int] = None, user=Depends(get_verified_us
             r.raise_for_status()
 
             response_data = r.json()
+
             if "api.openai.com" in url:
-                response_data["data"] = list(
-                    filter(lambda model: "gpt" in model["id"], response_data["data"])
-                )
+                # Filter the response data
+                response_data["data"] = [
+                    model
+                    for model in response_data["data"]
+                    if not any(
+                        name in model["id"]
+                        for name in [
+                            "babbage",
+                            "dall-e",
+                            "davinci",
+                            "embedding",
+                            "tts",
+                            "whisper",
+                        ]
+                    )
+                ]
 
             return response_data
         except Exception as e:
@@ -386,14 +404,32 @@ async def generate_chat_completion(
             "role": user.role,
         }
 
+    url = app.state.config.OPENAI_API_BASE_URLS[idx]
+    key = app.state.config.OPENAI_API_KEYS[idx]
+    is_o1 = payload["model"].lower().startswith("o1-")
+
+    # Change max_completion_tokens to max_tokens (Backward compatible)
+    if "api.openai.com" not in url and not is_o1:
+        if "max_completion_tokens" in payload:
+            # Remove "max_completion_tokens" from the payload
+            payload["max_tokens"] = payload["max_completion_tokens"]
+            del payload["max_completion_tokens"]
+    else:
+        if is_o1 and "max_tokens" in payload:
+            payload["max_completion_tokens"] = payload["max_tokens"]
+            del payload["max_tokens"]
+        if "max_tokens" in payload and "max_completion_tokens" in payload:
+            del payload["max_tokens"]
+
+    # Fix: O1 does not support the "system" parameter, Modify "system" to "user"
+    if is_o1 and payload["messages"][0]["role"] == "system":
+        payload["messages"][0]["role"] = "user"
+
     # Convert the modified body back to JSON
     payload = json.dumps(payload)
 
     log.debug(payload)
 
-    url = app.state.config.OPENAI_API_BASE_URLS[idx]
-    key = app.state.config.OPENAI_API_KEYS[idx]
-
     headers = {}
     headers["Authorization"] = f"Bearer {key}"
     headers["Content-Type"] = "application/json"
@@ -404,6 +440,7 @@ async def generate_chat_completion(
     r = None
     session = None
     streaming = False
+    response = None
 
     try:
         session = aiohttp.ClientSession(
@@ -416,8 +453,6 @@ async def generate_chat_completion(
             headers=headers,
         )
 
-        r.raise_for_status()
-
         # Check if response is SSE
         if "text/event-stream" in r.headers.get("Content-Type", ""):
             streaming = True
@@ -430,19 +465,23 @@ async def generate_chat_completion(
                 ),
             )
         else:
-            response_data = await r.json()
-            return response_data
+            try:
+                response = await r.json()
+            except Exception as e:
+                log.error(e)
+                response = await r.text()
+
+            r.raise_for_status()
+            return response
     except Exception as e:
         log.exception(e)
         error_detail = "Open WebUI: Server Connection Error"
-        if r is not None:
-            try:
-                res = await r.json()
-                print(res)
-                if "error" in res:
-                    error_detail = f"External: {res['error']['message'] if 'message' in res['error'] else res['error']}"
-            except Exception:
-                error_detail = f"External: {e}"
+        if isinstance(response, dict):
+            if "error" in response:
+                error_detail = f"{response['error']['message'] if 'message' in response['error'] else response['error']}"
+        elif isinstance(response, str):
+            error_detail = response
+
         raise HTTPException(status_code=r.status if r else 500, detail=error_detail)
     finally:
         if not streaming and session:

backend/open_webui/apps/retrieval/loaders/main.py

@@ -0,0 +1,190 @@
+import requests
+import logging
+import ftfy
+
+from langchain_community.document_loaders import (
+    BSHTMLLoader,
+    CSVLoader,
+    Docx2txtLoader,
+    OutlookMessageLoader,
+    PyPDFLoader,
+    TextLoader,
+    UnstructuredEPubLoader,
+    UnstructuredExcelLoader,
+    UnstructuredMarkdownLoader,
+    UnstructuredPowerPointLoader,
+    UnstructuredRSTLoader,
+    UnstructuredXMLLoader,
+    YoutubeLoader,
+)
+from langchain_core.documents import Document
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+known_source_ext = [
+    "go",
+    "py",
+    "java",
+    "sh",
+    "bat",
+    "ps1",
+    "cmd",
+    "js",
+    "ts",
+    "css",
+    "cpp",
+    "hpp",
+    "h",
+    "c",
+    "cs",
+    "sql",
+    "log",
+    "ini",
+    "pl",
+    "pm",
+    "r",
+    "dart",
+    "dockerfile",
+    "env",
+    "php",
+    "hs",
+    "hsc",
+    "lua",
+    "nginxconf",
+    "conf",
+    "m",
+    "mm",
+    "plsql",
+    "perl",
+    "rb",
+    "rs",
+    "db2",
+    "scala",
+    "bash",
+    "swift",
+    "vue",
+    "svelte",
+    "msg",
+    "ex",
+    "exs",
+    "erl",
+    "tsx",
+    "jsx",
+    "hs",
+    "lhs",
+]
+
+
+class TikaLoader:
+    def __init__(self, url, file_path, mime_type=None):
+        self.url = url
+        self.file_path = file_path
+        self.mime_type = mime_type
+
+    def load(self) -> list[Document]:
+        with open(self.file_path, "rb") as f:
+            data = f.read()
+
+        if self.mime_type is not None:
+            headers = {"Content-Type": self.mime_type}
+        else:
+            headers = {}
+
+        endpoint = self.url
+        if not endpoint.endswith("/"):
+            endpoint += "/"
+        endpoint += "tika/text"
+
+        r = requests.put(endpoint, data=data, headers=headers)
+
+        if r.ok:
+            raw_metadata = r.json()
+            text = raw_metadata.get("X-TIKA:content", "<No text content found>")
+
+            if "Content-Type" in raw_metadata:
+                headers["Content-Type"] = raw_metadata["Content-Type"]
+
+            log.info("Tika extracted text: %s", text)
+
+            return [Document(page_content=text, metadata=headers)]
+        else:
+            raise Exception(f"Error calling Tika: {r.reason}")
+
+
+class Loader:
+    def __init__(self, engine: str = "", **kwargs):
+        self.engine = engine
+        self.kwargs = kwargs
+
+    def load(
+        self, filename: str, file_content_type: str, file_path: str
+    ) -> list[Document]:
+        loader = self._get_loader(filename, file_content_type, file_path)
+        docs = loader.load()
+
+        return [
+            Document(
+                page_content=ftfy.fix_text(doc.page_content), metadata=doc.metadata
+            )
+            for doc in docs
+        ]
+
+    def _get_loader(self, filename: str, file_content_type: str, file_path: str):
+        file_ext = filename.split(".")[-1].lower()
+
+        if self.engine == "tika" and self.kwargs.get("TIKA_SERVER_URL"):
+            if file_ext in known_source_ext or (
+                file_content_type and file_content_type.find("text/") >= 0
+            ):
+                loader = TextLoader(file_path, autodetect_encoding=True)
+            else:
+                loader = TikaLoader(
+                    url=self.kwargs.get("TIKA_SERVER_URL"),
+                    file_path=file_path,
+                    mime_type=file_content_type,
+                )
+        else:
+            if file_ext == "pdf":
+                loader = PyPDFLoader(
+                    file_path, extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES")
+                )
+            elif file_ext == "csv":
+                loader = CSVLoader(file_path)
+            elif file_ext == "rst":
+                loader = UnstructuredRSTLoader(file_path, mode="elements")
+            elif file_ext == "xml":
+                loader = UnstructuredXMLLoader(file_path)
+            elif file_ext in ["htm", "html"]:
+                loader = BSHTMLLoader(file_path, open_encoding="unicode_escape")
+            elif file_ext == "md":
+                loader = UnstructuredMarkdownLoader(file_path)
+            elif file_content_type == "application/epub+zip":
+                loader = UnstructuredEPubLoader(file_path)
+            elif (
+                file_content_type
+                == "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+                or file_ext == "docx"
+            ):
+                loader = Docx2txtLoader(file_path)
+            elif file_content_type in [
+                "application/vnd.ms-excel",
+                "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+            ] or file_ext in ["xls", "xlsx"]:
+                loader = UnstructuredExcelLoader(file_path)
+            elif file_content_type in [
+                "application/vnd.ms-powerpoint",
+                "application/vnd.openxmlformats-officedocument.presentationml.presentation",
+            ] or file_ext in ["ppt", "pptx"]:
+                loader = UnstructuredPowerPointLoader(file_path)
+            elif file_ext == "msg":
+                loader = OutlookMessageLoader(file_path)
+            elif file_ext in known_source_ext or (
+                file_content_type and file_content_type.find("text/") >= 0
+            ):
+                loader = TextLoader(file_path, autodetect_encoding=True)
+            else:
+                loader = TextLoader(file_path, autodetect_encoding=True)
+
+        return loader

backend/open_webui/apps/retrieval/models/colbert.py

@@ -0,0 +1,81 @@
+import os
+import torch
+import numpy as np
+from colbert.infra import ColBERTConfig
+from colbert.modeling.checkpoint import Checkpoint
+
+
+class ColBERT:
+    def __init__(self, name, **kwargs) -> None:
+        print("ColBERT: Loading model", name)
+        self.device = "cuda" if torch.cuda.is_available() else "cpu"
+
+        DOCKER = kwargs.get("env") == "docker"
+        if DOCKER:
+            # This is a workaround for the issue with the docker container
+            # where the torch extension is not loaded properly
+            # and the following error is thrown:
+            # /root/.cache/torch_extensions/py311_cpu/segmented_maxsim_cpp/segmented_maxsim_cpp.so: cannot open shared object file: No such file or directory
+
+            lock_file = (
+                "/root/.cache/torch_extensions/py311_cpu/segmented_maxsim_cpp/lock"
+            )
+            if os.path.exists(lock_file):
+                os.remove(lock_file)
+
+        self.ckpt = Checkpoint(
+            name,
+            colbert_config=ColBERTConfig(model_name=name),
+        ).to(self.device)
+        pass
+
+    def calculate_similarity_scores(self, query_embeddings, document_embeddings):
+
+        query_embeddings = query_embeddings.to(self.device)
+        document_embeddings = document_embeddings.to(self.device)
+
+        # Validate dimensions to ensure compatibility
+        if query_embeddings.dim() != 3:
+            raise ValueError(
+                f"Expected query embeddings to have 3 dimensions, but got {query_embeddings.dim()}."
+            )
+        if document_embeddings.dim() != 3:
+            raise ValueError(
+                f"Expected document embeddings to have 3 dimensions, but got {document_embeddings.dim()}."
+            )
+        if query_embeddings.size(0) not in [1, document_embeddings.size(0)]:
+            raise ValueError(
+                "There should be either one query or queries equal to the number of documents."
+            )
+
+        # Transpose the query embeddings to align for matrix multiplication
+        transposed_query_embeddings = query_embeddings.permute(0, 2, 1)
+        # Compute similarity scores using batch matrix multiplication
+        computed_scores = torch.matmul(document_embeddings, transposed_query_embeddings)
+        # Apply max pooling to extract the highest semantic similarity across each document's sequence
+        maximum_scores = torch.max(computed_scores, dim=1).values
+
+        # Sum up the maximum scores across features to get the overall document relevance scores
+        final_scores = maximum_scores.sum(dim=1)
+
+        normalized_scores = torch.softmax(final_scores, dim=0)
+
+        return normalized_scores.detach().cpu().numpy().astype(np.float32)
+
+    def predict(self, sentences):
+
+        query = sentences[0][0]
+        docs = [i[1] for i in sentences]
+
+        # Embedding the documents
+        embedded_docs = self.ckpt.docFromText(docs, bsize=32)[0]
+        # Embedding the queries
+        embedded_queries = self.ckpt.queryFromText([query], bsize=32)
+        embedded_query = embedded_queries[0]
+
+        # Calculate retrieval scores for the query against all documents
+        scores = self.calculate_similarity_scores(
+            embedded_query.unsqueeze(0), embedded_docs
+        )
+
+        return scores

backend/open_webui/apps/retrieval/utils.py

@@ -1,50 +1,84 @@
-import os
 import logging
+import os
+import uuid
+from typing import Optional, Union
+
 import requests
 
-from typing import Union
-
-from apps.ollama.main import (
-    generate_ollama_embeddings,
-    GenerateEmbeddingsForm,
-)
-
 from huggingface_hub import snapshot_download
-
-from langchain_core.documents import Document
+from langchain.retrievers import ContextualCompressionRetriever, EnsembleRetriever
 from langchain_community.retrievers import BM25Retriever
-from langchain.retrievers import (
-    ContextualCompressionRetriever,
-    EnsembleRetriever,
+from langchain_core.documents import Document
+
+
+from open_webui.apps.ollama.main import (
+    GenerateEmbeddingsForm,
+    generate_ollama_embeddings,
 )
+from open_webui.apps.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.utils.misc import get_last_user_message
 
-from typing import Optional
+from open_webui.env import SRC_LOG_LEVELS
 
-from utils.misc import get_last_user_message, add_or_update_system_message
-from config import SRC_LOG_LEVELS, CHROMA_CLIENT
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
+from typing import Any
+
+from langchain_core.callbacks import CallbackManagerForRetrieverRun
+from langchain_core.retrievers import BaseRetriever
+
+
+class VectorSearchRetriever(BaseRetriever):
+    collection_name: Any
+    embedding_function: Any
+    top_k: int
+
+    def _get_relevant_documents(
+        self,
+        query: str,
+        *,
+        run_manager: CallbackManagerForRetrieverRun,
+    ) -> list[Document]:
+        result = VECTOR_DB_CLIENT.search(
+            collection_name=self.collection_name,
+            vectors=[self.embedding_function(query)],
+            limit=self.top_k,
+        )
+
+        ids = result.ids[0]
+        metadatas = result.metadatas[0]
+        documents = result.documents[0]
+
+        results = []
+        for idx in range(len(ids)):
+            results.append(
+                Document(
+                    metadata=metadatas[idx],
+                    page_content=documents[idx],
+                )
+            )
+        return results
+
+
 def query_doc(
     collection_name: str,
-    query: str,
-    embedding_function,
+    query_embedding: list[float],
     k: int,
 ):
     try:
-        collection = CHROMA_CLIENT.get_collection(name=collection_name)
-        query_embeddings = embedding_function(query)
-
-        result = collection.query(
-            query_embeddings=[query_embeddings],
-            n_results=k,
+        result = VECTOR_DB_CLIENT.search(
+            collection_name=collection_name,
+            vectors=[query_embedding],
+            limit=k,
         )
 
         log.info(f"query_doc:result {result}")
         return result
     except Exception as e:
+        print(e)
         raise e
 
 
@@ -55,27 +89,25 @@ def query_doc_with_hybrid_search(
     k: int,
     reranking_function,
     r: float,
-):
+) -> dict:
     try:
-        collection = CHROMA_CLIENT.get_collection(name=collection_name)
-        documents = collection.get()  # get all documents
+        result = VECTOR_DB_CLIENT.get(collection_name=collection_name)
 
         bm25_retriever = BM25Retriever.from_texts(
-            texts=documents.get("documents"),
-            metadatas=documents.get("metadatas"),
+            texts=result.documents[0],
+            metadatas=result.metadatas[0],
         )
         bm25_retriever.k = k
 
-        chroma_retriever = ChromaRetriever(
-            collection=collection,
+        vector_search_retriever = VectorSearchRetriever(
+            collection_name=collection_name,
             embedding_function=embedding_function,
-            top_n=k,
+            top_k=k,
         )
 
         ensemble_retriever = EnsembleRetriever(
-            retrievers=[bm25_retriever, chroma_retriever], weights=[0.5, 0.5]
+            retrievers=[bm25_retriever, vector_search_retriever], weights=[0.5, 0.5]
         )
-
         compressor = RerankCompressor(
             embedding_function=embedding_function,
             top_n=k,
@@ -100,7 +132,9 @@ def query_doc_with_hybrid_search(
         raise e
 
 
-def merge_and_sort_query_results(query_results, k, reverse=False):
+def merge_and_sort_query_results(
+    query_results: list[dict], k: int, reverse: bool = False
+) -> list[dict]:
     # Initialize lists to store combined data
     combined_distances = []
     combined_documents = []
@@ -146,20 +180,22 @@ def query_collection(
     query: str,
     embedding_function,
     k: int,
-):
+) -> dict:
+
     results = []
+    query_embedding = embedding_function(query)
+
     for collection_name in collection_names:
         if collection_name:
             try:
                 result = query_doc(
                     collection_name=collection_name,
-                    query=query,
                     k=k,
-                    embedding_function=embedding_function,
+                    query_embedding=query_embedding,
                 )
-                results.append(result)
-            except Exception:
-                pass
+                results.append(result.model_dump())
+            except Exception as e:
+                log.exception(f"Error when querying the collection: {e}")
         else:
             pass
 
@@ -173,8 +209,9 @@ def query_collection_with_hybrid_search(
     k: int,
     reranking_function,
     r: float,
-):
+) -> dict:
     results = []
+    error = False
     for collection_name in collection_names:
         try:
             result = query_doc_with_hybrid_search(
@@ -186,14 +223,39 @@ def query_collection_with_hybrid_search(
                 r=r,
             )
             results.append(result)
-        except Exception:
-            pass
+        except Exception as e:
+            log.exception(
+                "Error when querying the collection with " f"hybrid_search: {e}"
+            )
+            error = True
+
+    if error:
+        raise Exception(
+            "Hybrid search failed for all collections. Using Non hybrid search as fallback."
+        )
+
     return merge_and_sort_query_results(results, k=k, reverse=True)
 
 
 def rag_template(template: str, context: str, query: str):
-    template = template.replace("[context]", context)
-    template = template.replace("[query]", query)
+    count = template.count("[context]")
+    assert "[context]" in template, "RAG template does not contain '[context]'"
+
+    if "<context>" in context and "</context>" in context:
+        log.debug(
+            "WARNING: Potential prompt injection attack: the RAG "
+            "context contains '<context>' and '</context>'. This might be "
+            "nothing, or the user might be trying to hack something."
+        )
+
+    if "[query]" in context:
+        query_placeholder = f"[query-{str(uuid.uuid4())}]"
+        template = template.replace("[query]", query_placeholder)
+        template = template.replace("[context]", context)
+        template = template.replace(query_placeholder, query)
+    else:
+        template = template.replace("[context]", context)
+        template = template.replace("[query]", query)
     return template
 
 
@@ -256,51 +318,71 @@ def get_rag_context(
     relevant_contexts = []
 
     for file in files:
-        context = None
-
-        collection_names = (
-            file["collection_names"]
-            if file["type"] == "collection"
-            else [file["collection_name"]] if file["collection_name"] else []
-        )
-
-        collection_names = set(collection_names).difference(extracted_collections)
-        if not collection_names:
-            log.debug(f"skipping {file} as it has already been extracted")
-            continue
-
-        try:
-            if file["type"] == "text":
-                context = file["content"]
-            else:
-                if hybrid_search:
-                    context = query_collection_with_hybrid_search(
-                        collection_names=collection_names,
-                        query=query,
-                        embedding_function=embedding_function,
-                        k=k,
-                        reranking_function=reranking_function,
-                        r=r,
-                    )
-                else:
-                    context = query_collection(
-                        collection_names=collection_names,
-                        query=query,
-                        embedding_function=embedding_function,
-                        k=k,
-                    )
-        except Exception as e:
-            log.exception(e)
+        if file.get("context") == "full":
+            context = {
+                "documents": [[file.get("file").get("data", {}).get("content")]],
+                "metadatas": [[{"file_id": file.get("id"), "name": file.get("name")}]],
+            }
+        else:
             context = None
 
-        if context:
-            relevant_contexts.append({**context, "source": file})
+            collection_names = []
+            if file.get("type") == "collection":
+                if file.get("legacy"):
+                    collection_names = file.get("collection_names", [])
+                else:
+                    collection_names.append(file["id"])
+            elif file.get("collection_name"):
+                collection_names.append(file["collection_name"])
+            elif file.get("id"):
+                if file.get("legacy"):
+                    collection_names.append(f"{file['id']}")
+                else:
+                    collection_names.append(f"file-{file['id']}")
 
-        extracted_collections.extend(collection_names)
+            collection_names = set(collection_names).difference(extracted_collections)
+            if not collection_names:
+                log.debug(f"skipping {file} as it has already been extracted")
+                continue
+
+            try:
+                context = None
+                if file.get("type") == "text":
+                    context = file["content"]
+                else:
+                    if hybrid_search:
+                        try:
+                            context = query_collection_with_hybrid_search(
+                                collection_names=collection_names,
+                                query=query,
+                                embedding_function=embedding_function,
+                                k=k,
+                                reranking_function=reranking_function,
+                                r=r,
+                            )
+                        except Exception as e:
+                            log.debug(
+                                "Error when using hybrid search, using"
+                                " non hybrid search as fallback."
+                            )
+
+                    if (not hybrid_search) or (context is None):
+                        context = query_collection(
+                            collection_names=collection_names,
+                            query=query,
+                            embedding_function=embedding_function,
+                            k=k,
+                        )
+            except Exception as e:
+                log.exception(e)
+
+            extracted_collections.extend(collection_names)
+
+        if context:
+            relevant_contexts.append({**context, "file": file})
 
     contexts = []
     citations = []
-
     for context in relevant_contexts:
         try:
             if "documents" in context:
@@ -313,7 +395,7 @@ def get_rag_context(
                 if "metadatas" in context:
                     citations.append(
                         {
-                            "source": context["source"],
+                            "source": context["file"],
                             "document": context["documents"][0],
                             "metadata": context["metadatas"][0],
                         }
@@ -399,52 +481,11 @@ def generate_openai_batch_embeddings(
         return None
 
 
-from typing import Any
-
-from langchain_core.retrievers import BaseRetriever
-from langchain_core.callbacks import CallbackManagerForRetrieverRun
-
-
-class ChromaRetriever(BaseRetriever):
-    collection: Any
-    embedding_function: Any
-    top_n: int
-
-    def _get_relevant_documents(
-        self,
-        query: str,
-        *,
-        run_manager: CallbackManagerForRetrieverRun,
-    ) -> list[Document]:
-        query_embeddings = self.embedding_function(query)
-
-        results = self.collection.query(
-            query_embeddings=[query_embeddings],
-            n_results=self.top_n,
-        )
-
-        ids = results["ids"][0]
-        metadatas = results["metadatas"][0]
-        documents = results["documents"][0]
-
-        results = []
-        for idx in range(len(ids)):
-            results.append(
-                Document(
-                    metadata=metadatas[idx],
-                    page_content=documents[idx],
-                )
-            )
-        return results
-
-
 import operator
-
 from typing import Optional, Sequence
 
-from langchain_core.documents import BaseDocumentCompressor, Document
 from langchain_core.callbacks import Callbacks
-from langchain_core.pydantic_v1 import Extra
+from langchain_core.documents import BaseDocumentCompressor, Document
 
 
 class RerankCompressor(BaseDocumentCompressor):
@@ -454,7 +495,7 @@ class RerankCompressor(BaseDocumentCompressor):
     r_score: float
 
     class Config:
-        extra = Extra.forbid
+        extra = "forbid"
         arbitrary_types_allowed = True
 
     def compress_documents(

backend/open_webui/apps/retrieval/vector/connector.py

@@ -0,0 +1,10 @@
+from open_webui.config import VECTOR_DB
+
+if VECTOR_DB == "milvus":
+    from open_webui.apps.retrieval.vector.dbs.milvus import MilvusClient
+
+    VECTOR_DB_CLIENT = MilvusClient()
+else:
+    from open_webui.apps.retrieval.vector.dbs.chroma import ChromaClient
+
+    VECTOR_DB_CLIENT = ChromaClient()

backend/open_webui/apps/retrieval/vector/dbs/chroma.py

@@ -0,0 +1,157 @@
+import chromadb
+from chromadb import Settings
+from chromadb.utils.batch_utils import create_batches
+
+from typing import Optional
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import (
+    CHROMA_DATA_PATH,
+    CHROMA_HTTP_HOST,
+    CHROMA_HTTP_PORT,
+    CHROMA_HTTP_HEADERS,
+    CHROMA_HTTP_SSL,
+    CHROMA_TENANT,
+    CHROMA_DATABASE,
+)
+
+
+class ChromaClient:
+    def __init__(self):
+        if CHROMA_HTTP_HOST != "":
+            self.client = chromadb.HttpClient(
+                host=CHROMA_HTTP_HOST,
+                port=CHROMA_HTTP_PORT,
+                headers=CHROMA_HTTP_HEADERS,
+                ssl=CHROMA_HTTP_SSL,
+                tenant=CHROMA_TENANT,
+                database=CHROMA_DATABASE,
+                settings=Settings(allow_reset=True, anonymized_telemetry=False),
+            )
+        else:
+            self.client = chromadb.PersistentClient(
+                path=CHROMA_DATA_PATH,
+                settings=Settings(allow_reset=True, anonymized_telemetry=False),
+                tenant=CHROMA_TENANT,
+                database=CHROMA_DATABASE,
+            )
+
+    def has_collection(self, collection_name: str) -> bool:
+        # Check if the collection exists based on the collection name.
+        collections = self.client.list_collections()
+        return collection_name in [collection.name for collection in collections]
+
+    def delete_collection(self, collection_name: str):
+        # Delete the collection based on the collection name.
+        return self.client.delete_collection(name=collection_name)
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
+        try:
+            collection = self.client.get_collection(name=collection_name)
+            if collection:
+                result = collection.query(
+                    query_embeddings=vectors,
+                    n_results=limit,
+                )
+
+                return SearchResult(
+                    **{
+                        "ids": result["ids"],
+                        "distances": result["distances"],
+                        "documents": result["documents"],
+                        "metadatas": result["metadatas"],
+                    }
+                )
+            return None
+        except Exception as e:
+            return None
+
+    def query(
+        self, collection_name: str, filter: dict, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        # Query the items from the collection based on the filter.
+        try:
+            collection = self.client.get_collection(name=collection_name)
+            if collection:
+                result = collection.get(
+                    where=filter,
+                    limit=limit,
+                )
+
+                return GetResult(
+                    **{
+                        "ids": [result["ids"]],
+                        "documents": [result["documents"]],
+                        "metadatas": [result["metadatas"]],
+                    }
+                )
+            return None
+        except Exception as e:
+            print(e)
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        # Get all the items in the collection.
+        collection = self.client.get_collection(name=collection_name)
+        if collection:
+            result = collection.get()
+            return GetResult(
+                **{
+                    "ids": [result["ids"]],
+                    "documents": [result["documents"]],
+                    "metadatas": [result["metadatas"]],
+                }
+            )
+        return None
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        # Insert the items into the collection, if the collection does not exist, it will be created.
+        collection = self.client.get_or_create_collection(name=collection_name)
+
+        ids = [item["id"] for item in items]
+        documents = [item["text"] for item in items]
+        embeddings = [item["vector"] for item in items]
+        metadatas = [item["metadata"] for item in items]
+
+        for batch in create_batches(
+            api=self.client,
+            documents=documents,
+            embeddings=embeddings,
+            ids=ids,
+            metadatas=metadatas,
+        ):
+            collection.add(*batch)
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
+        collection = self.client.get_or_create_collection(name=collection_name)
+
+        ids = [item["id"] for item in items]
+        documents = [item["text"] for item in items]
+        embeddings = [item["vector"] for item in items]
+        metadatas = [item["metadata"] for item in items]
+
+        collection.upsert(
+            ids=ids, documents=documents, embeddings=embeddings, metadatas=metadatas
+        )
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        # Delete the items from the collection based on the ids.
+        collection = self.client.get_collection(name=collection_name)
+        if collection:
+            if ids:
+                collection.delete(ids=ids)
+            elif filter:
+                collection.delete(where=filter)
+
+    def reset(self):
+        # Resets the database. This will delete all collections and item entries.
+        return self.client.reset()

backend/open_webui/apps/retrieval/vector/dbs/milvus.py

@@ -0,0 +1,286 @@
+from pymilvus import MilvusClient as Client
+from pymilvus import FieldSchema, DataType
+import json
+
+from typing import Optional
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import (
+    MILVUS_URI,
+)
+
+
+class MilvusClient:
+    def __init__(self):
+        self.collection_prefix = "open_webui"
+        self.client = Client(uri=MILVUS_URI)
+
+    def _result_to_get_result(self, result) -> GetResult:
+        ids = []
+        documents = []
+        metadatas = []
+
+        for match in result:
+            _ids = []
+            _documents = []
+            _metadatas = []
+            for item in match:
+                _ids.append(item.get("id"))
+                _documents.append(item.get("data", {}).get("text"))
+                _metadatas.append(item.get("metadata"))
+
+            ids.append(_ids)
+            documents.append(_documents)
+            metadatas.append(_metadatas)
+
+        return GetResult(
+            **{
+                "ids": ids,
+                "documents": documents,
+                "metadatas": metadatas,
+            }
+        )
+
+    def _result_to_search_result(self, result) -> SearchResult:
+        ids = []
+        distances = []
+        documents = []
+        metadatas = []
+
+        for match in result:
+            _ids = []
+            _distances = []
+            _documents = []
+            _metadatas = []
+
+            for item in match:
+                _ids.append(item.get("id"))
+                _distances.append(item.get("distance"))
+                _documents.append(item.get("entity", {}).get("data", {}).get("text"))
+                _metadatas.append(item.get("entity", {}).get("metadata"))
+
+            ids.append(_ids)
+            distances.append(_distances)
+            documents.append(_documents)
+            metadatas.append(_metadatas)
+
+        return SearchResult(
+            **{
+                "ids": ids,
+                "distances": distances,
+                "documents": documents,
+                "metadatas": metadatas,
+            }
+        )
+
+    def _create_collection(self, collection_name: str, dimension: int):
+        schema = self.client.create_schema(
+            auto_id=False,
+            enable_dynamic_field=True,
+        )
+        schema.add_field(
+            field_name="id",
+            datatype=DataType.VARCHAR,
+            is_primary=True,
+            max_length=65535,
+        )
+        schema.add_field(
+            field_name="vector",
+            datatype=DataType.FLOAT_VECTOR,
+            dim=dimension,
+            description="vector",
+        )
+        schema.add_field(field_name="data", datatype=DataType.JSON, description="data")
+        schema.add_field(
+            field_name="metadata", datatype=DataType.JSON, description="metadata"
+        )
+
+        index_params = self.client.prepare_index_params()
+        index_params.add_index(
+            field_name="vector",
+            index_type="HNSW",
+            metric_type="COSINE",
+            params={"M": 16, "efConstruction": 100},
+        )
+
+        self.client.create_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            schema=schema,
+            index_params=index_params,
+        )
+
+    def has_collection(self, collection_name: str) -> bool:
+        # Check if the collection exists based on the collection name.
+        collection_name = collection_name.replace("-", "_")
+        return self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def delete_collection(self, collection_name: str):
+        # Delete the collection based on the collection name.
+        collection_name = collection_name.replace("-", "_")
+        return self.client.drop_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
+        collection_name = collection_name.replace("-", "_")
+        result = self.client.search(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=vectors,
+            limit=limit,
+            output_fields=["data", "metadata"],
+        )
+
+        return self._result_to_search_result(result)
+
+    def query(self, collection_name: str, filter: dict, limit: Optional[int] = None):
+        # Construct the filter string for querying
+        collection_name = collection_name.replace("-", "_")
+        if not self.has_collection(collection_name):
+            return None
+
+        filter_string = " && ".join(
+            [
+                f'metadata["{key}"] == {json.dumps(value)}'
+                for key, value in filter.items()
+            ]
+        )
+
+        max_limit = 16383  # The maximum number of records per request
+        all_results = []
+
+        if limit is None:
+            limit = float("inf")  # Use infinity as a placeholder for no limit
+
+        # Initialize offset and remaining to handle pagination
+        offset = 0
+        remaining = limit
+
+        try:
+            # Loop until there are no more items to fetch or the desired limit is reached
+            while remaining > 0:
+                print("remaining", remaining)
+                current_fetch = min(
+                    max_limit, remaining
+                )  # Determine how many items to fetch in this iteration
+
+                results = self.client.query(
+                    collection_name=f"{self.collection_prefix}_{collection_name}",
+                    filter=filter_string,
+                    output_fields=["*"],
+                    limit=current_fetch,
+                    offset=offset,
+                )
+
+                if not results:
+                    break
+
+                all_results.extend(results)
+                results_count = len(results)
+                remaining -= (
+                    results_count  # Decrease remaining by the number of items fetched
+                )
+                offset += results_count
+
+                # Break the loop if the results returned are less than the requested fetch count
+                if results_count < current_fetch:
+                    break
+
+            print(all_results)
+            return self._result_to_get_result([all_results])
+        except Exception as e:
+            print(e)
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        # Get all the items in the collection.
+        collection_name = collection_name.replace("-", "_")
+        result = self.client.query(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            filter='id != ""',
+        )
+        return self._result_to_get_result([result])
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        # Insert the items into the collection, if the collection does not exist, it will be created.
+        collection_name = collection_name.replace("-", "_")
+        if not self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        ):
+            self._create_collection(
+                collection_name=collection_name, dimension=len(items[0]["vector"])
+            )
+
+        return self.client.insert(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=[
+                {
+                    "id": item["id"],
+                    "vector": item["vector"],
+                    "data": {"text": item["text"]},
+                    "metadata": item["metadata"],
+                }
+                for item in items
+            ],
+        )
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
+        collection_name = collection_name.replace("-", "_")
+        if not self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        ):
+            self._create_collection(
+                collection_name=collection_name, dimension=len(items[0]["vector"])
+            )
+
+        return self.client.upsert(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=[
+                {
+                    "id": item["id"],
+                    "vector": item["vector"],
+                    "data": {"text": item["text"]},
+                    "metadata": item["metadata"],
+                }
+                for item in items
+            ],
+        )
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        # Delete the items from the collection based on the ids.
+        collection_name = collection_name.replace("-", "_")
+        if ids:
+            return self.client.delete(
+                collection_name=f"{self.collection_prefix}_{collection_name}",
+                ids=ids,
+            )
+        elif filter:
+            # Convert the filter dictionary to a string using JSON_CONTAINS.
+            filter_string = " && ".join(
+                [
+                    f'metadata["{key}"] == {json.dumps(value)}'
+                    for key, value in filter.items()
+                ]
+            )
+
+            return self.client.delete(
+                collection_name=f"{self.collection_prefix}_{collection_name}",
+                filter=filter_string,
+            )
+
+    def reset(self):
+        # Resets the database. This will delete all collections and item entries.
+        collection_names = self.client.list_collections()
+        for collection_name in collection_names:
+            if collection_name.startswith(self.collection_prefix):
+                self.client.drop_collection(collection_name=collection_name)

backend/open_webui/apps/retrieval/vector/main.py

@@ -0,0 +1,19 @@
+from pydantic import BaseModel
+from typing import Optional, List, Any
+
+
+class VectorItem(BaseModel):
+    id: str
+    text: str
+    vector: List[float | int]
+    metadata: Any
+
+
+class GetResult(BaseModel):
+    ids: Optional[List[List[str]]]
+    documents: Optional[List[List[str]]]
+    metadatas: Optional[List[List[Any]]]
+
+
+class SearchResult(GetResult):
+    distances: Optional[List[List[float | int]]]

backend/open_webui/apps/retrieval/web/brave.py

@@ -1,9 +1,9 @@
 import logging
 from typing import Optional
-import requests
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/duckduckgo.py

@@ -1,8 +1,9 @@
 import logging
 from typing import Optional
-from apps.rag.search.main import SearchResult, get_filtered_results
+
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
 from duckduckgo_search import DDGS
-from config import SRC_LOG_LEVELS
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/google_pse.py

@@ -1,10 +1,9 @@
-import json
 import logging
 from typing import Optional
-import requests
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/jina_search.py

@@ -1,9 +1,9 @@
 import logging
-import requests
-from yarl import URL
 
-from apps.rag.search.main import SearchResult
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult
+from open_webui.env import SRC_LOG_LEVELS
+from yarl import URL
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/main.py

@@ -1,5 +1,6 @@
 from typing import Optional
 from urllib.parse import urlparse
+
 from pydantic import BaseModel
 
 
@@ -8,7 +9,8 @@ def get_filtered_results(results, filter_list):
         return results
     filtered_results = []
     for result in results:
-        domain = urlparse(result["url"]).netloc
+        url = result.get("url") or result.get("link", "")
+        domain = urlparse(url).netloc
         if any(domain.endswith(filtered_domain) for filtered_domain in filter_list):
             filtered_results.append(result)
     return filtered_results

backend/open_webui/apps/retrieval/web/searchapi.py

@@ -0,0 +1,48 @@
+import logging
+from typing import Optional
+from urllib.parse import urlencode
+
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_searchapi(
+    api_key: str,
+    engine: str,
+    query: str,
+    count: int,
+    filter_list: Optional[list[str]] = None,
+) -> list[SearchResult]:
+    """Search using searchapi.io's API and return the results as a list of SearchResult objects.
+
+    Args:
+      api_key (str): A searchapi.io API key
+      query (str): The query to search for
+    """
+    url = "https://www.searchapi.io/api/v1/search"
+
+    engine = engine or "google"
+
+    payload = {"engine": engine, "q": query, "api_key": api_key}
+
+    url = f"{url}?{urlencode(payload)}"
+    response = requests.request("GET", url)
+
+    json_response = response.json()
+    log.info(f"results from searchapi search: {json_response}")
+
+    results = sorted(
+        json_response.get("organic_results", []), key=lambda x: x.get("position", 0)
+    )
+    if filter_list:
+        results = get_filtered_results(results, filter_list)
+    return [
+        SearchResult(
+            link=result["link"], title=result["title"], snippet=result["snippet"]
+        )
+        for result in results[:count]
+    ]

backend/open_webui/apps/retrieval/web/searxng.py

@@ -1,10 +1,9 @@
 import logging
-import requests
-
 from typing import Optional
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/serper.py

@@ -1,10 +1,10 @@
 import json
 import logging
 from typing import Optional
-import requests
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/serply.py

@@ -1,11 +1,10 @@
-import json
 import logging
 from typing import Optional
-import requests
 from urllib.parse import urlencode
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/serpstack.py

@@ -1,10 +1,9 @@
-import json
 import logging
 from typing import Optional
-import requests
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/tavily.py

@@ -1,9 +1,8 @@
 import logging
 
 import requests
-
-from apps.rag.search.main import SearchResult
-from config import SRC_LOG_LEVELS
+from open_webui.apps.retrieval.web.main import SearchResult
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])

backend/open_webui/apps/retrieval/web/utils.py

@@ -0,0 +1,97 @@
+import socket
+import urllib.parse
+import validators
+from typing import Union, Sequence, Iterator
+
+from langchain_community.document_loaders import (
+    WebBaseLoader,
+)
+from langchain_core.documents import Document
+
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.config import ENABLE_RAG_LOCAL_WEB_FETCH
+from open_webui.env import SRC_LOG_LEVELS
+
+import logging
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def validate_url(url: Union[str, Sequence[str]]):
+    if isinstance(url, str):
+        if isinstance(validators.url(url), validators.ValidationError):
+            raise ValueError(ERROR_MESSAGES.INVALID_URL)
+        if not ENABLE_RAG_LOCAL_WEB_FETCH:
+            # Local web fetch is disabled, filter out any URLs that resolve to private IP addresses
+            parsed_url = urllib.parse.urlparse(url)
+            # Get IPv4 and IPv6 addresses
+            ipv4_addresses, ipv6_addresses = resolve_hostname(parsed_url.hostname)
+            # Check if any of the resolved addresses are private
+            # This is technically still vulnerable to DNS rebinding attacks, as we don't control WebBaseLoader
+            for ip in ipv4_addresses:
+                if validators.ipv4(ip, private=True):
+                    raise ValueError(ERROR_MESSAGES.INVALID_URL)
+            for ip in ipv6_addresses:
+                if validators.ipv6(ip, private=True):
+                    raise ValueError(ERROR_MESSAGES.INVALID_URL)
+        return True
+    elif isinstance(url, Sequence):
+        return all(validate_url(u) for u in url)
+    else:
+        return False
+
+
+def resolve_hostname(hostname):
+    # Get address information
+    addr_info = socket.getaddrinfo(hostname, None)
+
+    # Extract IP addresses from address information
+    ipv4_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET]
+    ipv6_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET6]
+
+    return ipv4_addresses, ipv6_addresses
+
+
+class SafeWebBaseLoader(WebBaseLoader):
+    """WebBaseLoader with enhanced error handling for URLs."""
+
+    def lazy_load(self) -> Iterator[Document]:
+        """Lazy load text from the url(s) in web_path with error handling."""
+        for path in self.web_paths:
+            try:
+                soup = self._scrape(path, bs_kwargs=self.bs_kwargs)
+                text = soup.get_text(**self.bs_get_text_kwargs)
+
+                # Build metadata
+                metadata = {"source": path}
+                if title := soup.find("title"):
+                    metadata["title"] = title.get_text()
+                if description := soup.find("meta", attrs={"name": "description"}):
+                    metadata["description"] = description.get(
+                        "content", "No description found."
+                    )
+                if html := soup.find("html"):
+                    metadata["language"] = html.get("lang", "No language found.")
+
+                yield Document(page_content=text, metadata=metadata)
+            except Exception as e:
+                # Log the error and continue with the next URL
+                log.error(f"Error loading {path}: {e}")
+
+
+def get_web_loader(
+    url: Union[str, Sequence[str]],
+    verify_ssl: bool = True,
+    requests_per_second: int = 2,
+):
+    # Check if the URL is valid
+    if not validate_url(url):
+        raise ValueError(ERROR_MESSAGES.INVALID_URL)
+    return SafeWebBaseLoader(
+        url,
+        verify_ssl=verify_ssl,
+        requests_per_second=requests_per_second,
+        continue_on_failure=True,
+    )

backend/open_webui/apps/socket/main.py

@@ -0,0 +1,219 @@
+import asyncio
+import socketio
+import logging
+import sys
+import time
+
+from open_webui.apps.webui.models.users import Users
+from open_webui.env import (
+    ENABLE_WEBSOCKET_SUPPORT,
+    WEBSOCKET_MANAGER,
+    WEBSOCKET_REDIS_URL,
+)
+from open_webui.utils.utils import decode_token
+from open_webui.apps.socket.utils import RedisDict
+
+from open_webui.env import (
+    GLOBAL_LOG_LEVEL,
+    SRC_LOG_LEVELS,
+)
+
+
+logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["SOCKET"])
+
+
+if WEBSOCKET_MANAGER == "redis":
+    mgr = socketio.AsyncRedisManager(WEBSOCKET_REDIS_URL)
+    sio = socketio.AsyncServer(
+        cors_allowed_origins=[],
+        async_mode="asgi",
+        transports=(
+            ["polling", "websocket"] if ENABLE_WEBSOCKET_SUPPORT else ["polling"]
+        ),
+        allow_upgrades=ENABLE_WEBSOCKET_SUPPORT,
+        always_connect=True,
+        client_manager=mgr,
+    )
+else:
+    sio = socketio.AsyncServer(
+        cors_allowed_origins=[],
+        async_mode="asgi",
+        transports=(
+            ["polling", "websocket"] if ENABLE_WEBSOCKET_SUPPORT else ["polling"]
+        ),
+        allow_upgrades=ENABLE_WEBSOCKET_SUPPORT,
+        always_connect=True,
+    )
+
+
+# Dictionary to maintain the user pool
+
+if WEBSOCKET_MANAGER == "redis":
+    SESSION_POOL = RedisDict("open-webui:session_pool", redis_url=WEBSOCKET_REDIS_URL)
+    USER_POOL = RedisDict("open-webui:user_pool", redis_url=WEBSOCKET_REDIS_URL)
+    USAGE_POOL = RedisDict("open-webui:usage_pool", redis_url=WEBSOCKET_REDIS_URL)
+else:
+    SESSION_POOL = {}
+    USER_POOL = {}
+    USAGE_POOL = {}
+
+
+# Timeout duration in seconds
+TIMEOUT_DURATION = 3
+
+
+async def periodic_usage_pool_cleanup():
+    while True:
+        now = int(time.time())
+        for model_id, connections in list(USAGE_POOL.items()):
+            # Creating a list of sids to remove if they have timed out
+            expired_sids = [
+                sid
+                for sid, details in connections.items()
+                if now - details["updated_at"] > TIMEOUT_DURATION
+            ]
+
+            for sid in expired_sids:
+                del connections[sid]
+
+            if not connections:
+                log.debug(f"Cleaning up model {model_id} from usage pool")
+                del USAGE_POOL[model_id]
+            else:
+                USAGE_POOL[model_id] = connections
+
+            # Emit updated usage information after cleaning
+            await sio.emit("usage", {"models": get_models_in_use()})
+
+        await asyncio.sleep(TIMEOUT_DURATION)
+
+
+app = socketio.ASGIApp(
+    sio,
+    socketio_path="/ws/socket.io",
+)
+
+
+def get_models_in_use():
+    # List models that are currently in use
+    models_in_use = list(USAGE_POOL.keys())
+    return models_in_use
+
+
+@sio.on("usage")
+async def usage(sid, data):
+    model_id = data["model"]
+    # Record the timestamp for the last update
+    current_time = int(time.time())
+
+    # Store the new usage data and task
+    USAGE_POOL[model_id] = {
+        **(USAGE_POOL[model_id] if model_id in USAGE_POOL else {}),
+        sid: {"updated_at": current_time},
+    }
+
+    # Broadcast the usage data to all clients
+    await sio.emit("usage", {"models": get_models_in_use()})
+
+
+@sio.event
+async def connect(sid, environ, auth):
+    user = None
+    if auth and "token" in auth:
+        data = decode_token(auth["token"])
+
+        if data is not None and "id" in data:
+            user = Users.get_user_by_id(data["id"])
+
+        if user:
+            SESSION_POOL[sid] = user.id
+            if user.id in USER_POOL:
+                USER_POOL[user.id].append(sid)
+            else:
+                USER_POOL[user.id] = [sid]
+
+            # print(f"user {user.name}({user.id}) connected with session ID {sid}")
+            await sio.emit("user-count", {"count": len(USER_POOL.items())})
+            await sio.emit("usage", {"models": get_models_in_use()})
+
+
+@sio.on("user-join")
+async def user_join(sid, data):
+    # print("user-join", sid, data)
+
+    auth = data["auth"] if "auth" in data else None
+    if not auth or "token" not in auth:
+        return
+
+    data = decode_token(auth["token"])
+    if data is None or "id" not in data:
+        return
+
+    user = Users.get_user_by_id(data["id"])
+    if not user:
+        return
+
+    SESSION_POOL[sid] = user.id
+    if user.id in USER_POOL:
+        USER_POOL[user.id].append(sid)
+    else:
+        USER_POOL[user.id] = [sid]
+
+    # print(f"user {user.name}({user.id}) connected with session ID {sid}")
+
+    await sio.emit("user-count", {"count": len(USER_POOL.items())})
+
+
+@sio.on("user-count")
+async def user_count(sid):
+    await sio.emit("user-count", {"count": len(USER_POOL.items())})
+
+
+@sio.event
+async def disconnect(sid):
+    if sid in SESSION_POOL:
+        user_id = SESSION_POOL[sid]
+        del SESSION_POOL[sid]
+
+        USER_POOL[user_id] = [_sid for _sid in USER_POOL[user_id] if _sid != sid]
+
+        if len(USER_POOL[user_id]) == 0:
+            del USER_POOL[user_id]
+
+        await sio.emit("user-count", {"count": len(USER_POOL)})
+    else:
+        pass
+        # print(f"Unknown session ID {sid} disconnected")
+
+
+def get_event_emitter(request_info):
+    async def __event_emitter__(event_data):
+        await sio.emit(
+            "chat-events",
+            {
+                "chat_id": request_info["chat_id"],
+                "message_id": request_info["message_id"],
+                "data": event_data,
+            },
+            to=request_info["session_id"],
+        )
+
+    return __event_emitter__
+
+
+def get_event_call(request_info):
+    async def __event_call__(event_data):
+        response = await sio.call(
+            "chat-events",
+            {
+                "chat_id": request_info["chat_id"],
+                "message_id": request_info["message_id"],
+                "data": event_data,
+            },
+            to=request_info["session_id"],
+        )
+        return response
+
+    return __event_call__

backend/open_webui/apps/socket/utils.py

@@ -0,0 +1,59 @@
+import json
+import redis
+
+
+class RedisDict:
+    def __init__(self, name, redis_url):
+        self.name = name
+        self.redis = redis.Redis.from_url(redis_url, decode_responses=True)
+
+    def __setitem__(self, key, value):
+        serialized_value = json.dumps(value)
+        self.redis.hset(self.name, key, serialized_value)
+
+    def __getitem__(self, key):
+        value = self.redis.hget(self.name, key)
+        if value is None:
+            raise KeyError(key)
+        return json.loads(value)
+
+    def __delitem__(self, key):
+        result = self.redis.hdel(self.name, key)
+        if result == 0:
+            raise KeyError(key)
+
+    def __contains__(self, key):
+        return self.redis.hexists(self.name, key)
+
+    def __len__(self):
+        return self.redis.hlen(self.name)
+
+    def keys(self):
+        return self.redis.hkeys(self.name)
+
+    def values(self):
+        return [json.loads(v) for v in self.redis.hvals(self.name)]
+
+    def items(self):
+        return [(k, json.loads(v)) for k, v in self.redis.hgetall(self.name).items()]
+
+    def get(self, key, default=None):
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def clear(self):
+        self.redis.delete(self.name)
+
+    def update(self, other=None, **kwargs):
+        if other is not None:
+            for k, v in other.items() if hasattr(other, "items") else other:
+                self[k] = v
+        for k, v in kwargs.items():
+            self[k] = v
+
+    def setdefault(self, key, default=None):
+        if key not in self:
+            self[key] = default
+        return self[key]

backend/open_webui/apps/webui/internal/db.py

@@ -1,21 +1,25 @@
-import os
-import logging
 import json
+import logging
 from contextlib import contextmanager
+from typing import Any, Optional
 
-
-from typing import Optional, Any
-from typing_extensions import Self
-
-from sqlalchemy import create_engine, types, Dialect
-from sqlalchemy.sql.type_api import _T
-from sqlalchemy.ext.declarative import declarative_base
-from sqlalchemy.orm import sessionmaker, scoped_session
-
-
+from open_webui.apps.webui.internal.wrappers import register_connection
+from open_webui.env import (
+    OPEN_WEBUI_DIR,
+    DATABASE_URL,
+    SRC_LOG_LEVELS,
+    DATABASE_POOL_MAX_OVERFLOW,
+    DATABASE_POOL_RECYCLE,
+    DATABASE_POOL_SIZE,
+    DATABASE_POOL_TIMEOUT,
+)
 from peewee_migrate import Router
-from apps.webui.internal.wrappers import register_connection
-from env import SRC_LOG_LEVELS, BACKEND_DIR, DATABASE_URL
+from sqlalchemy import Dialect, create_engine, types
+from sqlalchemy.ext.declarative import declarative_base
+from sqlalchemy.orm import scoped_session, sessionmaker
+from sqlalchemy.pool import QueuePool, NullPool
+from sqlalchemy.sql.type_api import _T
+from typing_extensions import Self
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["DB"])
@@ -50,7 +54,7 @@ def handle_peewee_migration(DATABASE_URL):
     try:
         # Replace the postgresql:// with postgres:// to handle the peewee migration
         db = register_connection(DATABASE_URL.replace("postgresql://", "postgres://"))
-        migrate_dir = BACKEND_DIR / "apps" / "webui" / "internal" / "migrations"
+        migrate_dir = OPEN_WEBUI_DIR / "apps" / "webui" / "internal" / "migrations"
         router = Router(db, logger=log, migrate_dir=migrate_dir)
         router.run()
         db.close()
@@ -76,7 +80,20 @@ if "sqlite" in SQLALCHEMY_DATABASE_URL:
         SQLALCHEMY_DATABASE_URL, connect_args={"check_same_thread": False}
     )
 else:
-    engine = create_engine(SQLALCHEMY_DATABASE_URL, pool_pre_ping=True)
+    if DATABASE_POOL_SIZE > 0:
+        engine = create_engine(
+            SQLALCHEMY_DATABASE_URL,
+            pool_size=DATABASE_POOL_SIZE,
+            max_overflow=DATABASE_POOL_MAX_OVERFLOW,
+            pool_timeout=DATABASE_POOL_TIMEOUT,
+            pool_recycle=DATABASE_POOL_RECYCLE,
+            pool_pre_ping=True,
+            poolclass=QueuePool,
+        )
+    else:
+        engine = create_engine(
+            SQLALCHEMY_DATABASE_URL, pool_pre_ping=True, poolclass=NullPool
+        )
 
 
 SessionLocal = sessionmaker(

backend/open_webui/apps/webui/internal/migrations/010_migrate_modelfiles_to_models.py

@@ -30,7 +30,7 @@ import peewee as pw
 from peewee_migrate import Migrator
 import json
 
-from utils.misc import parse_ollama_modelfile
+from open_webui.utils.misc import parse_ollama_modelfile
 
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext

backend/open_webui/apps/webui/internal/wrappers.py

@@ -1,13 +1,13 @@
-from contextvars import ContextVar
-from peewee import *
-from peewee import PostgresqlDatabase, InterfaceError as PeeWeeInterfaceError
-
 import logging
+from contextvars import ContextVar
+
+from open_webui.env import SRC_LOG_LEVELS
+from peewee import *
+from peewee import InterfaceError as PeeWeeInterfaceError
+from peewee import PostgresqlDatabase
 from playhouse.db_url import connect, parse
 from playhouse.shortcuts import ReconnectMixin
 
-from env import SRC_LOG_LEVELS
-
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["DB"])
 

backend/open_webui/apps/webui/main.py

@@ -1,65 +1,66 @@
-from fastapi import FastAPI
-from fastapi.responses import StreamingResponse
-from fastapi.middleware.cors import CORSMiddleware
-from apps.webui.routers import (
+import inspect
+import json
+import logging
+from typing import AsyncGenerator, Generator, Iterator
+
+from open_webui.apps.socket.main import get_event_call, get_event_emitter
+from open_webui.apps.webui.models.functions import Functions
+from open_webui.apps.webui.models.models import Models
+from open_webui.apps.webui.routers import (
     auths,
-    users,
     chats,
-    documents,
-    tools,
-    models,
-    prompts,
     configs,
-    memories,
-    utils,
     files,
     functions,
+    memories,
+    models,
+    knowledge,
+    prompts,
+    tools,
+    users,
+    utils,
 )
-from apps.webui.models.functions import Functions
-from apps.webui.models.models import Models
-from apps.webui.utils import load_function_module_by_id
-
-from utils.misc import (
+from open_webui.apps.webui.utils import load_function_module_by_id
+from open_webui.config import (
+    ADMIN_EMAIL,
+    CORS_ALLOW_ORIGIN,
+    DEFAULT_MODELS,
+    DEFAULT_PROMPT_SUGGESTIONS,
+    DEFAULT_USER_ROLE,
+    ENABLE_COMMUNITY_SHARING,
+    ENABLE_LOGIN_FORM,
+    ENABLE_MESSAGE_RATING,
+    ENABLE_SIGNUP,
+    JWT_EXPIRES_IN,
+    OAUTH_EMAIL_CLAIM,
+    OAUTH_PICTURE_CLAIM,
+    OAUTH_USERNAME_CLAIM,
+    SHOW_ADMIN_DETAILS,
+    USER_PERMISSIONS,
+    WEBHOOK_URL,
+    WEBUI_AUTH,
+    WEBUI_BANNERS,
+    AppConfig,
+)
+from open_webui.env import (
+    WEBUI_AUTH_TRUSTED_EMAIL_HEADER,
+    WEBUI_AUTH_TRUSTED_NAME_HEADER,
+)
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel
+from open_webui.utils.misc import (
     openai_chat_chunk_message_template,
     openai_chat_completion_message_template,
+)
+from open_webui.utils.payload import (
     apply_model_params_to_body_openai,
     apply_model_system_prompt_to_body,
 )
 
-from utils.tools import get_tools
 
-from config import (
-    SHOW_ADMIN_DETAILS,
-    ADMIN_EMAIL,
-    WEBUI_AUTH,
-    DEFAULT_MODELS,
-    DEFAULT_PROMPT_SUGGESTIONS,
-    DEFAULT_USER_ROLE,
-    ENABLE_SIGNUP,
-    ENABLE_LOGIN_FORM,
-    USER_PERMISSIONS,
-    WEBHOOK_URL,
-    WEBUI_AUTH_TRUSTED_EMAIL_HEADER,
-    WEBUI_AUTH_TRUSTED_NAME_HEADER,
-    JWT_EXPIRES_IN,
-    WEBUI_BANNERS,
-    ENABLE_COMMUNITY_SHARING,
-    ENABLE_MESSAGE_RATING,
-    AppConfig,
-    OAUTH_USERNAME_CLAIM,
-    OAUTH_PICTURE_CLAIM,
-    OAUTH_EMAIL_CLAIM,
-    CORS_ALLOW_ORIGIN,
-)
-
-from apps.socket.main import get_event_call, get_event_emitter
-
-import inspect
-import json
-import logging
-
-from typing import Iterator, Generator, AsyncGenerator
-from pydantic import BaseModel
+from open_webui.utils.tools import get_tools
 
 app = FastAPI()
 
@@ -110,15 +111,15 @@ app.include_router(auths.router, prefix="/auths", tags=["auths"])
 app.include_router(users.router, prefix="/users", tags=["users"])
 app.include_router(chats.router, prefix="/chats", tags=["chats"])
 
-app.include_router(documents.router, prefix="/documents", tags=["documents"])
 app.include_router(models.router, prefix="/models", tags=["models"])
+app.include_router(knowledge.router, prefix="/knowledge", tags=["knowledge"])
 app.include_router(prompts.router, prefix="/prompts", tags=["prompts"])
 
-app.include_router(memories.router, prefix="/memories", tags=["memories"])
 app.include_router(files.router, prefix="/files", tags=["files"])
 app.include_router(tools.router, prefix="/tools", tags=["tools"])
 app.include_router(functions.router, prefix="/functions", tags=["functions"])
 
+app.include_router(memories.router, prefix="/memories", tags=["memories"])
 app.include_router(utils.router, prefix="/utils", tags=["utils"])
 
 
@@ -155,29 +156,33 @@ async def get_pipe_models():
 
         # Check if function is a manifold
         if hasattr(function_module, "pipes"):
-            manifold_pipes = []
+            sub_pipes = []
 
             # Check if pipes is a function or a list
-            if callable(function_module.pipes):
-                manifold_pipes = function_module.pipes()
-            else:
-                manifold_pipes = function_module.pipes
 
-            for p in manifold_pipes:
-                manifold_pipe_id = f'{pipe.id}.{p["id"]}'
-                manifold_pipe_name = p["name"]
+            try:
+                if callable(function_module.pipes):
+                    sub_pipes = function_module.pipes()
+                else:
+                    sub_pipes = function_module.pipes
+            except Exception as e:
+                log.exception(e)
+                sub_pipes = []
+
+            print(sub_pipes)
+
+            for p in sub_pipes:
+                sub_pipe_id = f'{pipe.id}.{p["id"]}'
+                sub_pipe_name = p["name"]
 
                 if hasattr(function_module, "name"):
-                    manifold_pipe_name = f"{function_module.name}{manifold_pipe_name}"
+                    sub_pipe_name = f"{function_module.name}{sub_pipe_name}"
 
                 pipe_flag = {"type": pipe.type}
-                if hasattr(function_module, "ChatValves"):
-                    pipe_flag["valves_spec"] = function_module.ChatValves.schema()
-
                 pipe_models.append(
                     {
-                        "id": manifold_pipe_id,
-                        "name": manifold_pipe_name,
+                        "id": sub_pipe_id,
+                        "name": sub_pipe_name,
                         "object": "model",
                         "created": pipe.created_at,
                         "owned_by": "openai",
@@ -186,8 +191,6 @@ async def get_pipe_models():
                 )
         else:
             pipe_flag = {"type": "pipe"}
-            if hasattr(function_module, "ChatValves"):
-                pipe_flag["valves_spec"] = function_module.ChatValves.schema()
 
             pipe_models.append(
                 {
@@ -284,17 +287,20 @@ async def generate_function_chat_completion(form_data, user):
     __event_emitter__ = None
     __event_call__ = None
     __task__ = None
+    __task_body__ = None
 
     if metadata:
         if all(k in metadata for k in ("session_id", "chat_id", "message_id")):
             __event_emitter__ = get_event_emitter(metadata)
             __event_call__ = get_event_call(metadata)
         __task__ = metadata.get("task", None)
+        __task_body__ = metadata.get("task_body", None)
 
     extra_params = {
         "__event_emitter__": __event_emitter__,
         "__event_call__": __event_call__,
         "__task__": __task__,
+        "__task_body__": __task_body__,
         "__files__": files,
         "__user__": {
             "id": user.id,

backend/open_webui/apps/webui/models/auths.py

@@ -1,15 +1,13 @@
-from pydantic import BaseModel
-from typing import Optional
-import uuid
 import logging
-from sqlalchemy import String, Column, Boolean, Text
+import uuid
+from typing import Optional
 
-from utils.utils import verify_password
-
-from apps.webui.models.users import UserModel, Users
-from apps.webui.internal.db import Base, get_db
-
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, get_db
+from open_webui.apps.webui.models.users import UserModel, Users
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel
+from sqlalchemy import Boolean, Column, String, Text
+from open_webui.utils.utils import verify_password
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -92,7 +90,6 @@ class AddUserForm(SignupForm):
 
 
 class AuthsTable:
-
     def insert_new_auth(
         self,
         email: str,
@@ -103,7 +100,6 @@ class AuthsTable:
         oauth_sub: Optional[str] = None,
     ) -> Optional[UserModel]:
         with get_db() as db:
-
             log.info("insert_new_auth")
 
             id = str(uuid.uuid4())
@@ -130,7 +126,6 @@ class AuthsTable:
         log.info(f"authenticate_user: {email}")
         try:
             with get_db() as db:
-
                 auth = db.query(Auth).filter_by(email=email, active=True).first()
                 if auth:
                     if verify_password(password, auth.password):
@@ -189,7 +184,6 @@ class AuthsTable:
     def delete_auth_by_id(self, id: str) -> bool:
         try:
             with get_db() as db:
-
                 # Delete User
                 result = Users.delete_user_by_id(id)
 

backend/open_webui/apps/webui/models/chats.py

@@ -1,14 +1,11 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Union, Optional
-
 import json
-import uuid
 import time
+import uuid
+from typing import Optional
 
-from sqlalchemy import Column, String, BigInteger, Boolean, Text
-
-from apps.webui.internal.db import Base, get_db
-
+from open_webui.apps.webui.internal.db import Base, get_db
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Boolean, Column, String, Text
 
 ####################
 # Chat DB Schema
@@ -77,10 +74,8 @@ class ChatTitleIdResponse(BaseModel):
 
 
 class ChatTable:
-
     def insert_new_chat(self, user_id: str, form_data: ChatForm) -> Optional[ChatModel]:
         with get_db() as db:
-
             id = str(uuid.uuid4())
             chat = ChatModel(
                 **{
@@ -106,7 +101,6 @@ class ChatTable:
     def update_chat_by_id(self, id: str, chat: dict) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat_obj = db.get(Chat, id)
                 chat_obj.chat = json.dumps(chat)
                 chat_obj.title = chat["title"] if "title" in chat else "New Chat"
@@ -115,12 +109,11 @@ class ChatTable:
                 db.refresh(chat_obj)
 
                 return ChatModel.model_validate(chat_obj)
-        except Exception as e:
+        except Exception:
             return None
 
     def insert_shared_chat_by_chat_id(self, chat_id: str) -> Optional[ChatModel]:
         with get_db() as db:
-
             # Get the existing chat to share
             chat = db.get(Chat, chat_id)
             # Check if the chat is already shared
@@ -154,7 +147,6 @@ class ChatTable:
     def update_shared_chat_by_chat_id(self, chat_id: str) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 print("update_shared_chat_by_id")
                 chat = db.get(Chat, chat_id)
                 print(chat)
@@ -170,7 +162,6 @@ class ChatTable:
     def delete_shared_chat_by_chat_id(self, chat_id: str) -> bool:
         try:
             with get_db() as db:
-
                 db.query(Chat).filter_by(user_id=f"shared-{chat_id}").delete()
                 db.commit()
 
@@ -183,7 +174,6 @@ class ChatTable:
     ) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat = db.get(Chat, id)
                 chat.share_id = share_id
                 db.commit()
@@ -195,7 +185,6 @@ class ChatTable:
     def toggle_chat_archive_by_id(self, id: str) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat = db.get(Chat, id)
                 chat.archived = not chat.archived
                 db.commit()
@@ -217,7 +206,6 @@ class ChatTable:
         self, user_id: str, skip: int = 0, limit: int = 50
     ) -> list[ChatModel]:
         with get_db() as db:
-
             all_chats = (
                 db.query(Chat)
                 .filter_by(user_id=user_id, archived=True)
@@ -297,7 +285,6 @@ class ChatTable:
     def get_chat_by_id(self, id: str) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat = db.get(Chat, id)
                 return ChatModel.model_validate(chat)
         except Exception:
@@ -306,20 +293,18 @@ class ChatTable:
     def get_chat_by_share_id(self, id: str) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat = db.query(Chat).filter_by(share_id=id).first()
 
                 if chat:
                     return self.get_chat_by_id(id)
                 else:
                     return None
-        except Exception as e:
+        except Exception:
             return None
 
     def get_chat_by_id_and_user_id(self, id: str, user_id: str) -> Optional[ChatModel]:
         try:
             with get_db() as db:
-
                 chat = db.query(Chat).filter_by(id=id, user_id=user_id).first()
                 return ChatModel.model_validate(chat)
         except Exception:
@@ -327,7 +312,6 @@ class ChatTable:
 
     def get_chats(self, skip: int = 0, limit: int = 50) -> list[ChatModel]:
         with get_db() as db:
-
             all_chats = (
                 db.query(Chat)
                 # .limit(limit).offset(skip)
@@ -337,7 +321,6 @@ class ChatTable:
 
     def get_chats_by_user_id(self, user_id: str) -> list[ChatModel]:
         with get_db() as db:
-
             all_chats = (
                 db.query(Chat)
                 .filter_by(user_id=user_id)
@@ -347,7 +330,6 @@ class ChatTable:
 
     def get_archived_chats_by_user_id(self, user_id: str) -> list[ChatModel]:
         with get_db() as db:
-
             all_chats = (
                 db.query(Chat)
                 .filter_by(user_id=user_id, archived=True)
@@ -358,7 +340,6 @@ class ChatTable:
     def delete_chat_by_id(self, id: str) -> bool:
         try:
             with get_db() as db:
-
                 db.query(Chat).filter_by(id=id).delete()
                 db.commit()
 
@@ -369,7 +350,6 @@ class ChatTable:
     def delete_chat_by_id_and_user_id(self, id: str, user_id: str) -> bool:
         try:
             with get_db() as db:
-
                 db.query(Chat).filter_by(id=id, user_id=user_id).delete()
                 db.commit()
 
@@ -379,9 +359,7 @@ class ChatTable:
 
     def delete_chats_by_user_id(self, user_id: str) -> bool:
         try:
-
             with get_db() as db:
-
                 self.delete_shared_chats_by_user_id(user_id)
 
                 db.query(Chat).filter_by(user_id=user_id).delete()
@@ -393,9 +371,7 @@ class ChatTable:
 
     def delete_shared_chats_by_user_id(self, user_id: str) -> bool:
         try:
-
             with get_db() as db:
-
                 chats_by_user = db.query(Chat).filter_by(user_id=user_id).all()
                 shared_chat_ids = [f"shared-{chat.id}" for chat in chats_by_user]
 

backend/open_webui/apps/webui/models/documents.py

@@ -1,15 +1,12 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Optional
-import time
-import logging
-
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
 import json
+import logging
+import time
+from typing import Optional
 
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, get_db
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -70,12 +67,10 @@ class DocumentForm(DocumentUpdateForm):
 
 
 class DocumentsTable:
-
     def insert_new_doc(
         self, user_id: str, form_data: DocumentForm
     ) -> Optional[DocumentModel]:
         with get_db() as db:
-
             document = DocumentModel(
                 **{
                     **form_data.model_dump(),
@@ -99,7 +94,6 @@ class DocumentsTable:
     def get_doc_by_name(self, name: str) -> Optional[DocumentModel]:
         try:
             with get_db() as db:
-
                 document = db.query(Document).filter_by(name=name).first()
                 return DocumentModel.model_validate(document) if document else None
         except Exception:
@@ -107,7 +101,6 @@ class DocumentsTable:
 
     def get_docs(self) -> list[DocumentModel]:
         with get_db() as db:
-
             return [
                 DocumentModel.model_validate(doc) for doc in db.query(Document).all()
             ]
@@ -117,7 +110,6 @@ class DocumentsTable:
     ) -> Optional[DocumentModel]:
         try:
             with get_db() as db:
-
                 db.query(Document).filter_by(name=name).update(
                     {
                         "title": form_data.title,
@@ -140,7 +132,6 @@ class DocumentsTable:
             doc_content = {**doc_content, **updated}
 
             with get_db() as db:
-
                 db.query(Document).filter_by(name=name).update(
                     {
                         "content": json.dumps(doc_content),
@@ -156,7 +147,6 @@ class DocumentsTable:
     def delete_doc_by_name(self, name: str) -> bool:
         try:
             with get_db() as db:
-
                 db.query(Document).filter_by(name=name).delete()
                 db.commit()
                 return True

backend/open_webui/apps/webui/models/files.py

@@ -1,15 +1,11 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Union, Optional
-import time
 import logging
+import time
+from typing import Optional
 
-from sqlalchemy import Column, String, BigInteger, Text
-
-from apps.webui.internal.db import JSONField, Base, get_db
-
-import json
-
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, JSONField, get_db
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text, JSON
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -24,19 +20,29 @@ class File(Base):
 
     id = Column(String, primary_key=True)
     user_id = Column(String)
+    hash = Column(Text, nullable=True)
+
     filename = Column(Text)
+    data = Column(JSON, nullable=True)
     meta = Column(JSONField)
+
     created_at = Column(BigInteger)
+    updated_at = Column(BigInteger)
 
 
 class FileModel(BaseModel):
+    model_config = ConfigDict(from_attributes=True)
+
     id: str
     user_id: str
-    filename: str
-    meta: dict
-    created_at: int  # timestamp in epoch
+    hash: Optional[str] = None
 
-    model_config = ConfigDict(from_attributes=True)
+    filename: str
+    data: Optional[dict] = None
+    meta: dict
+
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
 
 
 ####################
@@ -47,27 +53,33 @@ class FileModel(BaseModel):
 class FileModelResponse(BaseModel):
     id: str
     user_id: str
+    hash: Optional[str] = None
+
     filename: str
+    data: Optional[dict] = None
     meta: dict
+
     created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
 
 
 class FileForm(BaseModel):
     id: str
+    hash: Optional[str] = None
     filename: str
+    data: dict = {}
     meta: dict = {}
 
 
 class FilesTable:
-
     def insert_new_file(self, user_id: str, form_data: FileForm) -> Optional[FileModel]:
         with get_db() as db:
-
             file = FileModel(
                 **{
                     **form_data.model_dump(),
                     "user_id": user_id,
                     "created_at": int(time.time()),
+                    "updated_at": int(time.time()),
                 }
             )
 
@@ -86,7 +98,6 @@ class FilesTable:
 
     def get_file_by_id(self, id: str) -> Optional[FileModel]:
         with get_db() as db:
-
             try:
                 file = db.get(File, id)
                 return FileModel.model_validate(file)
@@ -95,9 +106,18 @@ class FilesTable:
 
     def get_files(self) -> list[FileModel]:
         with get_db() as db:
-
             return [FileModel.model_validate(file) for file in db.query(File).all()]
 
+    def get_files_by_ids(self, ids: list[str]) -> list[FileModel]:
+        with get_db() as db:
+            return [
+                FileModel.model_validate(file)
+                for file in db.query(File)
+                .filter(File.id.in_(ids))
+                .order_by(File.updated_at.desc())
+                .all()
+            ]
+
     def get_files_by_user_id(self, user_id: str) -> list[FileModel]:
         with get_db() as db:
             return [
@@ -105,10 +125,40 @@ class FilesTable:
                 for file in db.query(File).filter_by(user_id=user_id).all()
             ]
 
-    def delete_file_by_id(self, id: str) -> bool:
-
+    def update_file_hash_by_id(self, id: str, hash: str) -> Optional[FileModel]:
         with get_db() as db:
+            try:
+                file = db.query(File).filter_by(id=id).first()
+                file.hash = hash
+                db.commit()
 
+                return FileModel.model_validate(file)
+            except Exception:
+                return None
+
+    def update_file_data_by_id(self, id: str, data: dict) -> Optional[FileModel]:
+        with get_db() as db:
+            try:
+                file = db.query(File).filter_by(id=id).first()
+                file.data = {**(file.data if file.data else {}), **data}
+                db.commit()
+                return FileModel.model_validate(file)
+            except Exception as e:
+
+                return None
+
+    def update_file_metadata_by_id(self, id: str, meta: dict) -> Optional[FileModel]:
+        with get_db() as db:
+            try:
+                file = db.query(File).filter_by(id=id).first()
+                file.meta = {**(file.meta if file.meta else {}), **meta}
+                db.commit()
+                return FileModel.model_validate(file)
+            except Exception:
+                return None
+
+    def delete_file_by_id(self, id: str) -> bool:
+        with get_db() as db:
             try:
                 db.query(File).filter_by(id=id).delete()
                 db.commit()
@@ -118,9 +168,7 @@ class FilesTable:
                 return False
 
     def delete_all_files(self) -> bool:
-
         with get_db() as db:
-
             try:
                 db.query(File).delete()
                 db.commit()

backend/open_webui/apps/webui/models/functions.py

@@ -1,18 +1,12 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Union, Optional
-import time
 import logging
+import time
+from typing import Optional
 
-from sqlalchemy import Column, String, Text, BigInteger, Boolean
-
-from apps.webui.internal.db import JSONField, Base, get_db
-from apps.webui.models.users import Users
-
-import json
-import copy
-
-
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, JSONField, get_db
+from open_webui.apps.webui.models.users import Users
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Boolean, Column, String, Text
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -87,11 +81,9 @@ class FunctionValves(BaseModel):
 
 
 class FunctionsTable:
-
     def insert_new_function(
         self, user_id: str, type: str, form_data: FunctionForm
     ) -> Optional[FunctionModel]:
-
         function = FunctionModel(
             **{
                 **form_data.model_dump(),
@@ -119,7 +111,6 @@ class FunctionsTable:
     def get_function_by_id(self, id: str) -> Optional[FunctionModel]:
         try:
             with get_db() as db:
-
                 function = db.get(Function, id)
                 return FunctionModel.model_validate(function)
         except Exception:
@@ -127,7 +118,6 @@ class FunctionsTable:
 
     def get_functions(self, active_only=False) -> list[FunctionModel]:
         with get_db() as db:
-
             if active_only:
                 return [
                     FunctionModel.model_validate(function)
@@ -143,7 +133,6 @@ class FunctionsTable:
         self, type: str, active_only=False
     ) -> list[FunctionModel]:
         with get_db() as db:
-
             if active_only:
                 return [
                     FunctionModel.model_validate(function)
@@ -159,7 +148,6 @@ class FunctionsTable:
 
     def get_global_filter_functions(self) -> list[FunctionModel]:
         with get_db() as db:
-
             return [
                 FunctionModel.model_validate(function)
                 for function in db.query(Function)
@@ -178,7 +166,6 @@ class FunctionsTable:
 
     def get_function_valves_by_id(self, id: str) -> Optional[dict]:
         with get_db() as db:
-
             try:
                 function = db.get(Function, id)
                 return function.valves if function.valves else {}
@@ -190,7 +177,6 @@ class FunctionsTable:
         self, id: str, valves: dict
     ) -> Optional[FunctionValves]:
         with get_db() as db:
-
             try:
                 function = db.get(Function, id)
                 function.valves = valves
@@ -204,7 +190,6 @@ class FunctionsTable:
     def get_user_valves_by_id_and_user_id(
         self, id: str, user_id: str
     ) -> Optional[dict]:
-
         try:
             user = Users.get_user_by_id(user_id)
             user_settings = user.settings.model_dump() if user.settings else {}
@@ -223,7 +208,6 @@ class FunctionsTable:
     def update_user_valves_by_id_and_user_id(
         self, id: str, user_id: str, valves: dict
     ) -> Optional[dict]:
-
         try:
             user = Users.get_user_by_id(user_id)
             user_settings = user.settings.model_dump() if user.settings else {}
@@ -246,7 +230,6 @@ class FunctionsTable:
 
     def update_function_by_id(self, id: str, updated: dict) -> Optional[FunctionModel]:
         with get_db() as db:
-
             try:
                 db.query(Function).filter_by(id=id).update(
                     {
@@ -261,7 +244,6 @@ class FunctionsTable:
 
     def deactivate_all_functions(self) -> Optional[bool]:
         with get_db() as db:
-
             try:
                 db.query(Function).update(
                     {

backend/open_webui/apps/webui/models/knowledge.py

@@ -0,0 +1,152 @@
+import json
+import logging
+import time
+from typing import Optional
+import uuid
+
+from open_webui.apps.webui.internal.db import Base, get_db
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text, JSON
+
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["MODELS"])
+
+####################
+# Knowledge DB Schema
+####################
+
+
+class Knowledge(Base):
+    __tablename__ = "knowledge"
+
+    id = Column(Text, unique=True, primary_key=True)
+    user_id = Column(Text)
+
+    name = Column(Text)
+    description = Column(Text)
+
+    data = Column(JSON, nullable=True)
+    meta = Column(JSON, nullable=True)
+
+    created_at = Column(BigInteger)
+    updated_at = Column(BigInteger)
+
+
+class KnowledgeModel(BaseModel):
+    model_config = ConfigDict(from_attributes=True)
+
+    id: str
+    user_id: str
+
+    name: str
+    description: str
+
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+
+
+####################
+# Forms
+####################
+
+
+class KnowledgeResponse(BaseModel):
+    id: str
+    name: str
+    description: str
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+
+
+class KnowledgeForm(BaseModel):
+    name: str
+    description: str
+    data: Optional[dict] = None
+
+
+class KnowledgeUpdateForm(BaseModel):
+    name: Optional[str] = None
+    description: Optional[str] = None
+    data: Optional[dict] = None
+
+
+class KnowledgeTable:
+    def insert_new_knowledge(
+        self, user_id: str, form_data: KnowledgeForm
+    ) -> Optional[KnowledgeModel]:
+        with get_db() as db:
+            knowledge = KnowledgeModel(
+                **{
+                    **form_data.model_dump(),
+                    "id": str(uuid.uuid4()),
+                    "user_id": user_id,
+                    "created_at": int(time.time()),
+                    "updated_at": int(time.time()),
+                }
+            )
+
+            try:
+                result = Knowledge(**knowledge.model_dump())
+                db.add(result)
+                db.commit()
+                db.refresh(result)
+                if result:
+                    return KnowledgeModel.model_validate(result)
+                else:
+                    return None
+            except Exception:
+                return None
+
+    def get_knowledge_items(self) -> list[KnowledgeModel]:
+        with get_db() as db:
+            return [
+                KnowledgeModel.model_validate(knowledge)
+                for knowledge in db.query(Knowledge)
+                .order_by(Knowledge.updated_at.desc())
+                .all()
+            ]
+
+    def get_knowledge_by_id(self, id: str) -> Optional[KnowledgeModel]:
+        try:
+            with get_db() as db:
+                knowledge = db.query(Knowledge).filter_by(id=id).first()
+                return KnowledgeModel.model_validate(knowledge) if knowledge else None
+        except Exception:
+            return None
+
+    def update_knowledge_by_id(
+        self, id: str, form_data: KnowledgeUpdateForm, overwrite: bool = False
+    ) -> Optional[KnowledgeModel]:
+        try:
+            with get_db() as db:
+                knowledge = self.get_knowledge_by_id(id=id)
+                db.query(Knowledge).filter_by(id=id).update(
+                    {
+                        **form_data.model_dump(exclude_none=True),
+                        "updated_at": int(time.time()),
+                    }
+                )
+                db.commit()
+                return self.get_knowledge_by_id(id=id)
+        except Exception as e:
+            log.exception(e)
+            return None
+
+    def delete_knowledge_by_id(self, id: str) -> bool:
+        try:
+            with get_db() as db:
+                db.query(Knowledge).filter_by(id=id).delete()
+                db.commit()
+                return True
+        except Exception:
+            return False
+
+
+Knowledges = KnowledgeTable()

backend/open_webui/apps/webui/models/memories.py

@@ -1,12 +1,10 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Union, Optional
-
-from sqlalchemy import Column, String, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
 import time
 import uuid
+from typing import Optional
+
+from open_webui.apps.webui.internal.db import Base, get_db
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 ####################
 # Memory DB Schema
@@ -39,13 +37,11 @@ class MemoryModel(BaseModel):
 
 
 class MemoriesTable:
-
     def insert_new_memory(
         self,
         user_id: str,
         content: str,
     ) -> Optional[MemoryModel]:
-
         with get_db() as db:
             id = str(uuid.uuid4())
 
@@ -73,7 +69,6 @@ class MemoriesTable:
         content: str,
     ) -> Optional[MemoryModel]:
         with get_db() as db:
-
             try:
                 db.query(Memory).filter_by(id=id).update(
                     {"content": content, "updated_at": int(time.time())}
@@ -85,7 +80,6 @@ class MemoriesTable:
 
     def get_memories(self) -> list[MemoryModel]:
         with get_db() as db:
-
             try:
                 memories = db.query(Memory).all()
                 return [MemoryModel.model_validate(memory) for memory in memories]
@@ -94,7 +88,6 @@ class MemoriesTable:
 
     def get_memories_by_user_id(self, user_id: str) -> list[MemoryModel]:
         with get_db() as db:
-
             try:
                 memories = db.query(Memory).filter_by(user_id=user_id).all()
                 return [MemoryModel.model_validate(memory) for memory in memories]
@@ -103,7 +96,6 @@ class MemoriesTable:
 
     def get_memory_by_id(self, id: str) -> Optional[MemoryModel]:
         with get_db() as db:
-
             try:
                 memory = db.get(Memory, id)
                 return MemoryModel.model_validate(memory)
@@ -112,7 +104,6 @@ class MemoriesTable:
 
     def delete_memory_by_id(self, id: str) -> bool:
         with get_db() as db:
-
             try:
                 db.query(Memory).filter_by(id=id).delete()
                 db.commit()
@@ -124,7 +115,6 @@ class MemoriesTable:
 
     def delete_memories_by_user_id(self, user_id: str) -> bool:
         with get_db() as db:
-
             try:
                 db.query(Memory).filter_by(user_id=user_id).delete()
                 db.commit()
@@ -135,7 +125,6 @@ class MemoriesTable:
 
     def delete_memory_by_id_and_user_id(self, id: str, user_id: str) -> bool:
         with get_db() as db:
-
             try:
                 db.query(Memory).filter_by(id=id, user_id=user_id).delete()
                 db.commit()

backend/open_webui/apps/webui/models/models.py

@@ -1,14 +1,11 @@
 import logging
-from typing import Optional, List
-
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, JSONField, get_db
-
-from env import SRC_LOG_LEVELS
-
 import time
+from typing import Optional
+
+from open_webui.apps.webui.internal.db import Base, JSONField, get_db
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, Text
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])

backend/open_webui/apps/webui/models/prompts.py

@@ -1,12 +1,9 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Optional
 import time
+from typing import Optional
 
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
-import json
+from open_webui.apps.webui.internal.db import Base, get_db
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 ####################
 # Prompts DB Schema
@@ -45,7 +42,6 @@ class PromptForm(BaseModel):
 
 
 class PromptsTable:
-
     def insert_new_prompt(
         self, user_id: str, form_data: PromptForm
     ) -> Optional[PromptModel]:
@@ -61,7 +57,6 @@ class PromptsTable:
 
         try:
             with get_db() as db:
-
                 result = Prompt(**prompt.dict())
                 db.add(result)
                 db.commit()
@@ -70,13 +65,12 @@ class PromptsTable:
                     return PromptModel.model_validate(result)
                 else:
                     return None
-        except Exception as e:
+        except Exception:
             return None
 
     def get_prompt_by_command(self, command: str) -> Optional[PromptModel]:
         try:
             with get_db() as db:
-
                 prompt = db.query(Prompt).filter_by(command=command).first()
                 return PromptModel.model_validate(prompt)
         except Exception:
@@ -84,7 +78,6 @@ class PromptsTable:
 
     def get_prompts(self) -> list[PromptModel]:
         with get_db() as db:
-
             return [
                 PromptModel.model_validate(prompt) for prompt in db.query(Prompt).all()
             ]
@@ -94,7 +87,6 @@ class PromptsTable:
     ) -> Optional[PromptModel]:
         try:
             with get_db() as db:
-
                 prompt = db.query(Prompt).filter_by(command=command).first()
                 prompt.title = form_data.title
                 prompt.content = form_data.content
@@ -107,7 +99,6 @@ class PromptsTable:
     def delete_prompt_by_command(self, command: str) -> bool:
         try:
             with get_db() as db:
-
                 db.query(Prompt).filter_by(command=command).delete()
                 db.commit()
 

backend/open_webui/apps/webui/models/tags.py

@@ -1,16 +1,12 @@
-from pydantic import BaseModel, ConfigDict
+import logging
+import time
+import uuid
 from typing import Optional
 
-import json
-import uuid
-import time
-import logging
-
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, get_db
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -77,10 +73,8 @@ class ChatTagsResponse(BaseModel):
 
 
 class TagTable:
-
     def insert_new_tag(self, name: str, user_id: str) -> Optional[TagModel]:
         with get_db() as db:
-
             id = str(uuid.uuid4())
             tag = TagModel(**{"id": id, "user_id": user_id, "name": name})
             try:
@@ -92,7 +86,7 @@ class TagTable:
                     return TagModel.model_validate(result)
                 else:
                     return None
-            except Exception as e:
+            except Exception:
                 return None
 
     def get_tag_by_name_and_user_id(
@@ -102,7 +96,7 @@ class TagTable:
             with get_db() as db:
                 tag = db.query(Tag).filter_by(name=name, user_id=user_id).first()
                 return TagModel.model_validate(tag)
-        except Exception as e:
+        except Exception:
             return None
 
     def add_tag_to_chat(
@@ -161,7 +155,6 @@ class TagTable:
         self, chat_id: str, user_id: str
     ) -> list[TagModel]:
         with get_db() as db:
-
             tag_names = [
                 chat_id_tag.tag_name
                 for chat_id_tag in (
@@ -186,7 +179,6 @@ class TagTable:
         self, tag_name: str, user_id: str
     ) -> list[ChatIdTagModel]:
         with get_db() as db:
-
             return [
                 ChatIdTagModel.model_validate(chat_id_tag)
                 for chat_id_tag in (
@@ -201,7 +193,6 @@ class TagTable:
         self, tag_name: str, user_id: str
     ) -> int:
         with get_db() as db:
-
             return (
                 db.query(ChatIdTag)
                 .filter_by(tag_name=tag_name, user_id=user_id)
@@ -236,7 +227,6 @@ class TagTable:
     ) -> bool:
         try:
             with get_db() as db:
-
                 res = (
                     db.query(ChatIdTag)
                     .filter_by(tag_name=tag_name, chat_id=chat_id, user_id=user_id)

backend/open_webui/apps/webui/models/tools.py

@@ -1,17 +1,12 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Optional
-import time
 import logging
-from sqlalchemy import String, Column, BigInteger, Text
+import time
+from typing import Optional
 
-from apps.webui.internal.db import Base, JSONField, get_db
-from apps.webui.models.users import Users
-
-import json
-import copy
-
-
-from env import SRC_LOG_LEVELS
+from open_webui.apps.webui.internal.db import Base, JSONField, get_db
+from open_webui.apps.webui.models.users import Users
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -79,13 +74,10 @@ class ToolValves(BaseModel):
 
 
 class ToolsTable:
-
     def insert_new_tool(
         self, user_id: str, form_data: ToolForm, specs: list[dict]
     ) -> Optional[ToolModel]:
-
         with get_db() as db:
-
             tool = ToolModel(
                 **{
                     **form_data.model_dump(),
@@ -112,7 +104,6 @@ class ToolsTable:
     def get_tool_by_id(self, id: str) -> Optional[ToolModel]:
         try:
             with get_db() as db:
-
                 tool = db.get(Tool, id)
                 return ToolModel.model_validate(tool)
         except Exception:
@@ -125,7 +116,6 @@ class ToolsTable:
     def get_tool_valves_by_id(self, id: str) -> Optional[dict]:
         try:
             with get_db() as db:
-
                 tool = db.get(Tool, id)
                 return tool.valves if tool.valves else {}
         except Exception as e:
@@ -135,7 +125,6 @@ class ToolsTable:
     def update_tool_valves_by_id(self, id: str, valves: dict) -> Optional[ToolValves]:
         try:
             with get_db() as db:
-
                 db.query(Tool).filter_by(id=id).update(
                     {"valves": valves, "updated_at": int(time.time())}
                 )

backend/open_webui/apps/webui/models/users.py

@@ -1,11 +1,10 @@
-from pydantic import BaseModel, ConfigDict
-from typing import Optional
 import time
+from typing import Optional
 
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, JSONField, get_db
-from apps.webui.models.chats import Chats
+from open_webui.apps.webui.internal.db import Base, JSONField, get_db
+from open_webui.apps.webui.models.chats import Chats
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, String, Text
 
 ####################
 # User DB Schema
@@ -113,7 +112,7 @@ class UsersTable:
             with get_db() as db:
                 user = db.query(User).filter_by(id=id).first()
                 return UserModel.model_validate(user)
-        except Exception as e:
+        except Exception:
             return None
 
     def get_user_by_api_key(self, api_key: str) -> Optional[UserModel]:
@@ -221,7 +220,7 @@ class UsersTable:
                 user = db.query(User).filter_by(id=id).first()
                 return UserModel.model_validate(user)
                 # return UserModel(**user.dict())
-        except Exception as e:
+        except Exception:
             return None
 
     def delete_user_by_id(self, id: str) -> bool:
@@ -255,7 +254,7 @@ class UsersTable:
             with get_db() as db:
                 user = db.query(User).filter_by(id=id).first()
                 return user.api_key
-        except Exception as e:
+        except Exception:
             return None
 
 

backend/open_webui/apps/webui/routers/auths.py

@@ -1,43 +1,36 @@
-import logging
-
-from fastapi import Request, UploadFile, File
-from fastapi import Depends, HTTPException, status
-from fastapi.responses import Response
-
-from fastapi import APIRouter
-from pydantic import BaseModel
 import re
 import uuid
-import csv
 
-from apps.webui.models.auths import (
-    SigninForm,
-    SignupForm,
+from open_webui.apps.webui.models.auths import (
     AddUserForm,
-    UpdateProfileForm,
-    UpdatePasswordForm,
-    UserResponse,
-    SigninResponse,
-    Auths,
     ApiKey,
+    Auths,
+    SigninForm,
+    SigninResponse,
+    SignupForm,
+    UpdatePasswordForm,
+    UpdateProfileForm,
+    UserResponse,
 )
-from apps.webui.models.users import Users
-
-from utils.utils import (
-    get_password_hash,
-    get_current_user,
-    get_admin_user,
-    create_token,
-    create_api_key,
-)
-from utils.misc import parse_duration, validate_email_format
-from utils.webhook import post_webhook
-from constants import ERROR_MESSAGES, WEBHOOK_MESSAGES
-from config import (
-    WEBUI_AUTH,
+from open_webui.apps.webui.models.users import Users
+from open_webui.config import WEBUI_AUTH
+from open_webui.constants import ERROR_MESSAGES, WEBHOOK_MESSAGES
+from open_webui.env import (
     WEBUI_AUTH_TRUSTED_EMAIL_HEADER,
     WEBUI_AUTH_TRUSTED_NAME_HEADER,
 )
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from fastapi.responses import Response
+from pydantic import BaseModel
+from open_webui.utils.misc import parse_duration, validate_email_format
+from open_webui.utils.utils import (
+    create_api_key,
+    create_token,
+    get_admin_user,
+    get_current_user,
+    get_password_hash,
+)
+from open_webui.utils.webhook import post_webhook
 
 router = APIRouter()
 
@@ -195,14 +188,19 @@ async def signin(request: Request, response: Response, form_data: SigninForm):
 
 @router.post("/signup", response_model=SigninResponse)
 async def signup(request: Request, response: Response, form_data: SignupForm):
-    if (
-        not request.app.state.config.ENABLE_SIGNUP
-        and request.app.state.config.ENABLE_LOGIN_FORM
-        and WEBUI_AUTH
-    ):
-        raise HTTPException(
-            status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.ACCESS_PROHIBITED
-        )
+    if WEBUI_AUTH:
+        if (
+            not request.app.state.config.ENABLE_SIGNUP
+            or not request.app.state.config.ENABLE_LOGIN_FORM
+        ):
+            raise HTTPException(
+                status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.ACCESS_PROHIBITED
+            )
+    else:
+        if Users.get_num_users() != 0:
+            raise HTTPException(
+                status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.ACCESS_PROHIBITED
+            )
 
     if not validate_email_format(form_data.email.lower()):
         raise HTTPException(
@@ -273,7 +271,6 @@ async def signup(request: Request, response: Response, form_data: SignupForm):
 
 @router.post("/add", response_model=SigninResponse)
 async def add_user(form_data: AddUserForm, user=Depends(get_admin_user)):
-
     if not validate_email_format(form_data.email.lower()):
         raise HTTPException(
             status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.INVALID_EMAIL_FORMAT
@@ -283,7 +280,6 @@ async def add_user(form_data: AddUserForm, user=Depends(get_admin_user)):
         raise HTTPException(400, detail=ERROR_MESSAGES.EMAIL_TAKEN)
 
     try:
-
         print(form_data)
         hashed = get_password_hash(form_data.password)
         user = Auths.insert_new_auth(

backend/open_webui/apps/webui/routers/chats.py

@@ -1,34 +1,25 @@
-from fastapi import Depends, Request, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union, Optional
-from utils.utils import get_verified_user, get_admin_user
-from fastapi import APIRouter
-from pydantic import BaseModel
 import json
 import logging
+from typing import Optional
 
-from apps.webui.models.users import Users
-from apps.webui.models.chats import (
-    ChatModel,
-    ChatResponse,
-    ChatTitleForm,
+from open_webui.apps.webui.models.chats import (
     ChatForm,
-    ChatTitleIdResponse,
+    ChatResponse,
     Chats,
+    ChatTitleIdResponse,
 )
-
-
-from apps.webui.models.tags import (
-    TagModel,
-    ChatIdTagModel,
+from open_webui.apps.webui.models.tags import (
     ChatIdTagForm,
-    ChatTagsResponse,
+    ChatIdTagModel,
+    TagModel,
     Tags,
 )
-
-from constants import ERROR_MESSAGES
-
-from config import SRC_LOG_LEVELS, ENABLE_ADMIN_EXPORT, ENABLE_ADMIN_CHAT_ACCESS
+from open_webui.config import ENABLE_ADMIN_CHAT_ACCESS, ENABLE_ADMIN_EXPORT
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -61,11 +52,9 @@ async def get_session_user_chat_list(
 
 @router.delete("/", response_model=bool)
 async def delete_all_user_chats(request: Request, user=Depends(get_verified_user)):
-
-    if (
-        user.role == "user"
-        and not request.app.state.config.USER_PERMISSIONS["chat"]["deletion"]
-    ):
+    if user.role == "user" and not request.app.state.config.USER_PERMISSIONS.get(
+        "chat", {}
+    ).get("deletion", {}):
         raise HTTPException(
             status_code=status.HTTP_401_UNAUTHORIZED,
             detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
@@ -220,7 +209,6 @@ class TagNameForm(BaseModel):
 async def get_user_chat_list_by_tag_name(
     form_data: TagNameForm, user=Depends(get_verified_user)
 ):
-
     chat_ids = [
         chat_id_tag.chat_id
         for chat_id_tag in Tags.get_chat_ids_by_tag_name_and_user_id(
@@ -299,12 +287,13 @@ async def update_chat_by_id(
 
 @router.delete("/{id}", response_model=bool)
 async def delete_chat_by_id(request: Request, id: str, user=Depends(get_verified_user)):
-
     if user.role == "admin":
         result = Chats.delete_chat_by_id(id)
         return result
     else:
-        if not request.app.state.config.USER_PERMISSIONS["chat"]["deletion"]:
+        if not request.app.state.config.USER_PERMISSIONS.get("chat", {}).get(
+            "deletion", {}
+        ):
             raise HTTPException(
                 status_code=status.HTTP_401_UNAUTHORIZED,
                 detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
@@ -323,7 +312,6 @@ async def delete_chat_by_id(request: Request, id: str, user=Depends(get_verified
 async def clone_chat_by_id(id: str, user=Depends(get_verified_user)):
     chat = Chats.get_chat_by_id_and_user_id(id, user.id)
     if chat:
-
         chat_body = json.loads(chat.chat)
         updated_chat = {
             **chat_body,

backend/open_webui/apps/webui/routers/configs.py

@@ -1,29 +1,39 @@
-from fastapi import Response, Request
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union
-
-from fastapi import APIRouter
+from open_webui.config import BannerModel
+from fastapi import APIRouter, Depends, Request
 from pydantic import BaseModel
-import time
-import uuid
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
-from config import BannerModel
 
-from apps.webui.models.users import Users
-
-from utils.utils import (
-    get_password_hash,
-    get_verified_user,
-    get_admin_user,
-    create_token,
-)
-from utils.misc import get_gravatar_url, validate_email_format
-from constants import ERROR_MESSAGES
+from open_webui.config import get_config, save_config
 
 router = APIRouter()
 
 
+############################
+# ImportConfig
+############################
+
+
+class ImportConfigForm(BaseModel):
+    config: dict
+
+
+@router.post("/import", response_model=dict)
+async def import_config(form_data: ImportConfigForm, user=Depends(get_admin_user)):
+    save_config(form_data.config)
+    return get_config()
+
+
+############################
+# ExportConfig
+############################
+
+
+@router.get("/export", response_model=dict)
+async def export_config(user=Depends(get_admin_user)):
+    return get_config()
+
+
 class SetDefaultModelsForm(BaseModel):
     models: str
 

backend/open_webui/apps/webui/routers/documents.py

@@ -1,21 +1,16 @@
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union, Optional
-
-from fastapi import APIRouter
-from pydantic import BaseModel
 import json
+from typing import Optional
 
-from apps.webui.models.documents import (
-    Documents,
+from open_webui.apps.webui.models.documents import (
     DocumentForm,
-    DocumentUpdateForm,
-    DocumentModel,
     DocumentResponse,
+    Documents,
+    DocumentUpdateForm,
 )
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, status
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 router = APIRouter()
 

backend/open_webui/apps/webui/routers/files.py

@@ -1,42 +1,26 @@
-from fastapi import (
-    Depends,
-    FastAPI,
-    HTTPException,
-    status,
-    Request,
-    UploadFile,
-    File,
-    Form,
-)
-
-
-from datetime import datetime, timedelta
-from typing import Union, Optional
-from pathlib import Path
-
-from fastapi import APIRouter
-from fastapi.responses import StreamingResponse, JSONResponse, FileResponse
-
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.files import (
-    Files,
-    FileForm,
-    FileModel,
-    FileModelResponse,
-)
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-from importlib import util
+import logging
 import os
+import shutil
 import uuid
-import os, shutil, logging, re
+from pathlib import Path
+from typing import Optional
+from pydantic import BaseModel
+import mimetypes
 
 
-from config import SRC_LOG_LEVELS, UPLOAD_DIR
+from open_webui.apps.webui.models.files import FileForm, FileModel, Files
+from open_webui.apps.retrieval.main import process_file, ProcessFileForm
 
+from open_webui.config import UPLOAD_DIR
+from open_webui.env import SRC_LOG_LEVELS
+from open_webui.constants import ERROR_MESSAGES
+
+
+from fastapi import APIRouter, Depends, File, HTTPException, UploadFile, status
+from fastapi.responses import FileResponse, StreamingResponse
+
+
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -83,6 +67,13 @@ def upload_file(file: UploadFile = File(...), user=Depends(get_verified_user)):
             ),
         )
 
+        try:
+            process_file(ProcessFileForm(file_id=id))
+            file = Files.get_file_by_id(id=id)
+        except Exception as e:
+            log.exception(e)
+            log.error(f"Error processing file: {file.id}")
+
         if file:
             return file
         else:
@@ -168,6 +159,55 @@ async def get_file_by_id(id: str, user=Depends(get_verified_user)):
         )
 
 
+############################
+# Get File Data Content By Id
+############################
+
+
+@router.get("/{id}/data/content")
+async def get_file_data_content_by_id(id: str, user=Depends(get_verified_user)):
+    file = Files.get_file_by_id(id)
+
+    if file and (file.user_id == user.id or user.role == "admin"):
+        return {"content": file.data.get("content", "")}
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
+############################
+# Update File Data Content By Id
+############################
+
+
+class ContentForm(BaseModel):
+    content: str
+
+
+@router.post("/{id}/data/content/update")
+async def update_file_data_content_by_id(
+    id: str, form_data: ContentForm, user=Depends(get_verified_user)
+):
+    file = Files.get_file_by_id(id)
+
+    if file and (file.user_id == user.id or user.role == "admin"):
+        try:
+            process_file(ProcessFileForm(file_id=id, content=form_data.content))
+            file = Files.get_file_by_id(id=id)
+        except Exception as e:
+            log.exception(e)
+            log.error(f"Error processing file: {file.id}")
+
+        return {"content": file.data.get("content", "")}
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
 ############################
 # Get File Content By Id
 ############################
@@ -201,16 +241,32 @@ async def get_file_content_by_id(id: str, user=Depends(get_verified_user)):
     file = Files.get_file_by_id(id)
 
     if file and (file.user_id == user.id or user.role == "admin"):
-        file_path = Path(file.meta["path"])
+        file_path = file.meta.get("path")
+        if file_path:
+            file_path = Path(file_path)
 
-        # Check if the file already exists in the cache
-        if file_path.is_file():
-            print(f"file_path: {file_path}")
-            return FileResponse(file_path)
+            # Check if the file already exists in the cache
+            if file_path.is_file():
+                print(f"file_path: {file_path}")
+                return FileResponse(file_path)
+            else:
+                raise HTTPException(
+                    status_code=status.HTTP_404_NOT_FOUND,
+                    detail=ERROR_MESSAGES.NOT_FOUND,
+                )
         else:
-            raise HTTPException(
-                status_code=status.HTTP_404_NOT_FOUND,
-                detail=ERROR_MESSAGES.NOT_FOUND,
+            # File path doesn’t exist, return the content as .txt if possible
+            file_content = file.content.get("content", "")
+            file_name = file.filename
+
+            # Create a generator that encodes the file content
+            def generator():
+                yield file_content.encode("utf-8")
+
+            return StreamingResponse(
+                generator(),
+                media_type="text/plain",
+                headers={"Content-Disposition": f"attachment; filename={file_name}"},
             )
     else:
         raise HTTPException(

backend/open_webui/apps/webui/routers/functions.py

@@ -1,27 +1,18 @@
-from fastapi import Depends, FastAPI, HTTPException, status, Request
-from datetime import datetime, timedelta
-from typing import Union, Optional
+import os
+from pathlib import Path
+from typing import Optional
 
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.functions import (
-    Functions,
+from open_webui.apps.webui.models.functions import (
     FunctionForm,
     FunctionModel,
     FunctionResponse,
+    Functions,
 )
-from apps.webui.utils import load_function_module_by_id
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-from importlib import util
-import os
-from pathlib import Path
-
-from config import DATA_DIR, CACHE_DIR, FUNCTIONS_DIR
-
+from open_webui.apps.webui.utils import load_function_module_by_id, replace_imports
+from open_webui.config import CACHE_DIR, FUNCTIONS_DIR
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 router = APIRouter()
 
@@ -64,13 +55,11 @@ async def create_new_function(
 
     function = Functions.get_function_by_id(form_data.id)
     if function is None:
-        function_path = os.path.join(FUNCTIONS_DIR, f"{form_data.id}.py")
         try:
-            with open(function_path, "w") as function_file:
-                function_file.write(form_data.content)
-
+            form_data.content = replace_imports(form_data.content)
             function_module, function_type, frontmatter = load_function_module_by_id(
-                form_data.id
+                form_data.id,
+                content=form_data.content,
             )
             form_data.meta.manifest = frontmatter
 
@@ -183,13 +172,11 @@ async def toggle_global_by_id(id: str, user=Depends(get_admin_user)):
 async def update_function_by_id(
     request: Request, id: str, form_data: FunctionForm, user=Depends(get_admin_user)
 ):
-    function_path = os.path.join(FUNCTIONS_DIR, f"{id}.py")
-
     try:
-        with open(function_path, "w") as function_file:
-            function_file.write(form_data.content)
-
-        function_module, function_type, frontmatter = load_function_module_by_id(id)
+        form_data.content = replace_imports(form_data.content)
+        function_module, function_type, frontmatter = load_function_module_by_id(
+            id, content=form_data.content
+        )
         form_data.meta.manifest = frontmatter
 
         FUNCTIONS = request.app.state.FUNCTIONS
@@ -231,13 +218,6 @@ async def delete_function_by_id(
         if id in FUNCTIONS:
             del FUNCTIONS[id]
 
-        # delete the function file
-        function_path = os.path.join(FUNCTIONS_DIR, f"{id}.py")
-        try:
-            os.remove(function_path)
-        except Exception:
-            pass
-
     return result
 
 
@@ -304,7 +284,6 @@ async def update_function_valves_by_id(
 ):
     function = Functions.get_function_by_id(id)
     if function:
-
         if id in request.app.state.FUNCTIONS:
             function_module = request.app.state.FUNCTIONS[id]
         else:

backend/open_webui/apps/webui/routers/knowledge.py

@@ -0,0 +1,348 @@
+import json
+from typing import Optional, Union
+from pydantic import BaseModel
+from fastapi import APIRouter, Depends, HTTPException, status
+import logging
+
+from open_webui.apps.webui.models.knowledge import (
+    Knowledges,
+    KnowledgeUpdateForm,
+    KnowledgeForm,
+    KnowledgeResponse,
+)
+from open_webui.apps.webui.models.files import Files, FileModel
+from open_webui.apps.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.apps.retrieval.main import process_file, ProcessFileForm
+
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.utils.utils import get_admin_user, get_verified_user
+from open_webui.env import SRC_LOG_LEVELS
+
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["MODELS"])
+
+router = APIRouter()
+
+############################
+# GetKnowledgeItems
+############################
+
+
+@router.get(
+    "/", response_model=Optional[Union[list[KnowledgeResponse], KnowledgeResponse]]
+)
+async def get_knowledge_items(
+    id: Optional[str] = None, user=Depends(get_verified_user)
+):
+    if id:
+        knowledge = Knowledges.get_knowledge_by_id(id=id)
+
+        if knowledge:
+            return knowledge
+        else:
+            raise HTTPException(
+                status_code=status.HTTP_401_UNAUTHORIZED,
+                detail=ERROR_MESSAGES.NOT_FOUND,
+            )
+    else:
+        return [
+            KnowledgeResponse(**knowledge.model_dump())
+            for knowledge in Knowledges.get_knowledge_items()
+        ]
+
+
+############################
+# CreateNewKnowledge
+############################
+
+
+@router.post("/create", response_model=Optional[KnowledgeResponse])
+async def create_new_knowledge(form_data: KnowledgeForm, user=Depends(get_admin_user)):
+    knowledge = Knowledges.insert_new_knowledge(user.id, form_data)
+
+    if knowledge:
+        return knowledge
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.FILE_EXISTS,
+        )
+
+
+############################
+# GetKnowledgeById
+############################
+
+
+class KnowledgeFilesResponse(KnowledgeResponse):
+    files: list[FileModel]
+
+
+@router.get("/{id}", response_model=Optional[KnowledgeFilesResponse])
+async def get_knowledge_by_id(id: str, user=Depends(get_verified_user)):
+    knowledge = Knowledges.get_knowledge_by_id(id=id)
+
+    if knowledge:
+        file_ids = knowledge.data.get("file_ids", []) if knowledge.data else []
+        files = Files.get_files_by_ids(file_ids)
+
+        return KnowledgeFilesResponse(
+            **knowledge.model_dump(),
+            files=files,
+        )
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
+############################
+# UpdateKnowledgeById
+############################
+
+
+@router.post("/{id}/update", response_model=Optional[KnowledgeFilesResponse])
+async def update_knowledge_by_id(
+    id: str,
+    form_data: KnowledgeUpdateForm,
+    user=Depends(get_admin_user),
+):
+    knowledge = Knowledges.update_knowledge_by_id(id=id, form_data=form_data)
+
+    if knowledge:
+        file_ids = knowledge.data.get("file_ids", []) if knowledge.data else []
+        files = Files.get_files_by_ids(file_ids)
+
+        return KnowledgeFilesResponse(
+            **knowledge.model_dump(),
+            files=files,
+        )
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.ID_TAKEN,
+        )
+
+
+############################
+# AddFileToKnowledge
+############################
+
+
+class KnowledgeFileIdForm(BaseModel):
+    file_id: str
+
+
+@router.post("/{id}/file/add", response_model=Optional[KnowledgeFilesResponse])
+def add_file_to_knowledge_by_id(
+    id: str,
+    form_data: KnowledgeFileIdForm,
+    user=Depends(get_admin_user),
+):
+    knowledge = Knowledges.get_knowledge_by_id(id=id)
+    file = Files.get_file_by_id(form_data.file_id)
+    if not file:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+    if not file.data:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.FILE_NOT_PROCESSED,
+        )
+
+    # Add content to the vector database
+    try:
+        process_file(ProcessFileForm(file_id=form_data.file_id, collection_name=id))
+    except Exception as e:
+        log.debug(e)
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=str(e),
+        )
+
+    if knowledge:
+        data = knowledge.data or {}
+        file_ids = data.get("file_ids", [])
+
+        if form_data.file_id not in file_ids:
+            file_ids.append(form_data.file_id)
+            data["file_ids"] = file_ids
+
+            knowledge = Knowledges.update_knowledge_by_id(
+                id=id, form_data=KnowledgeUpdateForm(data=data)
+            )
+
+            if knowledge:
+                files = Files.get_files_by_ids(file_ids)
+
+                return KnowledgeFilesResponse(
+                    **knowledge.model_dump(),
+                    files=files,
+                )
+            else:
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT("knowledge"),
+                )
+        else:
+            raise HTTPException(
+                status_code=status.HTTP_400_BAD_REQUEST,
+                detail=ERROR_MESSAGES.DEFAULT("file_id"),
+            )
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
+@router.post("/{id}/file/update", response_model=Optional[KnowledgeFilesResponse])
+def update_file_from_knowledge_by_id(
+    id: str,
+    form_data: KnowledgeFileIdForm,
+    user=Depends(get_admin_user),
+):
+    knowledge = Knowledges.get_knowledge_by_id(id=id)
+    file = Files.get_file_by_id(form_data.file_id)
+    if not file:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    # Remove content from the vector database
+    VECTOR_DB_CLIENT.delete(
+        collection_name=knowledge.id, filter={"file_id": form_data.file_id}
+    )
+
+    # Add content to the vector database
+    try:
+        process_file(ProcessFileForm(file_id=form_data.file_id, collection_name=id))
+    except Exception as e:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=str(e),
+        )
+
+    if knowledge:
+        data = knowledge.data or {}
+        file_ids = data.get("file_ids", [])
+
+        files = Files.get_files_by_ids(file_ids)
+
+        return KnowledgeFilesResponse(
+            **knowledge.model_dump(),
+            files=files,
+        )
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
+############################
+# RemoveFileFromKnowledge
+############################
+
+
+@router.post("/{id}/file/remove", response_model=Optional[KnowledgeFilesResponse])
+def remove_file_from_knowledge_by_id(
+    id: str,
+    form_data: KnowledgeFileIdForm,
+    user=Depends(get_admin_user),
+):
+    knowledge = Knowledges.get_knowledge_by_id(id=id)
+    file = Files.get_file_by_id(form_data.file_id)
+    if not file:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    # Remove content from the vector database
+    VECTOR_DB_CLIENT.delete(
+        collection_name=knowledge.id, filter={"file_id": form_data.file_id}
+    )
+
+    result = VECTOR_DB_CLIENT.query(
+        collection_name=knowledge.id,
+        filter={"file_id": form_data.file_id},
+    )
+
+    Files.delete_file_by_id(form_data.file_id)
+
+    if knowledge:
+        data = knowledge.data or {}
+        file_ids = data.get("file_ids", [])
+
+        if form_data.file_id in file_ids:
+            file_ids.remove(form_data.file_id)
+            data["file_ids"] = file_ids
+
+            knowledge = Knowledges.update_knowledge_by_id(
+                id=id, form_data=KnowledgeUpdateForm(data=data)
+            )
+
+            if knowledge:
+                files = Files.get_files_by_ids(file_ids)
+
+                return KnowledgeFilesResponse(
+                    **knowledge.model_dump(),
+                    files=files,
+                )
+            else:
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT("knowledge"),
+                )
+        else:
+            raise HTTPException(
+                status_code=status.HTTP_400_BAD_REQUEST,
+                detail=ERROR_MESSAGES.DEFAULT("file_id"),
+            )
+    else:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
+############################
+# ResetKnowledgeById
+############################
+
+
+@router.post("/{id}/reset", response_model=Optional[KnowledgeResponse])
+async def reset_knowledge_by_id(id: str, user=Depends(get_admin_user)):
+    try:
+        VECTOR_DB_CLIENT.delete_collection(collection_name=id)
+    except Exception as e:
+        log.debug(e)
+        pass
+
+    knowledge = Knowledges.update_knowledge_by_id(
+        id=id, form_data=KnowledgeUpdateForm(data={"file_ids": []})
+    )
+    return knowledge
+
+
+############################
+# DeleteKnowledgeById
+############################
+
+
+@router.delete("/{id}/delete", response_model=bool)
+async def delete_knowledge_by_id(id: str, user=Depends(get_admin_user)):
+    try:
+        VECTOR_DB_CLIENT.delete_collection(collection_name=id)
+    except Exception as e:
+        log.debug(e)
+        pass
+    result = Knowledges.delete_knowledge_by_id(id=id)
+    return result

backend/open_webui/apps/webui/routers/memories.py

@@ -1,18 +1,13 @@
-from fastapi import Response, Request
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union, Optional
-
-from fastapi import APIRouter
+from fastapi import APIRouter, Depends, HTTPException, Request
 from pydantic import BaseModel
 import logging
+from typing import Optional
 
-from apps.webui.models.memories import Memories, MemoryModel
+from open_webui.apps.webui.models.memories import Memories, MemoryModel
+from open_webui.apps.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.utils.utils import get_verified_user
+from open_webui.env import SRC_LOG_LEVELS
 
-from utils.utils import get_verified_user
-from constants import ERROR_MESSAGES
-
-from config import SRC_LOG_LEVELS, CHROMA_CLIENT
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -55,14 +50,17 @@ async def add_memory(
     user=Depends(get_verified_user),
 ):
     memory = Memories.insert_new_memory(user.id, form_data.content)
-    memory_embedding = request.app.state.EMBEDDING_FUNCTION(memory.content)
 
-    collection = CHROMA_CLIENT.get_or_create_collection(name=f"user-memory-{user.id}")
-    collection.upsert(
-        documents=[memory.content],
-        ids=[memory.id],
-        embeddings=[memory_embedding],
-        metadatas=[{"created_at": memory.created_at}],
+    VECTOR_DB_CLIENT.upsert(
+        collection_name=f"user-memory-{user.id}",
+        items=[
+            {
+                "id": memory.id,
+                "text": memory.content,
+                "vector": request.app.state.EMBEDDING_FUNCTION(memory.content),
+                "metadata": {"created_at": memory.created_at},
+            }
+        ],
     )
 
     return memory
@@ -82,12 +80,10 @@ class QueryMemoryForm(BaseModel):
 async def query_memory(
     request: Request, form_data: QueryMemoryForm, user=Depends(get_verified_user)
 ):
-    query_embedding = request.app.state.EMBEDDING_FUNCTION(form_data.content)
-    collection = CHROMA_CLIENT.get_or_create_collection(name=f"user-memory-{user.id}")
-
-    results = collection.query(
-        query_embeddings=[query_embedding],
-        n_results=form_data.k,  # how many results to return
+    results = VECTOR_DB_CLIENT.search(
+        collection_name=f"user-memory-{user.id}",
+        vectors=[request.app.state.EMBEDDING_FUNCTION(form_data.content)],
+        limit=form_data.k,
     )
 
     return results
@@ -100,17 +96,25 @@ async def query_memory(
 async def reset_memory_from_vector_db(
     request: Request, user=Depends(get_verified_user)
 ):
-    CHROMA_CLIENT.delete_collection(f"user-memory-{user.id}")
-    collection = CHROMA_CLIENT.get_or_create_collection(name=f"user-memory-{user.id}")
+    VECTOR_DB_CLIENT.delete_collection(f"user-memory-{user.id}")
 
     memories = Memories.get_memories_by_user_id(user.id)
-    for memory in memories:
-        memory_embedding = request.app.state.EMBEDDING_FUNCTION(memory.content)
-        collection.upsert(
-            documents=[memory.content],
-            ids=[memory.id],
-            embeddings=[memory_embedding],
-        )
+    VECTOR_DB_CLIENT.upsert(
+        collection_name=f"user-memory-{user.id}",
+        items=[
+            {
+                "id": memory.id,
+                "text": memory.content,
+                "vector": request.app.state.EMBEDDING_FUNCTION(memory.content),
+                "metadata": {
+                    "created_at": memory.created_at,
+                    "updated_at": memory.updated_at,
+                },
+            }
+            for memory in memories
+        ],
+    )
+
     return True
 
 
@@ -125,7 +129,7 @@ async def delete_memory_by_user_id(user=Depends(get_verified_user)):
 
     if result:
         try:
-            CHROMA_CLIENT.delete_collection(f"user-memory-{user.id}")
+            VECTOR_DB_CLIENT.delete_collection(f"user-memory-{user.id}")
         except Exception as e:
             log.error(e)
         return True
@@ -150,16 +154,18 @@ async def update_memory_by_id(
         raise HTTPException(status_code=404, detail="Memory not found")
 
     if form_data.content is not None:
-        memory_embedding = request.app.state.EMBEDDING_FUNCTION(form_data.content)
-        collection = CHROMA_CLIENT.get_or_create_collection(
-            name=f"user-memory-{user.id}"
-        )
-        collection.upsert(
-            documents=[form_data.content],
-            ids=[memory.id],
-            embeddings=[memory_embedding],
-            metadatas=[
-                {"created_at": memory.created_at, "updated_at": memory.updated_at}
+        VECTOR_DB_CLIENT.upsert(
+            collection_name=f"user-memory-{user.id}",
+            items=[
+                {
+                    "id": memory.id,
+                    "text": memory.content,
+                    "vector": request.app.state.EMBEDDING_FUNCTION(memory.content),
+                    "metadata": {
+                        "created_at": memory.created_at,
+                        "updated_at": memory.updated_at,
+                    },
+                }
             ],
         )
 
@@ -176,10 +182,9 @@ async def delete_memory_by_id(memory_id: str, user=Depends(get_verified_user)):
     result = Memories.delete_memory_by_id_and_user_id(memory_id, user.id)
 
     if result:
-        collection = CHROMA_CLIENT.get_or_create_collection(
-            name=f"user-memory-{user.id}"
+        VECTOR_DB_CLIENT.delete(
+            collection_name=f"user-memory-{user.id}", ids=[memory_id]
         )
-        collection.delete(ids=[memory_id])
         return True
 
     return False

backend/open_webui/apps/webui/routers/models.py

@@ -1,15 +1,14 @@
-from fastapi import Depends, FastAPI, HTTPException, status, Request
-from datetime import datetime, timedelta
-from typing import Union, Optional
+from typing import Optional
 
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.models import Models, ModelModel, ModelForm, ModelResponse
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
+from open_webui.apps.webui.models.models import (
+    ModelForm,
+    ModelModel,
+    ModelResponse,
+    Models,
+)
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 router = APIRouter()
 
@@ -19,8 +18,18 @@ router = APIRouter()
 
 
 @router.get("/", response_model=list[ModelResponse])
-async def get_models(user=Depends(get_verified_user)):
-    return Models.get_all_models()
+async def get_models(id: Optional[str] = None, user=Depends(get_verified_user)):
+    if id:
+        model = Models.get_model_by_id(id)
+        if model:
+            return [model]
+        else:
+            raise HTTPException(
+                status_code=status.HTTP_401_UNAUTHORIZED,
+                detail=ERROR_MESSAGES.NOT_FOUND,
+            )
+    else:
+        return Models.get_all_models()
 
 
 ############################
@@ -51,24 +60,6 @@ async def add_new_model(
             )
 
 
-############################
-# GetModelById
-############################
-
-
-@router.get("/", response_model=Optional[ModelModel])
-async def get_model_by_id(id: str, user=Depends(get_verified_user)):
-    model = Models.get_model_by_id(id)
-
-    if model:
-        return model
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
 ############################
 # UpdateModelById
 ############################

backend/open_webui/apps/webui/routers/prompts.py

@@ -1,15 +1,9 @@
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union, Optional
+from typing import Optional
 
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.prompts import Prompts, PromptForm, PromptModel
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
+from open_webui.apps.webui.models.prompts import PromptForm, PromptModel, Prompts
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, status
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 router = APIRouter()
 

backend/open_webui/apps/webui/routers/tools.py

@@ -1,20 +1,14 @@
-from fastapi import Depends, HTTPException, status, Request
-from typing import Optional
-
-from fastapi import APIRouter
-
-from apps.webui.models.tools import Tools, ToolForm, ToolModel, ToolResponse
-from apps.webui.utils import load_toolkit_module_by_id
-
-from utils.utils import get_admin_user, get_verified_user
-from utils.tools import get_tools_specs
-from constants import ERROR_MESSAGES
-
 import os
 from pathlib import Path
+from typing import Optional
 
-from config import DATA_DIR, CACHE_DIR
-
+from open_webui.apps.webui.models.tools import ToolForm, ToolModel, ToolResponse, Tools
+from open_webui.apps.webui.utils import load_toolkit_module_by_id, replace_imports
+from open_webui.config import CACHE_DIR, DATA_DIR
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from open_webui.utils.tools import get_tools_specs
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 TOOLS_DIR = f"{DATA_DIR}/tools"
 os.makedirs(TOOLS_DIR, exist_ok=True)
@@ -65,12 +59,11 @@ async def create_new_toolkit(
 
     toolkit = Tools.get_tool_by_id(form_data.id)
     if toolkit is None:
-        toolkit_path = os.path.join(TOOLS_DIR, f"{form_data.id}.py")
         try:
-            with open(toolkit_path, "w") as tool_file:
-                tool_file.write(form_data.content)
-
-            toolkit_module, frontmatter = load_toolkit_module_by_id(form_data.id)
+            form_data.content = replace_imports(form_data.content)
+            toolkit_module, frontmatter = load_toolkit_module_by_id(
+                form_data.id, content=form_data.content
+            )
             form_data.meta.manifest = frontmatter
 
             TOOLS = request.app.state.TOOLS
@@ -132,13 +125,11 @@ async def update_toolkit_by_id(
     form_data: ToolForm,
     user=Depends(get_admin_user),
 ):
-    toolkit_path = os.path.join(TOOLS_DIR, f"{id}.py")
-
     try:
-        with open(toolkit_path, "w") as tool_file:
-            tool_file.write(form_data.content)
-
-        toolkit_module, frontmatter = load_toolkit_module_by_id(id)
+        form_data.content = replace_imports(form_data.content)
+        toolkit_module, frontmatter = load_toolkit_module_by_id(
+            id, content=form_data.content
+        )
         form_data.meta.manifest = frontmatter
 
         TOOLS = request.app.state.TOOLS
@@ -183,10 +174,6 @@ async def delete_toolkit_by_id(request: Request, id: str, user=Depends(get_admin
         if id in TOOLS:
             del TOOLS[id]
 
-        # delete the toolkit file
-        toolkit_path = os.path.join(TOOLS_DIR, f"{id}.py")
-        os.remove(toolkit_path)
-
     return result
 
 

backend/open_webui/apps/webui/routers/users.py

@@ -1,33 +1,20 @@
-from fastapi import Response, Request
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import Union, Optional
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import time
-import uuid
 import logging
+from typing import Optional
 
-from apps.webui.models.users import (
+from open_webui.apps.webui.models.auths import Auths
+from open_webui.apps.webui.models.chats import Chats
+from open_webui.apps.webui.models.users import (
     UserModel,
-    UserUpdateForm,
     UserRoleUpdateForm,
-    UserSettings,
     Users,
+    UserSettings,
+    UserUpdateForm,
 )
-from apps.webui.models.auths import Auths
-from apps.webui.models.chats import Chats
-
-from utils.utils import (
-    get_verified_user,
-    get_password_hash,
-    get_current_user,
-    get_admin_user,
-)
-from constants import ERROR_MESSAGES
-
-from config import SRC_LOG_LEVELS
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+from fastapi import APIRouter, Depends, HTTPException, Request, status
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_password_hash, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MODELS"])
@@ -69,7 +56,6 @@ async def update_user_permissions(
 
 @router.post("/update/role", response_model=Optional[UserModel])
 async def update_user_role(form_data: UserRoleUpdateForm, user=Depends(get_admin_user)):
-
     if user.id != form_data.id and form_data.id != Users.get_first_user().id:
         return Users.update_user_role_by_id(form_data.id, form_data.role)
 
@@ -173,7 +159,6 @@ class UserResponse(BaseModel):
 
 @router.get("/{user_id}", response_model=UserResponse)
 async def get_user_by_id(user_id: str, user=Depends(get_verified_user)):
-
     # Check if user_id is a shared chat
     # If it is, get the user_id from the chat
     if user_id.startswith("shared-"):

backend/open_webui/apps/webui/routers/utils.py

@@ -1,23 +1,17 @@
-from pathlib import Path
 import site
+from pathlib import Path
 
-from fastapi import APIRouter, UploadFile, File, Response
-from fastapi import Depends, HTTPException, status
-from starlette.responses import StreamingResponse, FileResponse
-from pydantic import BaseModel
-
-
-from fpdf import FPDF
-import markdown
 import black
-
-
-from utils.utils import get_admin_user
-from utils.misc import calculate_sha256, get_gravatar_url
-
-from config import OLLAMA_BASE_URLS, DATA_DIR, UPLOAD_DIR, ENABLE_ADMIN_EXPORT
-from constants import ERROR_MESSAGES
-
+import markdown
+from open_webui.config import DATA_DIR, ENABLE_ADMIN_EXPORT
+from open_webui.env import FONTS_DIR
+from open_webui.constants import ERROR_MESSAGES
+from fastapi import APIRouter, Depends, HTTPException, Response, status
+from fpdf import FPDF
+from pydantic import BaseModel
+from starlette.responses import FileResponse
+from open_webui.utils.misc import get_gravatar_url
+from open_webui.utils.utils import get_admin_user
 
 router = APIRouter()
 
@@ -64,14 +58,11 @@ class ChatForm(BaseModel):
 async def download_chat_as_pdf(
     form_data: ChatForm,
 ):
+    global FONTS_DIR
+
     pdf = FPDF()
     pdf.add_page()
 
-    # When running in docker, workdir is /app/backend, so fonts is in /app/backend/static/fonts
-    FONTS_DIR = Path("./static/fonts")
-
-    # Non Docker Installation
-
     # When running using `pip install` the static directory is in the site packages.
     if not FONTS_DIR.exists():
         FONTS_DIR = Path(site.getsitepackages()[0]) / "static/fonts"
@@ -115,7 +106,7 @@ async def download_chat_as_pdf(
     return Response(
         content=bytes(pdf_bytes),
         media_type="application/pdf",
-        headers={"Content-Disposition": f"attachment;filename=chat.pdf"},
+        headers={"Content-Disposition": "attachment;filename=chat.pdf"},
     )
 
 
@@ -126,7 +117,7 @@ async def download_db(user=Depends(get_admin_user)):
             status_code=status.HTTP_401_UNAUTHORIZED,
             detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
         )
-    from apps.webui.internal.db import engine
+    from open_webui.apps.webui.internal.db import engine
 
     if engine.name != "sqlite":
         raise HTTPException(

backend/open_webui/apps/webui/utils.py

@@ -0,0 +1,171 @@
+import os
+import re
+import subprocess
+import sys
+from importlib import util
+import types
+import tempfile
+
+from open_webui.apps.webui.models.functions import Functions
+from open_webui.apps.webui.models.tools import Tools
+from open_webui.config import FUNCTIONS_DIR, TOOLS_DIR
+
+
+def extract_frontmatter(content):
+    """
+    Extract frontmatter as a dictionary from the provided content string.
+    """
+    frontmatter = {}
+    frontmatter_started = False
+    frontmatter_ended = False
+    frontmatter_pattern = re.compile(r"^\s*([a-z_]+):\s*(.*)\s*$", re.IGNORECASE)
+
+    try:
+        lines = content.splitlines()
+        if len(lines) < 1 or lines[0].strip() != '"""':
+            # The content doesn't start with triple quotes
+            return {}
+
+        frontmatter_started = True
+
+        for line in lines[1:]:
+            if '"""' in line:
+                if frontmatter_started:
+                    frontmatter_ended = True
+                    break
+
+            if frontmatter_started and not frontmatter_ended:
+                match = frontmatter_pattern.match(line)
+                if match:
+                    key, value = match.groups()
+                    frontmatter[key.strip()] = value.strip()
+
+    except Exception as e:
+        print(f"An error occurred: {e}")
+        return {}
+
+    return frontmatter
+
+
+def replace_imports(content):
+    """
+    Replace the import paths in the content.
+    """
+    replacements = {
+        "from utils": "from open_webui.utils",
+        "from apps": "from open_webui.apps",
+        "from main": "from open_webui.main",
+        "from config": "from open_webui.config",
+    }
+
+    for old, new in replacements.items():
+        content = content.replace(old, new)
+
+    return content
+
+
+def load_toolkit_module_by_id(toolkit_id, content=None):
+
+    if content is None:
+        tool = Tools.get_tool_by_id(toolkit_id)
+        if not tool:
+            raise Exception(f"Toolkit not found: {toolkit_id}")
+
+        content = tool.content
+
+        content = replace_imports(content)
+        Tools.update_tool_by_id(toolkit_id, {"content": content})
+    else:
+        frontmatter = extract_frontmatter(content)
+        # Install required packages found within the frontmatter
+        install_frontmatter_requirements(frontmatter.get("requirements", ""))
+
+    module_name = f"tool_{toolkit_id}"
+    module = types.ModuleType(module_name)
+    sys.modules[module_name] = module
+
+    # Create a temporary file and use it to define `__file__` so
+    # that it works as expected from the module's perspective.
+    temp_file = tempfile.NamedTemporaryFile(delete=False)
+    temp_file.close()
+    try:
+        with open(temp_file.name, "w", encoding="utf-8") as f:
+            f.write(content)
+        module.__dict__["__file__"] = temp_file.name
+
+        # Executing the modified content in the created module's namespace
+        exec(content, module.__dict__)
+        frontmatter = extract_frontmatter(content)
+        print(f"Loaded module: {module.__name__}")
+
+        # Create and return the object if the class 'Tools' is found in the module
+        if hasattr(module, "Tools"):
+            return module.Tools(), frontmatter
+        else:
+            raise Exception("No Tools class found in the module")
+    except Exception as e:
+        print(f"Error loading module: {toolkit_id}: {e}")
+        del sys.modules[module_name]  # Clean up
+        raise e
+    finally:
+        os.unlink(temp_file.name)
+
+
+def load_function_module_by_id(function_id, content=None):
+    if content is None:
+        function = Functions.get_function_by_id(function_id)
+        if not function:
+            raise Exception(f"Function not found: {function_id}")
+        content = function.content
+
+        content = replace_imports(content)
+        Functions.update_function_by_id(function_id, {"content": content})
+    else:
+        frontmatter = extract_frontmatter(content)
+        install_frontmatter_requirements(frontmatter.get("requirements", ""))
+
+    module_name = f"function_{function_id}"
+    module = types.ModuleType(module_name)
+    sys.modules[module_name] = module
+
+    # Create a temporary file and use it to define `__file__` so
+    # that it works as expected from the module's perspective.
+    temp_file = tempfile.NamedTemporaryFile(delete=False)
+    temp_file.close()
+    try:
+        with open(temp_file.name, "w", encoding="utf-8") as f:
+            f.write(content)
+        module.__dict__["__file__"] = temp_file.name
+
+        # Execute the modified content in the created module's namespace
+        exec(content, module.__dict__)
+        frontmatter = extract_frontmatter(content)
+        print(f"Loaded module: {module.__name__}")
+
+        # Create appropriate object based on available class type in the module
+        if hasattr(module, "Pipe"):
+            return module.Pipe(), "pipe", frontmatter
+        elif hasattr(module, "Filter"):
+            return module.Filter(), "filter", frontmatter
+        elif hasattr(module, "Action"):
+            return module.Action(), "action", frontmatter
+        else:
+            raise Exception("No Function class found in the module")
+    except Exception as e:
+        print(f"Error loading module: {function_id}: {e}")
+        del sys.modules[module_name]  # Cleanup by removing the module in case of error
+
+        Functions.update_function_by_id(function_id, {"is_active": False})
+        raise e
+    finally:
+        os.unlink(temp_file.name)
+
+
+def install_frontmatter_requirements(requirements):
+    if requirements:
+        req_list = [req.strip() for req in requirements.split(",")]
+        for req in req_list:
+            print(f"Installing requirement: {req}")
+            subprocess.check_call([sys.executable, "-m", "pip", "install", req])
+    else:
+        print("No requirements found in frontmatter.")