* paperless ngx data connector
* wip resync paperless ngx
* fix generateChunkSource for resyncing paperless ngx
* lint
* Refactor Paperless-NGX connector
Fix issue with date rendering in tooltip + extended width
Move tooltip details to be column for more space
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* add option to disable swagger docs via env
* update .example.env for docker and server
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* feat: add web app manifest and mobile PWA meta tags
* feat: serve dynamic manifest.json with custom branding for pwa
* feat: add ios status bar theming for pwa
* fix: prevent overscroll behavior for mobile
* fix: prevent ios safari auto-zoom on chat input
* fix: remove theme-color meta tags conflicting with ios status bar
* fix: add missing apple-mobile-web-app-capable meta tag for ios pwa
* fix: move catch-all route after manifest endpoint to prevent interception
* feat: add pwa detection helper and conditional styling for standalone mode
* PWA refactor
* undo changes to native CSS
* class fix
* proper response obj
* fix patch for import
* fix manifest errors
---------
Co-authored-by: Christian De Santis <christian.constantino98@gmail.com>
* fix(server): correct Qdrant batching logic for large uploads
Closes#4544
* modify O(3N) process to O(N)
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Migrate gemini agents away from `Untooled`
* disable agents for gemma models as they are not supported for tool calling
* Dev build
resolve#4452 via function name prefix and then stripping within provider
* Reimplement Cohere models
- Redo LLM implementation to grab models from endpoint and pre-filter
- Migrate embedding models to also grab from remote
- Add records for easy context window lookup'
* fix comment
* auto model context limit detection for ollama llm provider
* auto model context limit detection for lmstudio llm provider
* Patch Ollama to function and sync context windows like Foundry
* normalize how model context windows are cached from endpoint service
todo: move this into global utility class with MODEL_MAP
eager load models on boot to pre-cache them
add performance model improvements into ollama agent as well as apply n_ctx
* remove debug log
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* add microsoft foundry local llm and agent providers
* minor change to fix early stop token + overloading of context window
always use user defined window _unless_ it is larger than the models real contenxt window
cache the context windows when we can from the API (0.7.*)+
Unload model forcefully on model change to prevent resource hogging
* add back token preference since some models have very large windows and can crash a machine
normalize cases
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* WIP agentic tool call streaming
- OpenAI
- Anthropic
- Azure OpenAI
* WIP rest of providers EXCLUDES Bedrock and GenericOpenAI
* patch untooled complete/streaming to use chatCallback provider from provider class and not assume OpenAI client struct
example: Ollama
* modify ollama to function with its own overrides
normalize completion/stream outputs across providers/untooled
* dev build
* fix message sanization for anthropic agent streaming
* wip fix anthropic agentic streaming sanitization
* patch gemini, webgenui, generic aibitat providers + disable providers unable to test
* refactor anthropic aibitat provider for empty message and tool call formatting
* Add frontend missing prop check
update Azure for streaming support
update Gemini to streamting support on gemini-* models
generic OpenAI disable streaming
verify localAI support
verify NVIDIA Nim support
* DPAIS, remove temp from call, support streaming'
* remove 0 temp to remove possibility of bad temp error/500s/400s
* Patch condition where model is non-streamable and no tools are present or called resulting in the provider `handleFunctionCallChat` being called - which returns a string.
This would then fail in Untooled.complete since response would be a string and not the expected `response.choices?.[0]?.message`
Modified this line to handle both conditions for stream/non-streaming and tool presence or lack thereof
* Allow generic Openai to be streamable since using untooled it should work fine
honor disabled streaming for provider where that concern may apply for regular chats
* rename function and more gemini-specific function to gemini provider
* add comments for readability
.complete on azure should be non-streaming as this is the sync response
* migrate CometAPI, but disable as we cannot test
---------
Co-authored-by: shatfield4 <seanhatfield5@gmail.com>
* Fix system prompt variable color logic by removing unused variable type from switch statement and adding new types.
* Add workspace id, name and user id as default system prompt variables
* Combine user and workspace variable evaluations into a single if statment, reducing redundant code.
* minor refactor
* add systemPromptVariable expandSystemPromptVariables test cases
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Fix JSDOC for updateOrCreateCollection
* Add sanitizeForJsonb method to PGVector for safe JSONB handling
This new method recursively sanitizes values intended for JSONB storage, removing disallowed control characters and ensuring safe insertion into PostgreSQL. The method is integrated into the vector insertion process to sanitize metadata before database operations.
* Add unit tests for PGVector.sanitizeForJsonb method
This commit introduces a comprehensive test suite for the PGVector.sanitizeForJsonb method, ensuring it correctly handles various input types, including null, undefined, strings with disallowed control characters, objects, arrays, and Date objects. The tests verify that the method sanitizes inputs without mutating the original data structures.
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* Add HTTP request logging middleware for development mode
- Introduced httpLogger middleware to log HTTP requests and responses.
- Enabled logging only in development mode to assist with debugging.
* Update httpLogger middleware to disable time logging by default
* Add httpLogger middleware for development mode in collector service
* Refactor httpLogger middleware to rename timeLogs parameter to enableTimestamps for clarity
* Make HTTP Logger only mount in development and environment flag is enabled.
* Update .env.example to clarify HTTP Logger configuration comments
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* Add className property to various LLM and embedder classes to fix logging bug after minification
* Fix bug with this.log method by applying the missing private field symbol
* migrate openai llm provider to use responses api
* add back image support
* dont recalc tokens from OpenAI since we get metrics back
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* Add ENV to configure api request delay for generic open ai embedding engine
* yarn lint formatting
* refactor
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Add User-Agent header on the requests sent by Generic OpenAI providers.
* Moved getAnythingLLMUserAgent helper fn to server/endpoints/utils.js and changed fallback version string to "unknown"
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* Added the ability to pass in metadata to the /document/upload/{folderName} endpoint
* Added the ability to pass in metadata to the /document/upload-link endpoint
* feat: added metadata to document/upload api endpoint
* simplify optional metadata in document dev api endpoints
* lint
* patch handling of metadata in dev api
* Linting, small comments
---------
Co-authored-by: jstawskigmi <jstawski@getmyinterns.org>
Co-authored-by: shatfield4 <seanhatfield5@gmail.com>
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* feat: Implement CometAPI integration for chat completions and model management
- Added CometApiLLM class for handling chat completions using CometAPI.
- Implemented model synchronization and caching mechanisms.
- Introduced streaming support for chat responses with timeout handling.
- Created CometApiProvider class for agent interactions with CometAPI.
- Enhanced error handling and logging throughout the integration.
- Established a structure for managing function calls and completions.
* linting
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* export image support for json and jsonl
* add tests and cleanup functionality
* add test for convertTo prepare function
* comment
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* add chroma cloud as new vector db provider
* update docker example env
* extend chroma class to chroma cloud
* update readme
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Added exa-search case to the search provider switch in web-browsing.js
* Added ExaSearchOptions component for API key input
* update
* Patch missing image crashing UI
Fix issue where ENV key did not exist or was saved on click
Update copy for provider
Add Docs for ENV keys for manual placements
update systemssettings for returning key saved to UI
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Create parse endpoint in collector (#4212)
* create parse endpoint in collector
* revert cleanup temp util call
* lint
* remove unused cleanupTempDocuments function
* revert slug change
minor change for destinations
---------
Co-authored-by: timothycarambat <rambat1010@gmail.com>
* Add parsed files table and parse server endpoints (#4222)
* add workspace_parsed_files table + parse endpoints/models
* remove dev api parse endpoint
* remove unneeded imports
* iterate over all files + remove unneeded update function + update telemetry debounce
* Upload UI/UX context window check + frontend alert (#4230)
* prompt user to embed if exceeds prompt window + handle embed + handle cancel
* add tokenCountEstimate to workspace_parsed_files + optimizations
* use util for path locations + use safeJsonParse
* add modal for user decision on overflow of context window
* lint
* dynamic fetching of provider/model combo + inject parsed documents
* remove unneeded comments
* popup ui for attaching/removing files + warning to embed + wip fetching states on update
* remove prop drilling, fetch files/limits directly in attach files popup
* rework ux of FE + BE optimizations
* fix ux of FE + BE optimizations
* Implement bidirectional sync for parsed file states
linting
small changes and comments
* move parse support to another endpoint file
simplify calls and loading of records
* button borders
* enable default users to upload parsed files but NOT embed
* delete cascade on user/workspace/thread deletion to remove parsedFileRecord
* enable bgworker with "always" jobs and optional document sync jobs
orphan document job: Will find any broken reference files to prevent overpollution of the storage folder. This will run 10s after boot and every 12hr after
* change run timeout for orphan job to 1m to allow settling before spawning a worker
* linting and cleanup pr
---------
Co-authored-by: Timothy Carambat <rambat1010@gmail.com>
* dev build
* fix tooltip hiding during embedding overflow files
* prevent crash log from ERRNO on parse files
* unused import
* update docs link
* Migrate parsed-files to GET endpoint
patch logic for grabbing models names from utils
better handling for undetermined context windows (null instead of Pos_INIFI)
UI placeholder for null context windows
* patch URL
---------
Co-authored-by: Sean Hatfield <seanhatfield5@gmail.com>
