Releases: cloudera/CML_AMP_RAG_Studio
Releases · cloudera/CML_AMP_RAG_Studio
1.20.0-beta
basic openai support (#215) * quick & dirty hardcoded openai support * fill out the backend config for OpenAi * FE support for Open AI * ruff! * rename fields to more closely match standard env vars
1.19.0
Allow users to provide a CDP token for access to CAII (#211) * Add new endpoint for writing a cdp token to /tmp/jwt * swap to using text area * revert a CAII integration point that was broken with Crew changes * wip on adding new nav * add ui for setting cdp token * styling token box * add error alert to chat page * move custom cdp_token file out of tmp so it persists across restarts * clean up pythonic issues * get rid of print * fix bugs with logging metrics and filtering ds * fix issue with streaming chat with empty kb * disable buttons until full response * bump qdrant-client version down * remove links from copied response * don't create a new line on handle chat submit via enter * WIP cdp auth token lastFile:ui/src/pages/Settings/AuthenticationFields.tsx * drop databases lastFile:ui/src/pages/Settings/AuthenticationFields.tsx * styled tooltip * tidying up CDP auth token UX lastFile:ui/src/routes/_layout/chats/_layout-chats/index.tsx * python checks * show error message on session route and reformat * add error handling for invalid auth token * move comment * try just disabling the tab * create second hook with different key for settings nav * fix naming of cdp_token * show error message on fallback * rename component and move * drop databases lastFile:ui/src/pages/RagChatTab/ChatOutput/ChatMessages/ChatMessageBody.tsx * make enhanced pdf processing optional * remove hardcoded event * fix unused * remove extra copyright * remove extra space from response and center citation --------- Co-authored-by: Elijah Williams <ewilliams@cloudera.com> Co-authored-by: Baasit Sharief <baasitsharief@gmail.com> Co-authored-by: Michael Liu <mliu@cloudera.com>
1.18.0
Update release version to 1.18.0-beta
1.18.0-beta
Streaming!!!! (#202) * wip on simple streaming * simple poc for streaming * remove usage from RagChatQueryInput.tsx * remove stream hypothetical * remove unused import * wip on doing something once the gen is done * progress on generators * go back to simple streaming only endpoint * wip lastFile:llm-service/app/services/chat.py * add response id on every chunk returned lastFile:llm-service/app/routers/index/sessions/__init__.py * remove duplicate calls, but still not rendering * getting there * Consolidate response_id generation * wip lastFile:ui/src/api/chatApi.ts * drop databases lastFile:ui/src/pages/RagChatTab/ChatOutput/Loaders/PendingRagOutputSkeleton.tsx * mob next [ci-skip] [ci skip] [skip ci] lastFile:ui/src/pages/RagChatTab/ChatOutput/Loaders/PendingRagOutputSkeleton.tsx * mob next [ci-skip] [ci skip] [skip ci] lastFile:llm-service/app/routers/index/sessions/__init__.py * small refactor * remove deps * things are getting close * wip lastFile:ui/src/pages/RagChatTab/ChatOutput/Placeholders/SuggestedQuestionsCards.tsx * drop databases lastFile:llm-service/app/services/chat.py * wip lastFile:llm-service/app/services/chat.py * mob next [ci-skip] [ci skip] [skip ci] lastFile:llm-service/app/services/chat.py * drop databases lastFile:llm-service/app/services/chat.py * wip lastFile:llm-service/app/services/chat.py * fixing scrolling * only show loading nodes if kb * remove unused * removing active loading state * fix mypy issues * ruff * Update release version to dev-testing * handle file not found error for summaries when local * remove log * renaming * better error handling * bump bedrock to use max tokens of 1024 * python refactoring lastFile:llm-service/app/routers/index/sessions/__init__.py * mob next [ci-skip] [ci skip] [skip ci] lastFile:llm-service/app/routers/index/sessions/__init__.py * nits --------- Co-authored-by: Elijah Williams <ewilliams@cloudera.com> Co-authored-by: Michael Liu <mliu@cloudera.com> Co-authored-by: Baasit Sharief <baasitsharief@gmail.com> Co-authored-by: actions-user <actions@github.com>
1.17.0
- We're externalizing storage options in an effort to modularize RAG Studio
a. Users can now configure both chat history and document summaries to be stored in S3 as well as the CML file system - To improve chat responsiveness, we now paginate chat history calls
- Additionally, we've performed work to allow us to break apart RAG Studio to run across multiple CML applications. RAG Studio will continue to run in a single application with this release; however, we have the option to turn it on with some minor changes.
1.17.0-beta
Merge pull request #182 from cloudera/mob/main Move Qdrant to separate app and Summaries to S3
1.16.0
Merge pull request #179 from cloudera/mob/main Manage settings in the RAG Studio UI itself
1.15.0-beta
Merge pull request #177 from cloudera/mob/main Clean up session moving UI, tweak the feedback UI for positive feedback
1.15.0
Project ID filtering on analytics. Move sessions between projects. (#…
1.14.0
Merge pull request #173 from cloudera/mob/main OpenAPI exposure, Updates to KB lists, and bug fixes