Refactor Langchain Handler #8998
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@paxcema - any major comments ? Or am I fine to merge this one |
paxcema
approved these changes
Mar 27, 2024
Merged
hamishfagg
added a commit
that referenced
this pull request
Apr 10, 2024
* safe extract for all tar files * fix * tests for: - create empty table - interval function * added module and skeleton for GoogleServiceAccountOauth2Utilities * added the NoCredentialsException * raised NoCredentialsException if no creds provided * added method of getting credentials: file and JSON * added the method for download creds file * refactored method for getting credentials: URL * renamed class to GoogleServiceAccountOAuth2Manager * added error handling * imported the lclass to the main utils package * added parsing for JSON credentials * separated parsing credentials JSON to new function * updated Vertex handler to use new credentials params * updated Vertex client to use GoogleServiceAccountOAuth2Manager * imported GoogleServiceAccountOAuth2Manager to main auth package * removing redundant package lists and sentence-transformers library (#8971) * removing redundant package lists and sentence-transformers library * Add input column to RAG tests and update requirements Added 'input_column' to RAG tests in `test_rag.py` to accommodate for changes in the function signatures. Furthermore, the RAG handler requirements have been updated to include 'sentence-transformers', required for HuggingFaceEmbeddings from the langchain-community package. * Add allow_dangerous_deserialization in FAISS loading and fix typo in unit tests. A new allow_dangerous_deserialization parameter was added to the FAISS loading in RAG handler settings, providing additional flexibility for deserialization. Additionally, a typo in the input column name within unit tests was fixed, which was causing the tests to fail. --------- Co-authored-by: dusvyat <13661123+dusvyat@users.noreply.github.com> * Fix tarfile safe extract * sql dep * Refactor Langchain Handler (#8998) * Refactor langchain_handler * Added more tests * Add back mdb_read tool by default * Remove prints * Use updated import path * Update test_llm_utils * Add pydantic to reqs * Remove pydantic from handler reqs not that it is in main reqs * Remove explicit pydantic versioning to avoid conflict * added method to get creds and used it in create and predict * updated credentials parsing logic * renamed GoogleOAuth2Manager to GoogleUserOAuth2Manager * renamed google_oauth_utilities to google_user_oauth_utilities * updated the README with new parameters * removed unused json import * Update query format in quickstart-tutorial.mdx * fix: delete from table * fix url in huggingface_inference_api.mdx * test * Fix hardcoded API base, read it from conn args too * added a raise_for_status() check when downloading creds * raised a ValueError if parsing JSON creds fail * updated the CREATE ML_ENGINE examples in README * refactored the error message for when parsing fails * added a check for import errors during model creation * added comment to explain change * split req file paths to flag and path * Add pred_args * updated hostname of sample dbs * fixed slack chatbot tutorial * added default values for creds params in client * updated the order of params passed to client * clean up handler install log messages (#8720) * Docker build improvements (#8999) * build once and push cache later * Separate cache job * add db seed and mssql disconnect (#8742) * fixe the init for the EventStoreDB handler * removed unusued imports from the Rockset handler * removed unusued deps from the Rockset handler * fixed types in connection args * refactored req files to be installed from abs path * removed unnecessary call to set the success of result * updated Snowflake docs with tip to install git * fixed grammar in last sentence * added space between the two lines * updated email integration docs * fix: dn.query could not support projection * added a check for standalone comments * added a check for inline comments * make alembic use mindsdb log level (#9028) * Bump version (#9041) * Updated LightFM Integration Docs with Tip to Install Linux Dev Pakcages (#9042) * added a tip for installing the Linux dev packages * fixed a couple of grammatical errors * updated ollama docs (#8995) * updated ollama docs * updated note * added tip for docker * updates * updates --------- Co-authored-by: Max Stepanov <stpmax@yandex.ru> Co-authored-by: andrew <elkin.andr@gmail.com> Co-authored-by: Minura Punchihewa <minurapunchihewa17@gmail.com> Co-authored-by: QuantumPlumber <44450703+QuantumPlumber@users.noreply.github.com> Co-authored-by: dusvyat <13661123+dusvyat@users.noreply.github.com> Co-authored-by: Ty <124617566+tmichaeldb@users.noreply.github.com> Co-authored-by: Andrey <andrey@mindsdb.com> Co-authored-by: Vlad Romanenko <vlad.romanenko@hotmail.com> Co-authored-by: Arnav K <arnavkaushal09@gmail.com> Co-authored-by: Zoran Pandovski <zoran.pandovski@gmail.com> Co-authored-by: Martyna <martyna@mindsdb.com> Co-authored-by: Minura Punchihewa <49385643+MinuraPunchihewa@users.noreply.github.com> Co-authored-by: martyna-mindsdb <109554435+martyna-mindsdb@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Fixes RAG-2.
This PR overhauls
langchain_handler:user_column,assisstant_column) as possiblemodeargument as it was untested & unusedagent_typeto customize behavior (see agent types)Type of change
Verification Process
To ensure the changes are working as expected:
./tests/unit/ml_handlers/test_langchain.pyREADME.mdto create and query a model using the langchain handler.Additional Media:
Checklist: