* Update pyproject.toml to newer Chroma releaseVersion [0.3.25](https://github.com/chroma-core/chroma/releases/tag/0.3.25) of Chroma removes a lot of heavyweight dependencies, which should considerably reduce install sizes for downstream projects. This PR bumps the required Chroma version to take advantage.* Update poetry.lock---------Co-authored-by: Chelsea Voss <csvoss@openai.com>

* added es datastore* updated readmy and toml* updated readme* clean up code + implemented a few more methods* add tests + fix issues found* update documentation* update notebook* clean up notebook---------Co-authored-by: Sebastian Montero <smonteroparis@icloud.com>

)* Updating Azure Cognitive Search client library* Update Azure Cognitive Search library to 11.4.0b8

In setup.md in the LlamaIndex setup section - Changed GPTSimpleVectorIndex to GPTVectorStoreIndex to match the updated field in LlamaIndex. Fixes issueopenai#307.

* Adding mongo vCore as a datastore* updating the readme file

When dealing with Pinecone occurs sporadically situations where `metadata["text"]` contains a datetime object somehow.This raises an Internal Error since Pydantic is unable to create a DocumentChunkWithScore object.We force the typing to str to fix this.Example:```{'matches': [{'id': '68_1',              'metadata': {'document_id': '68',                           'text': datetime.date(2265, 12, 1)},              'score': 0.760368824,              'values': []},```

…ustom GPTs / function calling (openai#411)* Update to support new embeddings models, custom GPTs, and function calling* Update README.md* Update README.md* Update README.md

* docker compose file.* search example.* mongodb atlas datastore.* refactor, docstring and notebook cleaning.* docstring.* fix attributes names.* Functional tests.* Example adjustement.* setup.md* remove some useless comments.* wrong docker image.* Minor documentation fixes.* Update example.* refactor.* default as a default collection.* TODO resolved.* Refactor delete.* fix readme and setup.md* add warning when delete without criteria.* rename private function.* replace pymongo to motor and fix integration test.* Refactor code and adjust tests* wait for assert function.* Update docs/providers/mongodb_atlas/setup.mdCo-authored-by: Jib <Jibzade@gmail.com>* Update datastore/providers/mongodb_atlas_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* Increase oversampling factor to 10.* Update tests/datastore/providers/mongodb_atlas/test_mongodb_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* Update tests/datastore/providers/mongodb_atlas/test_mongodb_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* Update datastore/providers/mongodb_atlas_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* Init docstring.* Default parameters* Update datastore/providers/mongodb_atlas_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* refactor sample_embeddings.* Apply suggestions from code reviewCo-authored-by: Jib <Jibzade@gmail.com>* refactor delete.* Version added.* Update datastore/providers/mongodb_atlas_datastore.pyCo-authored-by: Jib <Jibzade@gmail.com>* Removed _atlas from folder name to keep it simple and self-consistent* Expanded setup.md* Fixed a couple typos in docstrings* Add optional EMBEDDING_DIMENSION to get_embedding* Fixed typo in kwarg* Extended setup.md* Edits to environment variable table* Added authentication token descriptions* Removed hardcoded vector size* Added semantic search example* Added instructions to integration tests* Cleanup* Removed pathname from example.* Override DataStore.upsert in  MongoDBAtlasDataStore to increase performance.* upsert now returns ids of chunks, which is what each datastore document is* Added full integration test* test_integration now uses FastAPI TestClient* Retries query until response contains number requested---------Co-authored-by: Emanuel Lupi <emanuel.lupi91@gmail.com>Co-authored-by: Jib <Jibzade@gmail.com>