Skip to content

Latest commit

 

History

History
240 lines (182 loc) · 18.2 KB

CHANGELOG.md

File metadata and controls

240 lines (182 loc) · 18.2 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

vNext

Added

  • Added DB.ExportToWriter() to allow users to pass any io.Writer implementation for the DB export, not just a file. This allows for example to export the DB to AWS S3 or compatible services (like Ceph, MinIO etc.). (PR #71)
  • Added DB.ImportFromReader() to allow users to pass any io.ReadSeeker implementation for the DB import, not just a file. This allows for example to import the DB from AWS S3 or compatible services (like Ceph, MinIO etc.). (PR #72)
  • Added example code for S3 export/import with the ⬆️ new methods (PR #73)
  • Added Azure OpenAI compatibility (PR #74 by @iwilltry42)

Fixed

  • The Collection.QueryEmbedding() call assumed/expected the query embedding from the parameter to be normalized already, but it wasn't documented and it's also inconvenient for users who use an embedding model/API that doesn't return normalized embeddings. Now we check whether the embedding is normalized and if it's not then we normalize it. (PR #77)
    • (Currently chromem-go only does cosine similarity, and document embeddings are already being normalized, so the query embedding has to be normalized as well. In the future we might offer other distance functions or allow to inject your own and make the normalization optional)

v0.6.0 (2024-04-25)

Highlights in this release are an extended interface, experimental WebAssembly bindings, and the option to use a custom Ollama URL. But also the fact that three people contributed to this release! Thank you so much! 🙇‍♂️

Added

  • Added Collection.Delete() to delete documents from a collection (PR #63 by @iwilltry42)
  • Added an experimental WebAssembly binding (package wasm) and example (PR #69)

Improved

  • Use prefixes for nomic-embed-text model in RAG-Wikipedia-Ollama example (PR #49, #65)
    • Thanks @rinor for pointing out the bug!
  • Made Ollama URL configurable (PR #64 by @erikdubbelboer)
  • Added building of code examples to CI (PR #66)
  • Improved RAG template (PR #67)

Breaking changes

  • NewEmbeddingFuncOllama now requires a second parameter for the base URL. But it can be empty to use the default which was also used in the past.

New Contributors

Full Changelog: https://github.com/philippgille/chromem-go/compare/v0.5.0...v0.6.0

v0.5.0 (2024-03-23)

Highlights in this release are query performance improvements (5x faster, 98% fewer memory allocations), export/import of the entire DB to/from a single file with optional gzip-compression and AES-GCM encryption, optional gzip-compression for the regular persistence, a new code example for semantic search across 5,000 arXiv papers, and an embedding func for Cohere.

Added

  • Added arXiv semantic search example (PR #45)
  • Added basic query benchmark (PR #46)
  • Added unit test for collection query errors (PR #51)
  • Added Collection.QueryEmbedding() method for when you already have the embedding of your query (PR #52)
  • Added export and import of the entire DB to/from a single file, with optional gzip-compression and AES-GCM encryption (PR #58)
  • Added optional gzip-compression to the regular persistence (i.e. the DB from NewPersistentDB() which writes a file for each added collection and document) (PR #59)
  • Added minimal example (PR #60, #62)
  • Added embedding func for Cohere (PR #61)

Improved

  • Changed the example link target to directory instead of main.go file (PR #43)
  • Improved query performance (5x faster, 98% fewer memory allocations) (PR #47, #53, #54)
    • benchstat output 
      goos: linux
goarch: amd64
pkg: github.com/philippgille/chromem-go
cpu: 11th Gen Intel(R) Core(TM) i5-1135G7 @ 2.40GHz
                                    │    before     │               after                 │
                                    │    sec/op     │    sec/op     vs base               │
Collection_Query_NoContent_100-8      413.69µ ±  4%   90.79µ ±  2%  -78.05% (p=0.002 n=6)
Collection_Query_NoContent_1000-8     2759.4µ ±  0%   518.8µ ±  1%  -81.20% (p=0.002 n=6)
Collection_Query_NoContent_5000-8     12.980m ±  1%   2.144m ±  1%  -83.49% (p=0.002 n=6)
Collection_Query_NoContent_25000-8    66.559m ±  1%   9.947m ±  2%  -85.06% (p=0.002 n=6)
Collection_Query_NoContent_100000-8   282.41m ±  3%   39.75m ±  1%  -85.92% (p=0.002 n=6)
Collection_Query_100-8                416.75µ ±  2%   90.99µ ±  1%  -78.17% (p=0.002 n=6)
Collection_Query_1000-8               2792.8µ ± 23%   595.2µ ± 13%  -78.69% (p=0.002 n=6)
Collection_Query_5000-8               15.643m ±  1%   2.556m ±  1%  -83.66% (p=0.002 n=6)
Collection_Query_25000-8               78.29m ±  1%   11.66m ±  1%  -85.11% (p=0.002 n=6)
Collection_Query_100000-8             338.54m ±  5%   39.70m ± 12%  -88.27% (p=0.002 n=6)
geomean                                12.97m         2.192m        -83.10%

                                    │      before      │               after                 │
                                    │       B/op       │     B/op      vs base               │
Collection_Query_NoContent_100-8       1211.007Ki ± 0%   5.030Ki ± 0%  -99.58% (p=0.002 n=6)
Collection_Query_NoContent_1000-8      12082.16Ki ± 0%   13.24Ki ± 0%  -99.89% (p=0.002 n=6)
Collection_Query_NoContent_5000-8      60394.23Ki ± 0%   45.99Ki ± 0%  -99.92% (p=0.002 n=6)
Collection_Query_NoContent_25000-8     301962.1Ki ± 0%   206.7Ki ± 0%  -99.93% (p=0.002 n=6)
Collection_Query_NoContent_100000-8   1207818.1Ki ± 0%   791.4Ki ± 0%  -99.93% (p=0.002 n=6)
Collection_Query_100-8                 1211.006Ki ± 0%   5.033Ki ± 0%  -99.58% (p=0.002 n=6)
Collection_Query_1000-8                12082.11Ki ± 0%   13.25Ki ± 0%  -99.89% (p=0.002 n=6)
Collection_Query_5000-8                60394.10Ki ± 0%   46.04Ki ± 0%  -99.92% (p=0.002 n=6)
Collection_Query_25000-8               301962.1Ki ± 0%   206.8Ki ± 0%  -99.93% (p=0.002 n=6)
Collection_Query_100000-8             1207818.1Ki ± 0%   791.4Ki ± 0%  -99.93% (p=0.002 n=6)
geomean                                   49.13Mi        54.97Ki       -99.89%

                                    │    before     │              after                │
                                    │   allocs/op   │ allocs/op   vs base               │
Collection_Query_NoContent_100-8        238.00 ± 0%   94.00 ± 1%  -60.50% (p=0.002 n=6)
Collection_Query_NoContent_1000-8       2038.5 ± 0%   140.5 ± 0%  -93.11% (p=0.002 n=6)
Collection_Query_NoContent_5000-8      10039.0 ± 0%   172.0 ± 1%  -98.29% (p=0.002 n=6)
Collection_Query_NoContent_25000-8     50038.0 ± 0%   204.0 ± 1%  -99.59% (p=0.002 n=6)
Collection_Query_NoContent_100000-8   200038.0 ± 0%   232.0 ± 3%  -99.88% (p=0.002 n=6)
Collection_Query_100-8                  238.00 ± 0%   94.50 ± 1%  -60.29% (p=0.002 n=6)
Collection_Query_1000-8                 2038.0 ± 0%   141.0 ± 1%  -93.08% (p=0.002 n=6)
Collection_Query_5000-8                10038.0 ± 0%   174.5 ± 2%  -98.26% (p=0.002 n=6)
Collection_Query_25000-8               50038.0 ± 0%   205.5 ± 2%  -99.59% (p=0.002 n=6)
Collection_Query_100000-8             200038.5 ± 0%   233.0 ± 1%  -99.88% (p=0.002 n=6)
geomean                                 8.661k        161.4       -98.14%
  • Extended parameter validation (PR #50, #51)
  • Simplified unit tests (PR #55)
  • Improve NewPersistentDB() path handling (PR #56)
  • Improve loading of persistent DB (PR #57)
  • Increased unit test coverage in various of the other listed PRs

Fixed

  • Fixed path joining (PR #44)

Breaking changes

  • Due to vectors now being normalized at the time of adding the document to the collection instead of when querying, the persisted data from prior versions is incompatible with this version (PR #47)

v0.4.0 (2024-03-06)

Highlights in this release are optional persistence, an extended interface, support for creating embeddings with Ollama, the exporting of the Document struct, and more Go-idiomatic methods to add documents to collections.

Added

  • Extended the interface:
    • DB.ListCollections() (PR #12)
    • DB.GetCollection() (PR #13 + #19)
    • DB.DeleteCollection() (PR #14)
    • DB.Reset() (PR #15)
    • DB.GetOrCreateCollection() (PR #22)
    • Collection.Count() (PR #27)
    • Document struct, NewDocument() function, Collection.AddDocument() and Collection.AddDocuments() methods (PR #34)
      • More Go-idiomatic alternatives to Collection.Add()
  • Added various unit tests (PR #20, #39)
  • Added optional persistence! Via multiple PRs:
  • Added support for creating embeddings with Ollama (PR #32)
  • Added example documentation (PR #42)

Improved

  • Improved example (PR #11, #28, #33)
  • Stop exporting Collection.Metadata (PR #16)
    • Goal: Prevent direct modifications which could cause data races in case of the user doing a modification while chromem-go for example ranges over it during a Collection.Query() call.
  • Copy metadata in constructors (PR #17)
    • Goal: Prevent direct modifications which could cause data races in case of the user doing a modification while chromem-go for example ranges over it.
  • Improved CI (PR #18)
    • Add Go 1.22 to test matrix, update used GitHub Action from v4 to v5, use race detector during tests
  • Reorganize code internally (PR #21)
  • Switched to newer recommended check for file related ErrNotExist errors (PR #29)
  • Added more validations in several existing methods (PR #30)
  • Internal variable renamed (PR #37)
  • Fail unit tests immediately (PR #40)

Fixed

  • Fixed metadatas validation in Collection.AddConcurrently() (PR #35)
  • Fixed Godoc of Collection.Query() method (PR #36)
  • Fixed length of result slice (PR #38)
  • Fixed filter test (PR #41)

Breaking changes

  • Because functions can't be (de-)serialized, GetCollection requires a new parameter of type EmbeddingFunc, in order to set the correct func when using a DB with persistence and it just loaded the collections and documents from storage. (PR #25)
  • Some methods now return an error (due to file operations when persistence is used)
  • Additional validations will return an early error, but most (if not all) prior calls with the invalid parameters probably lead to some errors down the line anyway
  • Collection.Metadata is not exported anymore
  • Result.Document field was renamed to Result.Content, to avoid confusion with the now exported Document struct

v0.3.0 (2024-02-10)

Added

Improved

  • Improve concurrency when adding documents to collection (PR #2)
  • Rename Client to DB to better indicate that the database is embedded and there's no client-server separation (PR #3)
  • Change OpenAPI embedding model from "text-embedding-ada-002" to "text-embedding-3-small" (PR #4)
  • Allow custom base URL for OpenAI, enabling the use of Azure OpenAI, LiteLLM, ollama etc. (PR #7)
  • Renamed EmbeddingFunc constructors to follow best practice (PR #9)

Fixed

  • Don't allow nResults arg < 0 (PR #5)

Breaking changes

  • Several function names and signatures were changed in this release. This can happen as long as the version is at v0.x.y.

v0.2.0 (2024-01-01)

Added

  • Added GitHub Actions config (commit)
  • Added CHANGELOG.md (commit)
  • Exported embedding creation functions (commit)
  • Added Collection.AddConcurrently to add embeddings concurrently (commit)

Improved

  • Improved example code (commit)
  • Removed unused field in Client (commit)
  • Improved validation in Query method (commit)
  • Added and improved Godoc (commit)
  • Improved locking around a collection's documents (commit)
  • Removed dependency on third party library for OpenAI (commit)
  • Parallelized document querying (PR #1)

v0.1.0 (2023-12-29)

Initial release with a minimal Chroma-like interface and a working retrieval augmented generation (RAG) example.