v0.1.4

What's Changed

• Add GenerationInfo to GenerateContent by @eliben in #488
• Clean up GeneratePrompt -- unused code, and simplify caller by @eliben in #489
• fix unmarshalling of chatmessage by @ctrlaltdel121 in #491
• feature: add new gemini llm supporting new experimental interface by @mrothroc in #497
• Add Ollama test and remove custom prompt for chat by @eliben in #495
• ollama: Clean up Ollama options by @eliben in #496
• googleai: set up options and embeddings by @eliben in #498
• Extend Model / GenerateContent interface to support multi-content chats by @eliben in #500
• Add multi-turn tests for new GenerateContent capabilities by @eliben in #501
• ollama: implement Model interface with GenerateContent by @eliben in #502
• Test streaming support in GenerateContent in openai and ollama by @eliben in #503
• googleai: implement streaming and some generation options by @eliben in #504
• Add function invocation into GenerateContent responses by @eliben in #512
• googleai: fix streaming termination by @eliben in #515
• Add callback handler method for GenerateContent start and end by @eliben in #514
• Finalize the majority of the model/LLM interface refactoring by @eliben in #521
• all: clean up now unused types Generation and LLMResult by @eliben in #523
• all: clean up and update comments by @eliben in #524
• docs: add example to embeddings by @haochunchang in #518
• Deprecate the LLM interface by @eliben in #525
• googleai: fix test comparisons to use case-insensitive regexps by @eliben in #528
• googleai: add more model parameters and update comments by @eliben in #529
• googleai: move downloadImageData to separate file by @eliben in #530
• Link to LangChain Python (deserves credit) by @carsonkahn-external in #509
• vectorstores: Weaviate make additional fields configurable by @jvgrootveld in #522
• ollama: add WithHttpClient option by @elnoro in #505
• googleai: propagate more options and add test by @eliben in #532
• llms: configure multiple response candidates for models that support them by @eliben in #533
• vectorstores: added pgvector example by @elnoro in #534
• weaviate: add option to deduplicate documents while adding by @corani in #527
• weaviate: honor vectorstores.WithEmbedder by @corani in #526
• vectorstores: pgvector add option WithConn by @Abirdcfly in #499
• memory: fix cutMessages in ConversationWindowBuffer by @Abirdcfly in #506
• textsplitter: render code blocks and reference links by @corani in #493
• templates: Add f-string based templating implementation by @chyroc in #479
• agents: add openai functions agent by @lujingxuansc in #459
• bug fix: openai_functions_agent not compatible with the new llms.Model interface by @devinyf in #536
• llms: rename convenience function to GenerateFromSinglePrompt by @eliben in #537
• googleai: test GenerateFromSinglePrompt convenience function by @eliben in #538
• docs: add pgvector page by @elnoro in #535
• googleai: add initial Vertex (GCP) implementation of Model by @eliben in #540
• googleai: move the PaLM provider into googleai by @eliben in #541
• embeddings: add cybertron local embeddings by @corani in #531
• googleai: add embeddings to vertex by @eliben in #546
• googleai: refactor to better separate generated code by @eliben in #547
• huggingface: allows to set custom inference endpoint by @mheers in #545
• cybertron: add example for cybertron embeddings by @corani in #543
• Move utility function by @eliben in #552
• googleai: add codegen tool to generate vertex.go from googleai.go by @eliben in #553
• googleai: clean up comments and order files with consistent naming by @eliben in #554
• llms: officially deprecate Call method on the interface by @eliben in #555
• llms: Add convenience helper for text parts by @tmc in #549
• feat: Added Qdrant vectorstore support by @Anush008 in #544
• googleai: rename constructors to New for consistency by @eliben in #570
• llms: convenience methods to create Parts by @eliben in #571
• embeddings: fix variable typo in huggingface_test by @ion232 in #569
• vectorstores: Wrapped the PGVector store init function in a single transaction by @pieterclaerhout in #563
• vectorstores: Added a unique constraint for the collection name when using PGVector by @pieterclaerhout in #559
• vectorstores: Added an option to PGVector to set the collection metadata by @pieterclaerhout in #557
• docs: add example and update conceptual guide URL for agents by @haochunchang in #539
• examples: add new vertex and googleai examples by @eliben in #572
• docs: qdrant-examples by @Anush008 in #568
• vectorstores: Implemented support for deduplicating documents when adding for PGVector by @pieterclaerhout in #566
• Update openai examples to new APIs by @eliben in #573
• examples: use llms.TextPart helper when it's useful by @eliben in #574
• bump examples dependency to v0.1.4-alpha.0 by @eliben in #575
• examples: update all to new llms interfaces by @eliben in #576
• vertex: add tests by @eliben in #582
• googleai: share tests between the two main providers by @eliben in #584
• googleai: add tests by @eliben in #586
• vectorstores: Added the missing comment on the exported WithCollectionMetadata function by @pieterclaerhout in #578
• vectorstores: Added an option to specify the vector dimensions for the embeddings by @pieterclaerhout in #579
• googleai: add test with BinaryPart image by @eliben in #587
• readme: update example with non-deprecated call by @eliben in #588
• vectorstores: Removed the custom ID option for embeddings when using PGVector by @pieterclaerhout in #590

New Contributors

• @ctrlaltdel121 made their first contribution in #491
• @mrothroc made their first contribution in #497
• @haochunchang made their first contribution in #518
• @carsonkahn-external made their first contribution in #509
• @jvgrootveld made their first contribution in #522
• @elnoro made their first contribution in #505
• @lujingxuansc made their first contribution in #459
• @devinyf made their first contribution in #536
• @mheers made their first contribution in #545
• @Anush008 made their first contribution in #544
• @ion232 made their first contribution in #569
• @pieterclaerhout made their first contribution in #563

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

• prompts: move prompts to embedded text files by @chyroc in #451
• vectorstores: Add ids to the return when adding documents in any vectorstore by @jaylalakiya in #448
• Skip tests that are currently failing (#415) by @eliben in #466
• Refactor LanguageModel interface away by @eliben in #468
• templates: Fix up gonja/jinja2 handling by @tmc in #469
• vectorstores: Add opensearch by @ThomasP1988 in #464
• Run go.mod tidy to make IDE happy by @eliben in #470
• vectorstores: Add azureaisearch by @ThomasP1988 in #462
• textsplitter: handle singular markdown tokens in markdown_splitter by @corani in #467
• Remove duplicated and unused LLMResult type from schema by @eliben in #472
• Remove unused schema types by @eliben in #473
• embeddings: Fix issue caused by new implementation of BatchTexts function by @jaylalakiya in #475
• GenerateContent - initial experimental API with OpenAI implementation by @eliben in #480
• Docs: fix jinja typo by @chyroc in #478
• Fix: Add error handle for Chat Generate by @NileZhou in #481
• support agent finish callback by @h3l in #476
• Clear out GetNumTokens by @eliben in #483
• Remove use of GetNumToken from example (followup on #483) by @eliben in #484
• chore: bump chroma-go to latest version by @bjwswang in #485

New Contributors

• @ThomasP1988 made their first contribution in #464
• @corani made their first contribution in #467
• @NileZhou made their first contribution in #481

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

• embeddings: Fix BatchTexts function by @jaylalakiya in #453
• vectorstores: fix chroma not return score from SimilaritySearch by @Abirdcfly in #449
• Feat: use chains callback as streaming func by @chyroc in #447
• fix: avoid panic when call openai.NewChat() and openai.New() by @h3l in #460
• support callback handler to llm chain by @h3l in #461
• add StopReason field in Generation struct. by @me020523 in #458
• feat: callbacks.Handler support handle error by @chyroc in #457
• vectorstores: fix weaviate not return score from SimilaritySearch by @Abirdcfly in #454
• Feat: support render prompt template with jinjia2 by @chyroc in #450
• memory: add ConversationWindowBuffer by @Abirdcfly in #411
• feat: support combine multi callbacks by @chyroc in #446

New Contributors

• @jaylalakiya made their first contribution in #453
• @h3l made their first contribution in #460
• @me020523 made their first contribution in #458

Full Changelog: v0.1.1...v0.1.2

v0.1.1

Updates dependencies, adds some pgvector improvements, test improvements.

What's Changed

• Clean up comments in chat_message_history by @eliben in #413
• test: fix api chain test error (#414) by @byebyebruce in #434
• pgvector: support postgres schema by @tecnologer in #439
• pgvector: Fix DropTable Function by @mehulmshah in #441
• fix: remove suffix slash from openai base url by @chyroc in #436
• feat: use WithCallback to set openai callback option by @chyroc in #438
• feat: read OPENAI_API_BASE env to compatible with langchain by @chyroc in #437
• Docs: fix comment typo by @chyroc in #442
• deps: Update x/crypto by @tmc in #444

New Contributors

• @tecnologer made their first contribution in #439
• @mehulmshah made their first contribution in #441
• @chyroc made their first contribution in #436

Full Changelog: v0.1.0...v0.1.1

v0.1.0

Our first versioned release!

What's Changed

• Add output parsers, prompt templates, memory, llmChain and example by @FluffyKebab in #2
• Move namespace in pinecone index from a struct field to function argument by @FluffyKebab in #4
• exp: Move recent contributions into exp subdir for now by @tmc in #14
• fix: rename test exp packages and fix all test compiling error by @nekomeowww in #17
• docs: added CONTRIBUTING.md and CODE_OF_CONDUCT.md by @nekomeowww in #16
• Add support for local LLMs by @rdbell in #22
• schema: Add schema package by @tmc in #23
• chore: improve code qualities and use idiomatic way to implementation by @nekomeowww in #27
• docs: added Pull Request template and improved CONTRIBUTING.md by @nekomeowww in #26
• schema: add more upstream schema concepts by @tmc in #32
• exp/prompts: move exp/prompts closer to acceptable state by @tmc in #33
• chains: add support for stop words by @Abraxas-365 in #30
• llm type check fix by @anignx in #36
• *: fix space in .github folder by @Fraser-Isbester in #40
• ci: Add linting, building, and testing to ci by @Fraser-Isbester in #42
• exp: change names of files and start using actual schema by @FluffyKebab in #44
• exp/agents: flesh out agents implementation by @Abraxas-365 in #38
• lint: add more sophisticated lint rules by @tmc in #47
• lint: adjust the line limit of funlen linter to 90 by @nekomeowww in #49
• exp/prompts: fix missing return and unexpected token by @nekomeowww in #48
• lint: disable testpackage by @nekomeowww in #51
• test: introduce stretchr/testify package for better test development experience by @nekomeowww in #28
• exp/memory: add multiple fixes by @FluffyKebab in #53
• prompts: introduced PromptTemplate along with the relevant types by @nekomeowww in #34
• prompts: simplify prompts package by @tmc in #57
• exp: adopt prompts by @tmc in #58
• memory: add version outside of exp by @FluffyKebab in #60
• ci: trigger lint-build-test GitHub Actions within PRs by @nekomeowww in #54
• output parsers: add structured and list by @FluffyKebab in #55
• OpenAI model selection and Chat completions by @rdbell in #62
• context: add context throughout by @tmc in #66
• memory: remove exp version by @FluffyKebab in #67
• Make the MRKL agent functional by @rdbell in #69
• textsplitter: add version outside of exp by @FluffyKebab in #63
• exp: adopt new outputparser and textsplitter by @FluffyKebab in #71
• docs: add docs folder and parity matrix md file by @O-Roma in #73
• embeddings: refactor and move out of exp by @FluffyKebab in #75
• documentloader: add interface and text loader by @FluffyKebab in #72
• llms: Options refactor by @tmc in #80
• Update parity_matrix.md with document loaders by @O-Roma in #82
• openaillm: new openaillm support options by @yubing744 in #78
• typo: link directed to 404 page, updated typo by @aldarisbm in #81
• vectorstores: add interface and support for pinecone by @FluffyKebab in #76
• chain: add standard interface, run, call and llm by @FluffyKebab in #70
• chains: Fix call and improve errors by @tmc in #86
• llms: Enhance chat completion capabilities in openai implementation by @tmc in #85
• exp: remove chains, vectorstores, documentloaders and examples by @FluffyKebab in #84
• chains: add retrieval QA, stuff documents and prompt selector by @FluffyKebab in #83
• pinecone: fix panic in doRequest by @tarunKoyalwar in #88
• chains: Add LLMMathChain by @tmc in #89
• vectorstore: add optional namespace paramater by @FluffyKebab in #87
• vectorstores: add options to ToRetriever by @FluffyKebab in #91
• Feat(openaiclient): add Option for custom HTTP client by @obitech in #92
• agents: refactor to better match upstream and increase consistency by @FluffyKebab in #95
• chains: add refine documents chain and summarization by @FluffyKebab in #96
• Feat: Parse OpenAI status code, error message by @obitech in #93
• prompts: Add sprig templates by @tmc in #97
• llms: Add ChatLLM type by @tmc in #98
• Refactor(llms): simplify option handling by @obitech in #101
• Fix(pinecone): correct docstrings by @obitech in #100
• move agents and tools out of exp by @FluffyKebab in #99
• feat: add support for Google VertexAI based PaLM API by @alvarowolfx in #103
• pinecone similaritySearch: Include score threshold based filtering by @cduggn in #110
• Improving HuggingfaceLLM: Configurable RepoID option by @EvilFreelancer in #109
• llms: AddWithStreamingFunc by @tmc in #106
• agents: Fix the issue with the executor tool nil panic and align the prompt o… by @bdqfork in #105
• gpt2 used as default model of huggingface package by @EvilFreelancer in #111
• llm-openai-client: openai base url option by @swuecho in #117
• agents: refactor options by @FluffyKebab in #118
• feat: Options like WithSeed, WithTopK and so on by @EvilFreelancer in #115
• fix: sync go module to latest in openai-chat-example by @bjwswang in #122
• document loaders: add csv loader by @mvrilo in #127
• docs: add examples by @FluffyKebab in #126
• docs: added package level comments to codebase by @teezzan in #135
• vectorstores: add weaviate by @ryomak in #132
• agents: add conversational agent by @bdqfork in #133
• chains: add more call options by @FluffyKebab in #137
• prompts: add chat prompt template by @FluffyKebab in #136
• feat: LocalLLM improvments, llms.CallOptions as local-llm call arguments by @EvilFreelancer in #131
• pinecone restquery: Add filter logic to limit searches based on metadata by @cduggn in #129
• tools: add wikipedia by @FluffyKebab in #123
• agents: fix bug and add option for returning steps by @FluffyKebab in #143
• Docs: start documentation site by @FluffyKebab in #141
• vectorstores: Weaviate Add filter logic　 by @ryomak in #142
• chains: add api chain by @byebyebruce in #138
• llms: split LLMChat and LLM into different structs by @FluffyKebab in #128
• chains: use langauge model instead of LLM and add predict by @FluffyKebab in #148
• docs: add examples for Pinecone vectorstore operations: AddDocuments and SimilaritySearch by @cduggn in #147
• openai: Add all CallOptions to Chat and Completion Requests, add GenerationInfo to Chat response by @mcartagenah in #146
• CI: Add ci step and Makefile entry to ensure examples are healthy by @mvrilo in #151
• chains: add conversation chain and fix run by @FluffyKebab in #155
• agents: add memory to executor and add test fo...