‹ Reports
The Dispatch

OSS Report: langchain-ai/langchain

LangChain Development Focuses on Bug Fixes and Integration Enhancements

LangChain, an open-source framework for developing applications powered by large language models (LLMs), has seen active development with a focus on bug fixes and integration enhancements. The project simplifies building, deploying, and managing LLM applications.

Recent Activity

Recent issues and pull requests (PRs) highlight a focus on resolving bugs related to model integrations and improving documentation clarity. Notable issues include protobuf errors (#26745) and streamed response incompatibilities (#26719). PRs such as #26738 address JSON parsing issues, while releases like #26737 for version 0.3.5 indicate regular updates.

Development Team Activity

Of Note

  1. Frequent Releases: Regular version updates reflect active maintenance and responsiveness to user feedback.
  2. Collaborative Culture: Many co-authored commits indicate strong team collaboration.
  3. Documentation Focus: Significant efforts in updating documentation alongside code changes.
  4. Integration Testing: Emphasis on improving integration tests to maintain software quality.
  5. Tracing Enhancements: Improved tracing features aid in debugging complex applications.

The LangChain project is characterized by robust development practices, focusing on stability, user experience, and continuous improvement through community engagement and regular updates.

Quantified Reports

Quantify Issues



Recent GitHub Issues Activity

Timespan Opened Closed Comments Labeled Milestones
7 Days 56 13 60 3 1
30 Days 183 56 206 13 1
90 Days 379 183 658 23 1
All Time 7604 7011 - - -

Like all software activity quantification, these numbers are imperfect but sometimes useful. Comments, Labels, and Milestones refer to those issues opened in the timespan in question.

Quantify commits



Quantified Commit Activity Over 30 Days

Developer Avatar Branches PRs Commits Files Changes
Erick Friis 12 46/41/4 64 471 101267
Bagatur 19 36/33/1 144 379 94043
Eugene Yurtsev 13 24/17/6 70 396 19758
Christophe Bornet 4 6/2/3 13 237 7884
Nuno Campos 3 2/2/0 5 26 7443
ccurme 14 23/18/2 72 144 6423
Leonid Ganeline 5 0/0/0 14 166 2960
Yuki Watanabe 2 0/0/0 2 15 2480
Brian Sam-Bodden 1 0/0/0 1 4 1817
Richmond Alake 1 0/0/0 1 2 1595
Rajendra Kadam 2 1/1/0 3 8 1571
Jacob Lee 1 0/0/0 1 1 1465
Nithish Raghunandanan 1 0/0/0 1 10 1318
William FH 2 4/3/1 8 23 992
Scott Hurrey 2 0/0/0 2 10 755
Jorge Piedrahita Ortiz 2 0/0/0 3 7 702
ChengZi 2 0/0/0 2 4 665
Sharmistha S. Gupta 1 0/0/0 1 4 572
moyidang 1 0/0/0 1 4 556
Hyman 1 0/0/0 1 2 431
Yash Parmar 1 0/0/0 1 7 407
ZhangShenao 4 2/2/0 8 12 372
Yuwen Hu 1 0/0/0 1 4 341
Aditya 1 0/0/0 1 1 283
Isaac Francisco 3 2/1/0 5 5 275
Guangdong Liu 1 0/0/0 1 3 196
Tomaz Bratanic 1 2/2/0 5 3 189
Harutaka Kawamura 1 0/0/0 1 3 174
John 2 0/0/0 2 4 156
Bruno Alvisio 1 0/0/0 1 3 143
miri-bar 1 0/0/0 1 1 138
Harrison Chase 1 0/0/0 1 4 137
JonZeolla 1 0/0/0 1 19 126
langchain-infra 1 0/0/0 1 2 109
RUO 1 0/0/0 1 2 106
Erik Lindgren 1 0/0/0 1 2 103
Ohad Eytan 1 0/0/0 1 3 101
Parsa Abbasi 1 0/0/0 1 1 88
Alex Sherstinsky 1 0/0/0 1 3 86
maang-h 1 0/0/0 1 1 68
Sebastian Cherny 1 0/0/0 1 1 68
xander-art 1 0/0/0 1 3 65
Hasan Kumar 1 0/0/0 1 3 64
Tibor Reiss 1 2/1/0 2 3 62
Martin Triska 1 0/0/0 1 2 59
Erik 1 0/0/0 1 1 55
Allan Ascencio 1 0/0/0 1 2 55
liuhetian 1 0/0/0 1 2 43
CastaChick 1 0/0/0 1 3 43
Mohammad Mohtashim (keenborder786) 1 1/0/0 1 1 43
Friso H. Kingma 1 0/0/0 1 2 38
Miguel Grinberg 1 0/0/0 1 1 35
Alexander KIRILOV 1 0/0/0 1 2 32
Anton Dubovik 1 0/0/0 1 2 31
Eric Pinzur 1 0/0/0 1 1 29
Vadym Barda 1 0/0/0 2 2 28
Lucain 1 0/0/0 1 1 28
ℍ𝕠𝕝𝕝𝕠𝕨 π•„π•’π•Ÿ 1 0/0/0 1 1 27
sslee 1 0/0/0 1 3 26
Djordje 2 0/0/0 2 2 23
Qingchuan Hao 1 0/0/0 1 1 23
Alejandro RodrΓ­guez 1 0/0/0 1 2 23
Kyle Winkelman 1 0/0/0 2 3 23
Austin Burdette 1 0/0/0 1 2 22
Daniel Cooke 1 0/0/0 1 2 22
clement.l 1 0/0/0 1 1 22
Ivan 1 0/0/0 1 2 22
Krishna Kulkarni 2 0/0/0 2 2 21
Anush 1 0/0/0 1 2 21
Geoffrey HARRAZI 1 0/0/0 1 1 21
Matthew DeGenaro 1 0/0/0 1 2 21
Jakub W. 1 0/0/0 1 1 20
Swastik-Swarup-Dash 1 0/0/0 1 1 20
Ian 1 0/0/0 1 2 19
Alejandro Companioni 1 0/0/0 1 2 18
Param Singh 1 0/0/0 1 2 18
venkatram-dev 2 1/1/0 3 3 18
Noah Mayerhofer 1 0/0/0 1 1 17
Dylan 1 0/0/0 1 1 16
Ashvin (ashvin-a) 1 1/0/1 1 2 13
ogawa 1 0/0/0 1 1 13
k.muto 1 0/0/0 1 1 12
gbaian10 1 2/1/1 1 1 12
Michael Paciullo 1 0/0/0 1 1 11
Yoshitaka Fujii 1 1/1/0 1 1 10
James Espichan Vilca 1 0/0/0 2 2 9
Jinoos Lee 1 0/0/0 1 1 8
Eric 1 0/0/0 1 1 8
Emmanuel Leroy 1 0/0/0 1 1 8
Nobuhiko Otoba 1 0/0/0 1 2 6
jessicaou 1 1/1/0 2 2 6
rainsubtime 1 0/0/0 1 1 6
Emad Rad 1 0/0/0 1 1 6
Piyush Jain 1 2/2/0 1 1 5
Borahm Lee 1 0/0/0 1 2 5
mehdiosa 1 0/0/0 1 1 5
Brace Sproul 1 0/0/0 1 1 5
Serena Ruan 1 2/1/0 1 1 5
yahya-mouman 1 0/0/0 1 1 5
Tom Daniel Grande 1 0/0/0 1 1 4
Sanjay Parajuli 1 0/0/0 1 1 4
Matt Hull 1 0/0/0 1 1 4
Nicolas 1 1/1/0 1 2 4
mschoenb97IL 1 0/0/0 1 1 4
Marcelo Machado 1 0/0/0 1 1 4
Sam Jove 1 0/0/0 1 1 3
Maurits Bos 1 0/0/0 1 1 3
Nada Amin 1 0/0/0 1 1 2
ice yao 1 1/1/0 1 1 2
Zhanwei Zhang 1 1/1/0 1 1 2
Luiz F. G. dos Santos 1 0/0/0 1 1 2
conjuncts 1 0/0/0 1 1 2
Ikko Eltociear Ashimine 1 0/0/0 1 1 2
sanjay920 1 1/1/0 1 1 2
stein1988 1 1/1/0 1 1 2
Tyler Wray 1 0/0/0 1 1 2
Yuxi Zheng 1 0/0/0 1 1 2
이규민 1 0/0/0 1 1 2
SQpgducray 1 0/0/0 1 1 2
Patrick McGleenon 1 1/1/0 1 1 2
Sandeep Bhandari 1 0/0/0 1 1 2
wlleiiwang 1 0/0/0 1 1 2
Qasim Khan 1 0/0/0 1 1 2
Malik Ashar Khan 1 0/0/0 1 1 2
basirsedighi 1 0/0/0 1 1 2
guoqiang0401 1 1/1/0 1 1 2
Marco Rossi IT 1 1/1/0 1 1 2
Bhadresh Savani 1 1/1/0 1 1 2
Kirushikesh DB 1 0/0/0 1 1 1
Kevin Engelke 1 0/0/0 1 1 1
None (acho98) 0 1/0/0 0 0 0
Gabriel Altay (galtay) 0 1/0/0 0 0 0
Gor Hayrapetyan (gorros) 0 1/0/0 0 0 0
Enzo L F (lfenzo) 0 1/0/0 0 0 0
None (iboates) 0 1/0/0 0 0 0
Omri Eliyahu Levy (omriel1) 0 1/0/0 0 0 0
Hyejun An (jagaldol) 0 1/0/0 0 0 0
Paulo Roberto de Oliveira Castro (prcastro) 0 1/0/1 0 0 0
Jambo (JamboChen) 0 1/0/0 0 0 0
Tari Yekorogha (kingtroga) 0 1/0/0 0 0 0
SJ (sjjpo2002) 0 1/0/0 0 0 0
Siddharth Murching (smurching) 0 2/0/2 0 0 0
KoKo Mexcelsa (Ko-Ko-Kirk) 0 1/0/0 0 0 0
None (Sheepsta300) 0 1/0/0 0 0 0
V.Prasanna kumar (VpkPrasanna) 0 1/0/0 0 0 0
None (ZeevRispler) 0 1/0/0 0 0 0
Aksel Joonas Reedi (akseljoonas) 0 1/0/0 0 0 0
Anusha Kolan (anushakolan) 0 1/0/1 0 0 0
Arun Prakash (arunprakash) 0 1/0/0 0 0 0
Mark Perfect (markperfect) 0 1/0/1 0 0 0
None (FouadLbakali) 0 1/0/1 0 0 0
None (ZaraP-NSTARX) 0 1/0/0 0 0 0
Ivan (ivanmilevtues) 0 1/0/0 0 0 0
AndrΓ© Quintino (quintinoandre) 0 1/0/0 0 0 0
None (atanas-balevsky) 0 1/0/1 0 0 0
Rafael Miller (rafaelsideguide) 0 1/0/0 0 0 0
Rakesh Mehta (rakeshmehta0308) 0 1/0/0 0 0 0
Samuel Fernando (samuelfernandof) 0 1/0/1 0 0 0
Nicole Moiseyev (sfc-gh-nmoiseyev) 0 1/0/0 0 0 0

PRs: created by that dev and opened/merged/closed-unmerged during the period

Detailed Reports

Report On: Fetch issues



Recent Activity Analysis

The LangChain GitHub repository has seen significant recent activity, with a total of 593 open issues. Among these, several issues highlight critical bugs and feature requests, particularly related to integration with various models and tools. A recurring theme is the inconsistency in handling parameters across different components, particularly in relation to structured outputs and tool calls.

Notably, there are multiple reports about issues with specific models (e.g., ChatOllama, ChatOpenAI) not behaving as expected when certain parameters are set. Additionally, users have expressed concerns regarding the documentation's clarity on using tools and structured outputs, indicating a need for improved examples and explanations.

Issue Details

Recent Issues

  1. Issue #26745: Protobuf errors when using langchain-chroma with protobuf >= 4

    • Priority: πŸ€–:bug
    • Status: Open
    • Created: 0 days ago

  2. Issue #26734: Templates on templates.langchain.com don't work with the provided instructions

    • Priority: πŸ€–:bug
    • Status: Open
    • Created: 1 day ago

  3. Issue #26719: Streamed responses incompatible with multiple choices (n>1)

    • Priority: πŸ€–:bug
    • Status: Open
    • Created: 1 day ago

  4. Issue #26710: MHTMLLoader assumes files part encoding is utf-8 but Microsoft Word saves utf-16 parts in MHT files.

    • Priority: πŸ€–:bug
    • Status: Open
    • Created: 1 day ago

  5. Issue #26701: Anthropic's prompt caching in langchain does not work with ChatPromptTemplate.

    • Priority: πŸ€–:docs
    • Status: Open
    • Created: 2 days ago

  6. Issue #26697: Exactly one of file, filename and url must be specified.

    • Priority: πŸ€–:bug
    • Status: Open
    • Created: 2 days ago

  7. Issue #26685: kwargs not included in Runnable.invoke signature (yet included everywhere else)

    • Priority: πŸ€–:bug, β±­: core
    • Status: Open
    • Created: 2 days ago

  8. Issue #26680: AzureMLOnlineEndpoint (Serverless deployment) request body format is totally wrong.

    • Priority: investigate
    • Status: Open
    • Created: 2 days ago

  9. Issue #26664: pip Installation of only LangChain failing due to Hash mismatch (other packages are installing no issues)

    • Priority: investigate
    • Status: Open
    • Created: 2 days ago

  10. Issue #26659: Different behaviour of ChatBedrock for different models.

    • Priority: investigate, β±­: core
    • Status: Open
    • Created: 2 days ago

Important Issues Overview

  • Many recent issues relate to bugs in the integration of various models and tools, particularly concerning structured output and tool calling.
  • There is a noticeable lack of clarity in documentation regarding how to properly utilize these features, leading to user confusion and frustration.
  • The community has raised concerns about inconsistencies in how parameters are handled across different components, especially when switching between models or using specific features like caching or embeddings.

This analysis indicates that while the project is actively developed and maintained, there are significant areas for improvement in both functionality and documentation clarity to enhance user experience and reduce confusion among developers utilizing LangChain's capabilities.

Report On: Fetch pull requests



Overview

The analysis of the provided pull requests (PRs) for the LangChain project reveals a vibrant and active development environment. The PRs cover a wide range of enhancements, bug fixes, and new features across various components of the LangChain ecosystem. Notably, there is a strong focus on improving integrations with external services, enhancing existing functionalities, and expanding the framework's capabilities through community contributions.

Summary of Pull Requests

  1. PR #26738: Addresses a JSON parsing issue in langchain_core/output_parser/json.py, ensuring that escape characters in dictionaries do not throw exceptions.
  2. PR #26737: A release PR for version 0.3.5, indicating regular updates and maintenance.
  3. PR #26736: Enhances tracing functionality by propagating cancellation reasons to inner tasks, improving error handling and debugging capabilities.
  4. PR #26735: Introduces project name propagation in tracer runs, aiding in better tracking and management of different projects within the LangChain framework.
  5. PR #26729: Another release PR, this time for version 0.3.4, showcasing the project's active development and timely updates.

Analysis of Pull Requests

Themes and Commonalities

  • Active Maintenance and Updates: The frequent release PRs (e.g., #26737 for 0.3.5 and #26729 for 0.3.4) indicate a well-maintained project with regular updates, which is crucial for security, performance improvements, and new features.
  • Community Contributions: Many PRs are contributions from the community (e.g., #26738 by Ashvin), highlighting an engaged user base that actively participates in enhancing the project.
  • Focus on Enhancements and Bug Fixes: The PRs address both new features (like improved tracing in #26736) and bug fixes (such as JSON parsing issues in #26738), demonstrating a balanced approach to development that prioritizes both stability and growth.

Notable Features Being Worked On

  • Tracing Enhancements: Improvements in tracing (#26736) suggest an emphasis on better debugging tools within LangChain, which can significantly aid developers in building and maintaining complex applications.
  • Regular Version Releases: The consistent versioning (e.g., releases 0.3.5 and 0.3.4) reflects an organized release management process, ensuring that users have access to the latest features and fixes.

Anomalies

  • There are no significant anomalies observed in the provided PRs. The development process appears smooth with regular contributions, timely releases, and a focus on both new features and bug fixes.

Conclusion

The LangChain project exhibits a healthy development lifecycle characterized by active maintenance, community involvement, and a balanced focus on enhancements and stability. The regular updates and improvements in tracing functionalities are particularly noteworthy as they enhance the framework's usability for developers building complex applications powered by large language models (LLMs).

Report On: Fetch commits



Repo Commits Analysis

Development Team and Recent Activity

Team Members and Their Recent Activities

  1. Erick Friis (efriis)

    • Released versions 0.3.5, 0.3.4, and 0.3.3.
    • Collaborated on multiple PRs involving bug fixes and enhancements across various modules.
    • Active in discussions related to versioning and dependency management.

  2. William FH (hinthornw)

    • Omitted chunk in LangChainTracer.
    • Added project name to run from tracer.
    • Worked on avoiding copying runs and unifying run trees.

  3. Anton Dubovik (adubovik)

    • Fixed chunk size handling in OpenAI embeddings.
    • Co-authored with Erick Friis on the embedding fix.

  4. Nithish Raghunandanan (nithishr)

    • Added TTL support to Couchbase caches and chat message history.
    • Collaborated with Erick Friis on documentation updates.

  5. Krishna Kulkarni (kk9393)

    • Refined skip count calculation for MongoDB chat message histories.

  6. Tibor Reiss (tibor-reiss)

    • Fixed text splitter issues and added tests for it.

  7. Eugene Yurtsev (eyurtsev)

    • Made several contributions including fixing import issues, updating dependencies, and releasing versions.
    • Active in deprecating old memory abstractions.

  8. Chester Curme (chester.curme)

    • Worked on various integration tests and enhancements across different modules.
    • Contributed to fixing bugs related to AzureSearch vectorstore.

  9. Bagatur (baskaryan)

    • Involved in multiple patches, including support for JS chat model namespaces and releasing versions.
    • Active in enhancing integration tests for various components.

  10. Ccurme (ccurme)

    • Focused on adding tests, improving documentation, and fixing bugs across different integrations.
    • Contributed significantly to the community's testing framework.

  11. Lucain (Wauplin)

    • Fixed issues related to HuggingFace endpoint initialization.

  12. Others: Various contributors worked on minor fixes, documentation updates, and integration tests across the repository.

Patterns and Themes

  • Frequent Releases: The team is actively releasing new versions of the software, indicating a focus on continuous improvement and responsiveness to user feedback.
  • Collaborative Efforts: Many commits are co-authored, showcasing a collaborative culture among team members.
  • Focus on Bug Fixes: A significant portion of recent activities involves addressing bugs and enhancing existing features, particularly around integrations with external services like OpenAI, HuggingFace, and MongoDB.
  • Documentation Improvements: There is a strong emphasis on updating documentation alongside code changes, ensuring that users have access to accurate information about new features and fixes.
  • Integration Testing: The team is investing time in improving integration tests across various components, which is crucial for maintaining software quality as the project evolves.

Conclusions

The development team is highly active with a clear focus on enhancing the LangChain framework through regular releases, collaborative efforts, and comprehensive testing practices. The commitment to improving documentation alongside code changes reflects an understanding of user needs and a dedication to maintaining software quality.