feat(media): broaden content type support #1011
Conversation
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Expanded media content type support and improved code quality by enabling linting for all Python files in the langfuse/api directory.
- Added extensive media content type support in
langfuse/api/resources/media/types/media_content_type.pyincluding common image, audio, video, text and application formats - Removed linting exclusion for
langfuse/api/**/*.pyinci.ruff.tomlto enforce consistent code quality standards - Updated import paths across documentation and examples to use
langfuse.api.clientinstead offinto.clientfor better consistency
1 file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
The PR continues to improve code formatting across the codebase by breaking long lines into multiple lines and adding proper indentation for better readability.
- Reformatted dictionary definitions and function parameters in all resource files under
langfuse/api/resources/for improved readability - Split long import statements into multiple lines in various
__init__.pyfiles - Maintained consistent indentation and line breaks in error handling blocks across client files
- Added trailing commas to function parameters in
langfuse/api/core/client_wrapper.py
110 file(s) reviewed, 16 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Added support for Anthropic and Vertex media formats in the media manager while updating dependency constraints for better compatibility.
- Added handling for Anthropic media format with
type: base64andmedia_typefields inmedia_manager.py - Added support for Vertex media format with
type: mediaandmime_typefields inmedia_manager.py - Updated langchain-google-vertexai dependency to
>=2.0.0,<3.0.0inpyproject.toml - Added upper bounds to langchain packages (e.g.
<0.3,<0.4) inpyproject.tomlfor stability
2 file(s) reviewed, 1 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the provided information, I'll summarize the most recent changes that haven't been covered in previous reviews:
No major changes found since last review. The previous reviews have comprehensively covered:
- Media content type expansion
- Code formatting improvements
- Import path standardization
- Anthropic and Vertex media format support
- Dependency updates
All recent changes appear to be part of these already documented updates.
112 file(s) reviewed, 5 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Updated Python version support and import paths across the codebase while maintaining backward compatibility.
- Dropped Python 3.8 support and added Python 3.12 in
.github/workflows/ci.yml - Relaxed langchain-google-vertexai dependency from
>=2.0.0,<3.0.0to>=1.0.0,<3.0.0inpyproject.toml - Updated minimum Python version from 3.8.1 to 3.9 in
pyproject.toml - Excluded
langfuse/api/**/*.pyfrom linting inci.ruff.toml
19 file(s) reviewed, 2 comment(s)
Edit PR Review Bot Settings | Greptile
![ellipsis-dev[bot]](https://avatars.githubusercontent.com/in/64358?s=60&v=4){kind=small}
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes shown in the files and previous reviews, here's a summary of the latest updates:
Enhanced model extraction functionality and updated test coverage for new LangChain integrations.
- Added support for BedrockLLM and OllamaLLM model types in
langfuse/extract_model.py - Updated model name extraction logic for AzureOpenAI to check
invocation_params.modelbeforemodel_name - Moved langchain integration packages (ollama, cohere, huggingface) from dev to main dependencies in
pyproject.toml - Updated test cases in
test_extract_model.pyto use latest LangChain package structure and model configurations
These changes focus on improving model support and test coverage, which weren't covered in previous reviews that focused on media types, formatting, and Python version updates.
3 file(s) reviewed, 1 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes shown in the files and previous reviews, here's a summary of the latest updates:
Removed AzureOpenAI test support while maintaining AzureChatOpenAI functionality in test suite.
- Removed AzureOpenAI test case and import from
tests/test_extract_model.pywhile keeping AzureChatOpenAI tests - Maintained comprehensive test coverage for other model types (ChatGroq, Ollama, MistralAI, etc.)
- Test changes align with recent model extraction logic updates for AzureOpenAI
The changes focus specifically on test suite updates that weren't covered in previous reviews about media types, formatting, Python versions and model extraction logic.
1 file(s) reviewed, 1 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes, I'll provide a summary focusing only on the most recent updates not covered in previous reviews:
Updated dependency management and test configurations in pyproject.toml for improved package organization.
- Moved langchain integration packages (ollama, cohere, huggingface) back to dev dependencies in
pyproject.toml - Added
requestsandanyioas main dependencies inpyproject.toml - Updated
langchain-google-vertexaiversion constraint to>=1.0.0,<3.0.0for better compatibility - Added version constraints for new dependencies:
idna ^3.7andanyio ^4.4.0
The changes focus on dependency organization and version constraints that weren't covered in previous reviews about media types, formatting, Python versions and testing updates.
1 file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the provided information and previous reviews, I'll summarize only the most recent changes not covered in previous reviews:
Updated test configurations and added new media content types for broader format support.
- Added new media content types in
media_content_type.pyincludingaudio/oga,audio/aac,audio/mp4,audio/flac - Added support for additional document formats:
application/msword,application/vnd.ms-excel - Added
application/octet-streamfor generic binary data support - Added
text/csvandtext/cssfor additional text format support
The changes focus on expanding media format support that wasn't fully covered in previous reviews about dependency management, testing updates, and Python version changes.
No file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes shown in the files and previous reviews, here's a summary of the latest updates:
Updated test suite to accommodate changes in langchain-openai's behavior and model parameter handling.
- Modified model name assertions in
test_extract_model_langchain_openai.pyto use substring checks instead of exact matches - Updated
max_tokenstomax_completion_tokensin test cases to match new ChatOpenAI parameter naming - Added
refusal: Nonefield in OpenAI function call test expectations intest_langchain.py - Updated HumanMessage serialization test in
test_json.pyto verify full message structure
These changes focus on test suite adaptations that weren't covered in previous reviews about media types, Python versions, and dependency management.
6 file(s) reviewed, 1 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes and previous reviews, I'll summarize only the most recent updates:
Updated CI configuration and test parameters to align with langchain-openai changes.
- Changed environment variable from
LANGFUSE_READ_FROM_CLICKHOUSE_ONLYtoLANGFUSE_RETURN_FROM_CLICKHOUSE_ONLYinci.yml - Added required 'model' parameter to AzureChatOpenAI configuration in
test_extract_model_langchain_openai.py - Maintained Python version matrix up to 3.11 in CI despite broader version support in package requirements
These changes focus on CI/testing updates that weren't covered in previous reviews about media types, dependency management, and test suite adaptations.
2 file(s) reviewed, 1 comment(s)
Edit PR Review Bot Settings | Greptile
There was a problem hiding this comment.
Disclaimer: Experimental PR review
PR Summary
(updates since last review)
Based on the recent changes and previous reviews, I'll summarize only the most recent updates:
Fixed environment variable naming in CI configuration for Clickhouse integration.
- Fixed environment variable from
LANGFUSE_RETURN_FROM_CLICKHOUSE_ONLYtoLANGFUSE_RETURN_FROM_CLICKHOUSEin.github/workflows/ci.ymlfor proper Clickhouse integration - Maintained Python version matrix (3.9-3.11) in CI configuration despite broader package support
- Removed commented out
SERPAPI_API_KEYfrom CI environment variables
The changes focus on CI configuration improvements that weren't covered in previous reviews about media types, testing updates, and dependency management.
1 file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile
Important
Expand
MediaContentTypesupport and update import paths in documentation and examples.MediaContentTypeinmedia_content_type.pyto include more image, audio, video, text, and application types.README.md,reference.md, andclient.pyto uselangfuse.api.clientinstead offinto.client.comments/client.py,dataset_items/client.py,datasets/client.py, and 10 other files.ci.yml.pyproject.tomlto>=3.9,<4.0.This description was created by
for 18bf194. It will automatically update as commits are pushed.