agentjido / req_llm / 055cd74e7ae9d7c1ae23daa014bf11f56f36241b

49%

Build Type

push

github

Committed by web-flow

Commit Message

Refactor guides and update mix.exs for improved documentation and task organization (#164)

* Enhance CHANGELOG with new features, fixes, and improvements

- Added support for prompt caching and structured output for Bedrock Anthropic models.
- Introduced Google Search grounding for Gemini models and improved JSON Schema validation.
- Enhanced Bedrock provider with comprehensive fixes, including tool round-trip conversations and usage metrics.
- Refactored Meta/Llama support into a reusable generic provider.
- Updated documentation for clarity and improved organization, including new provider-specific guides.
- Fixed various bugs related to streaming, tool handling, and model compatibility.

This update significantly improves the functionality and usability of the ReqLLM library.

* Update documentation and examples to reflect model changes and improve clarity

- Changed primary model reference from `anthropic:claude-3-sonnet-20240229` to `anthropic:claude-haiku-4-5` across various files, including README, usage rules, and guides.
- Updated the main documentation entry point from "getting-started" to "overview" for better navigation.
- Enhanced the README to clarify the two-layer architecture and supported providers, including new examples and usage scenarios.
- Organized documentation structure to improve accessibility and understanding of the API features.

These updates ensure that users have the most accurate and helpful information for utilizing the ReqLLM library effectively.

* Update mix.exs and gen.ex to enhance task commands and documentation

- Added new Mix tasks for model synchronization and generation, improving command accessibility.
- Updated documentation for the `req_llm.gen` task to clarify default streaming behavior and provide clearer examples.
- Adjusted command options to include `--no-stream` for non-streaming mode, enhancing user control over output format.

These changes improve the usability and clarity of the ReqLLM library... (continued)

Run Details

0 of 4 new or added lines in 1 file covered. (0.0%)

8 existing lines in 5 files now uncovered.

3829 of 7092 relevant lines covered (53.99%)

460.63 hits per line