Module api

Structs§

anthropic_output_format
fallback_summary: Provide a deterministic fallback summary if model generation fails
format_types_description: Format commit types from config into a rich description for the prompt Order is preserved from config (first = highest priority)
generate_analysis_with_map_reduce: Generate conventional commit analysis, using map-reduce for large diffs
generate_conventional_analysis: Generate conventional commit analysis using OpenAI-compatible API
generate_fast_commit: Generate a complete commit in a single API call (fast mode).
generate_summary_from_analysis: Create commit summary using a smaller model focused on detail retention
get_client: Get (or create) the shared HTTP client with timeouts from config.
openai_prompt_cache_key: Generate a deterministic cache key for OpenAI prompt-prefix routing.
openai_response_format
retry_api_call: Retry an API call with exponential backoff
run_oneshot
strict_json_schema
strip_type_prefix: Strip conventional commit type prefix if LLM included it in summary.
summary_from_holistic_analysis: Build a commit summary from the holistic analysis response when present.
timed_send: Send an HTTP request with timing instrumentation.