Chroma MCP Server

# Validation Evidence Script **DEPRECATION NOTICE:** The `validate_evidence.sh` shell script is deprecated and will be removed in version 0.3.0. Please use the `chroma-mcp-client validate-evidence` console script installed via the Python package. This script simplifies the process of validating evidence for learning promotions. ## Overview Validation evidence provides objective metrics to qualify the impact and quality of code changes. The `validate_evidence.sh` script wraps the `chroma-mcp-client validate-evidence` command, making it easy to: - Load validation evidence from files or IDs - Calculate validation scores based on multiple evidence types - Check if evidence meets promotion threshold - Save validation reports to files ## Usage **Recommended**: Use the console script directly: ```bash chroma-mcp-client validate-evidence --file evidence.json ``` **Legacy wrapper script (deprecated)**: ```bash ./scripts/validate_evidence.sh --file evidence.json ``` ### Required Parameters At least one of these evidence sources must be provided: - `-f, --file`: Path to a JSON file containing validation evidence - `-t, --test-ids`: Comma-separated list of test transition evidence IDs - `-r, --runtime-ids`: Comma-separated list of runtime error evidence IDs - `-q, --quality-ids`: Comma-separated list of code quality evidence IDs ### Optional Parameters - `--threshold`: Validation score threshold for promotion eligibility (default: 0.7) - `-o, --output`: Path to save the validation results as JSON - `-v, --verbose`: Enable verbose output ## Examples ### Validating evidence from a file ```bash ./scripts/validate_evidence.sh --file evidence.json ``` ### Validating evidence from multiple sources ```bash ./scripts/validate_evidence.sh \ --test-ids test-123,test-456 \ --runtime-ids error-789 \ --quality-ids quality-012 ``` ### Setting a custom threshold and saving results ```bash ./scripts/validate_evidence.sh \ --file evidence.json \ --threshold 0.8 \ --output validation-report.json ``` ## Integration with Promotion Workflow This script plays a critical role in the promotion workflow: 1. Collect evidence using the various evidence logging scripts 2. Use this script to validate the collected evidence 3. If the evidence meets the threshold, use the validation report with the `promote-learning` command The validation score is calculated based on: - Test transitions (failing → passing) - Runtime error resolutions - Code quality improvements Each evidence type contributes to the overall score, with weights defined by the validation system. ## See Also - [log-error](log-error.md) - For logging runtime errors - [log-test-results](log-test-results.md) - For logging test results - [log-quality-check](log-quality-check.md) - For logging code quality metrics - [promote-learning](promote-learning.md) - For promoting validated learnings