Feat/metrics

[jay-m-dev]

This pull request introduces several important improvements to ASAREE, focusing on measurement and metrics terminology, enhanced documentation, test infrastructure, and frontend API updates. The changes clarify how built-in and custom metrics are handled, improve developer and user documentation, add test runner support in Docker and Compose, and expand frontend capabilities for measurement plan management and testing.

Measurement and Metrics Documentation:

• Added a comprehensive glossary to CONTEXT.md defining key terms for experiments, metrics, and runs, clarifying the distinction between built-in and custom metrics.
• Introduced an ADR (docs/adr/0001-measurements-have-explicit-producers.md) specifying that built-in metrics are evaluated by ASAREE, while custom metrics are reported from agent outputs without post-run evaluation.
• Expanded the README.md with detailed explanations of measurement plans, metric states, CSV export structure, and legacy value handling.

Test Infrastructure:

• Added a test image stage to the Dockerfile and a corresponding asaree-tests service to compose.yml, enabling backend tests to run in Docker Compose against a real Postgres instance, separate from the production image. [1] [2]
• Updated .dockerignore to include the tests/ directory, ensuring test code is not included in production images.
• Documented the new Compose-based test workflow in the README.md.

Frontend and API Enhancements:

• Extended the frontend API (frontend/src/api/client.ts) to support measurement plan validation, retrieval of measurement capabilities, and test runs for protocols. Also updated experiment creation and update methods to include measurement plan data. [1] [2] [3] [4] [5] [6]
• Added new frontend dev dependencies and scripts for testing with Vitest and Testing Library, preparing the frontend for component and integration tests. [1] [2]

Documentation and Project Metadata:

• Added project badges (release, CI, Python version, license) to the README.md for improved project visibility and status tracking.
• Updated CLAUDE.md with sections on agent skills, issue tracking, triage labels, and domain documentation.

Configuration and Limits:

• Introduced administrative environment variables for metric producer concurrency and timeout controls in .env.example.
• Changed the base Docker image to use a named stage (FROM python:3.13-slim AS application) for multi-stage builds.

These changes collectively improve clarity around metrics, strengthen testability, and enhance both developer and user experience.