decomposer: generate deliverable files for Implement the task REST API service with FastAPI, including standard CRUD endpoints, HTTP status semantics, request and response models, and structured error handling.; Add a PostgreSQL persistence layer for the task service using SQLAlchemy models, session management, database initialization or migrations, and durable task storage integrated with the existing API contract.; Define and integrate Pydantic request and response schemas for task fields, status values, identifiers, timestamps, and validation errors across the existing FastAPI service and persistence contract.; Add automated pytest coverage for the task API, including CRUD behavior, request and response validation failures, HTTP status semantics, and persistence interactions using isolated test data.; Containerize the FastAPI task service and PostgreSQL persistence service with Docker Compose, including environment-based configuration, health checks, networking, startup dependencies, and persistent database storage.; Create project documentation and operational guidance covering local setup, Docker Compose configuration, environment variables, API endpoints, validation and error responses, testing commands, persistence behavior, and relevant best-practice notes.

docs/api-contract.md

@@ -0,0 +1,5 @@
+# API contract
+
+All successful task responses use the `TaskRead` schema. `status` is one of `pending`, `in_progress`, or `completed`. IDs are UUID strings and timestamps are ISO-8601 values emitted by Pydantic. Create requires a non-blank title; update is partial and rejects an empty title. Query `limit` is 1–100 and `skip` is non-negative.
+
+Errors are always wrapped as `{ "error": { "code", "message", "details" } }`. Validation errors are 422, missing tasks are 404, delete success is 204 with no body, and database readiness failures are 503.