mellea-fy-fixtures

Melleafy Step 4: Fixture Generation

Version: 4.2.1 | Prereq: Step 3 complete (skeleton emitted with finalised run_pipeline signature) | Produces: fixtures_emission.json

Schema: Output intermediate/fixtures_emission.json MUST conform to schemas/fixtures_emission.schema.json

Output path rule (Rule OUT-4): Step 4 produces ONLY intermediate/fixtures_emission.json in the intermediate directory. This JSON file contains fixture specifications that will be used by downstream processes. DO NOT generate any Python source files — no fixtures/ directory, no .py files, no __init__.py. The fixture source code generation is handled by a separate process outside this skill.

CRITICAL: This step generates ONLY the JSON specification file.

Step 4 generates fixture specifications for 5–8 test fixtures covering ≥3 C-categories. The output is a single JSON file conforming to the schema.

Rule 4-1 — JSON-only fixture specification: Generate all fixture specifications in a single JSON object conforming to schemas/fixtures_emission.schema.json. The invocation receives the run_pipeline signature, the element mapping summary, and the C-category coverage requirement, and returns one JSON object — not Python source. This skill's output is ONLY the JSON file. Do not generate Python fixture files, do not create a fixtures/ directory, do not write __init__.py or individual fixture modules.

---

CRITICAL: Input parameter matching

The keys in every fixture's inputs object MUST be identical to the parameter names of the run_pipeline function in pipeline.py. This is not optional. A fixture with {"text": "..."} for a pipeline that expects run_pipeline(user_query=...) will fail at runtime with a TypeError. Before emitting JSON, verify the exact parameter names from the generated pipeline function. The schema cannot enforce this — it's the model's responsibility to match.

---

Fixture structure

The model emits one JSON object saved as intermediate/fixtures_emission.json. This is the only output of Step 4.