Testing Guide

weavster test verifies a project's flows with fixtures: an input document, the expected output, and a comparison between them. Each fixture's input is parsed, run through its flow, and compared to the expected output.

Fixture layout

Fixtures are grouped by the flow that transforms them. A fixture folder under fixtures/ names a flow in flows/, and each case inside it has one input and one expected file:

my-integration/
  flows/
    order.yaml          # the flow
  fixtures/
    order/              # cases for flows/order.yaml
      new-order/
        input.json      # source document
        expected.json   # output the flow should produce
      existing-order/
        input.json
        expected.json

• fixtures/<flow>/ maps to flows/<flow>.yaml.
• Each <case>/ has exactly one input.json and one expected.json.
• Both files are JSON. (XML and other format packs arrive in later milestones.)

Running

weavster test [path]

For each case the harness parses input.json, runs it through flows/<flow>.yaml, and compares the result to expected.json with a deep equality check. It prints ✓/✗ per case (labelled <flow>/<case>), a closing passed count, and exits 1 if any case fails.

A failing case shows a line-by-line diff of pretty-printed JSON — - is expected, + is actual:

✗ order/new-order
    {
      "id": "A-1001",
  -   "priority": "normal"
  +   "priority": "high"
    }

A flow that fails to load (missing or schema-invalid) fails every case under it; a flow that throws while running fails that case with the transform error.

Adding a fixture

1. Make sure the flow exists at flows/<flow>.yaml.
2. Create fixtures/<flow>/<case>/ named for the case.
3. Add input.json with a representative source document.
4. Add expected.json with the output the flow should produce for that input.
5. Run weavster test and confirm the case passes.