16 lines
893 B
Markdown
16 lines
893 B
Markdown
|
# tests/fixtures/
|
||
|
|
|
||
|
|
Pinned, anonymised data samples used by the test suite. Each entry below records what the fixture is, where it came from, and which tests it exercises. Add a row when you add a fixture; remove it when nothing references the file.
|
||
|
|
|
||
|
|
| Fixture | Origin | Used by |
|
||
|
|
|---|---|---|
|
||
|
|
|
||
|
|
_(empty — populated as Phase 1 items land)_
|
||
|
|
|
||
|
|
## Conventions
|
||
|
|
|
||
|
|
- **Anonymise before committing.** Strip emails, full names, raw lat/lon outside the run radius. Activity IDs and timestamps are fine.
|
||
|
|
- **Keep them small.** ~5 MB max per `.fit`; aim for the smallest sample that exercises the code path.
|
||
|
|
- **One fixture, one purpose.** If you need two FITs for different tests, commit two files — don't reuse one as both a "happy path" and an "edge case."
|
||
|
|
- **Don't regenerate.** Fixtures are inputs, not outputs. If a test asserts decoupling = 7.3%, that number is locked to *this exact* fixture.
|