Skip to content

Property sweep: dedicated captured corpus and per-domain capture tooling - #241

Open
NandaScott wants to merge 1 commit into
prd/property-sweep-enginefrom
sandcastle/issue-235
Open

Property sweep: dedicated captured corpus and per-domain capture tooling#241
NandaScott wants to merge 1 commit into
prd/property-sweep-enginefrom
sandcastle/issue-235

Conversation

@NandaScott

Copy link
Copy Markdown
Owner

Closes #235

Opened by Sandcastle. 1 commit(s) on sandcastle/issue-235.

#235)

Extend scripts/capture_fixtures.py with per-domain FIXTURE_MAP structure
({"usage": {...}, "sweep": {...}}) and DOMAIN_DIRS, writing each domain's
fixtures to its own directory. The usage domain's output is unchanged apart
from captured_at. capture() now writes discovered_via and truncated_to to
provenance when present in the spec; record_discovered_via is set per-domain
so usage provenance stays byte-identical.

Commit the sweep corpus (33 fixtures) in tests/sweep/fixtures/:
- 12 layout pins (same card IDs as usage; intentional duplication)
- 11 card accessor pins covering hand_modifier, life_modifier, promo_types,
  printed_*, attraction_lights, color_indicator, content_warning, defense,
  flavor_name, loyalty, tcgplayer_etched_id, variation_of, and watermark
- 4 face accessor pins covering face-level cmc/layout/oracle_id, defense,
  loyalty, and printed_* fields
- 1 set (onc; covers 21/21 set accessors vs 14 for lea)
- 1 bulk-data object and 1 catalog
- 3 list envelopes covering has_more/total_cards, warnings, and next_page
  (paginated; truncated_to: 2 recorded in provenance)

Add tests/sweep/fixtures/README.md documenting the selection criterion
(accessor coverage), the full corpus inventory, and how to refresh it.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant