feat(military): USPTO PatentsView defense/dual-use patent filing seeder (#2047) (#2091)

Browse Source

* feat(military): USPTO PatentsView defense/dual-use patent seeder (#2047)

* fix(military): correct PatentsView API field names, sort param, and total semantics

- Use nested dot-notation field names required by PatentsView v1 API:
  assignees.assignee_organization (was flat assignee_organization)
  cpc_at_issue.cpc_subclass_id (was non-existent cpc_subgroup_id)
- Split sort into separate s= param; fix per_page -> size in o= param
  (sort inside o= and per_page are both rejected by PatentsView v1)
- Fix listDefensePatents total to reflect full seeded dataset size,
  not the post-filter count (proto doc: "Total number of filings in
  the seeded dataset")

* feat(defense-patents): add DefensePatentsPanel — R&D Signal frontend

Panel displays weekly USPTO defense/dual-use patent filings from seed.
Five tabs: All | Comms (H04B) | Semiconductors (H01L) | Ammunition (F42B)
| AI (G06N) | Biotech (C12N). Shows assignee, title, CPC tag, date, USPTO link.

Wired into panel-layout.ts and registered as 'defense-patents' in panels.ts.

* fix(defense-patents): TTL 3x interval, recordCount, maxStaleMin 2x, wire into App.ts

- CACHE_TTL: 7d → 21d (3× weekly interval) so key survives missed runs
- recordCount: add (d) => d?.patents?.length ?? 0 so seed logs correct count
- health.js maxStaleMin: 10080 → 20160 (2× 7-day interval per gold standard)
- App.ts: import DefensePatentsPanel, add primeTask (initial load) and
  scheduleRefresh at 24h interval (daily poll for weekly data)

...

This commit is contained in:

Elie Habib

2026-03-26 17:50:37 +04:00

committed by GitHub

parent 1a2f1abf27

commit d4917cf00b

17 changed files with 593 additions and 1 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

2

docs/api/MilitaryService.openapi.json

File diff suppressed because one or more lines are too long