Respect Fed release dates and policy document layouts
Development CI / check-and-build (push) Successful in 2m12s
GitHub Git mirror / mirror (push) Successful in 7s
Development CI / deploy-development (push) Successful in 2m31s

Filter discovery links against both realtime and historical date boundaries, keep date-only cutoff days, and use the official calendar release date for FOMC minutes. Parse implementation note and projection headings and replay real Fed discovery and policy bodies without external I/O.
This commit is contained in:
eezd committed 2026-11-04 22:39:56 +08:00
1 parent cd7a314503
commit 328f1e91a1
5 files changed
+128 -10

No files matched your search

+5
View File
@@ -289,6 +289,11 @@ discovery responses were captured on 2026-11-04: the 2025/2026
press, speech and testimony archives, FOMC calendars and H.4.1 release dates.
Their offline regression checks configured entry URLs and article discovery;
Fed speech archives use the singular `speech/` directory.
Three captured policy bodies cover implementation notes, economic projections
and meeting minutes. Minutes use the calendar's explicit release date, with its
source URL recorded, rather than the meeting date encoded in the document URL.
Date-only documents remain eligible on the cutoff day; older discovery links
are filtered before fetching their bodies in both realtime and history lanes.
Additional API/PDF fixtures cover three 2025 DOJ institutional cases (OPA and USAO), their four attachments,
one excluded personal crime case, and a Federal Register daily XML containing 106
unique document numbers. The isolated database smoke checks idempotency and