Null Analysis ID Generation Breaking Analysis Links

← All Specs

Goal

Fix /analysis/None being generated in hypothesis pages when a hypothesis has a NULL analysis_id in the database.

Acceptance Criteria

☑ /hypothesis/{id} pages with NULL analysis_id do not contain /analysis/None links

☑ Source Analysis section handles missing analysis_id gracefully

☑ Page cache cleared for stale entries

Root Cause

Two issues:

h.get('analysis_id', '') returns None (not '') when the key exists in the dict but value is None — Python dict.get() only uses the default when key is missing

Stale page_cache entries contained the old HTML with /analysis/None

Fix Applied

Main api.py already had a fix using source_analysis_url variable:

analysis_id = (h.get("analysis_id") or "").strip()
source_analysis_url = f"/analysis/{html.escape(analysis_id)}" if analysis_id else ""

But the page cache still served old HTML with /analysis/None. Cleared hypothesis page cache.

Work Log

2026-04-04 12:42 UTC — Slot 4

Verified /analysis/None → 301 → /analyses/None → 404 (broken)
Found root cause: stale page cache with old HTML containing /analysis/None
Cleared hypothesis page cache: DELETE FROM page_cache WHERE cache_key LIKE 'hypothesis:%'
Tested h-de0d4364 (null analysis_id): /analysis/None no longer present
Result: Done — cache cleared, bug resolved

Tasks using this spec (1)

Null Analysis ID Generation Breaking Analysis Links

done P70

File: f7fa8abf_5fe0_null_analysis_id_spec.md

Modified: 2026-05-01 20:13

Size: 1.5 KB