ClassroomCopilot/api
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(exam): guarantee auto-map child rows reference an inserted question
Some checks failed
api-ci-deploy / test-build-deploy (push) Has been cancelled
Details

Browse Source 
On papers where band detection yields few/no questions but opencv/gemma still emit response
regions, those regions referenced a synthetic default_qid that was never inserted -> FK violation
(exam_response_areas/exam_boundaries -> exam_questions). Ensure the fallback container question
exists and reattach orphan child rows to it.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
CC Worker 2026-06-08 18:45:09 +00:00
parent e83873e822
commit 44ccba2151
1 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
routers/exam/templates.py
View File

@ -540,6 +540,23 @@ def _map_first_pass_to_rows(template_id: str, first_pass: Dict[str, Any], pdf_by
response_form = _response_form_from_region_type(region.get("region_type"))
if response_form:
response_areas.append({"id": _ai_id(template_id, "region", page_index, idx), "template_id": template_id, "question_id": first_part_by_page.get(page_index, default_qid), "page": page_index + 1, "bounds": bounds, "kind": "response", "response_form": response_form, "source": "ai", "confirmed": False, "confidence": _safe_confidence(region.get("confidence")), "derivation": region.get("detection_method") or "opencv-response-region"})
# Integrity guard: every response_area/boundary question_id must reference an inserted question
# (FK exam_response_areas/exam_boundaries -> exam_questions). On papers where band detection yields
# few/no questions but opencv/gemma still emit regions, those regions point at the synthetic
# default_qid which was never inserted. Ensure that fallback container question exists and reattach
# any orphan child rows to it, so persistence can't violate the FK.
qid_set = {q["id"] for q in questions}
orphans = [r for r in (response_areas + boundaries) if r.get("question_id") not in qid_set]
if orphans:
if default_qid not in qid_set:
questions.insert(0, {"id": default_qid, "template_id": template_id, "label": "Unassigned",
"order": 0, "max_marks": 0, "is_container": True, "source": "ai",
"confirmed": False, "confidence": 0.5,
"derivation": "auto-map-fallback-container"})
qid_set.add(default_qid)
for r in orphans:
r["question_id"] = default_qid
return {"questions": questions, "response_areas": response_areas, "boundaries": boundaries, "layout": layout}