Commit ·
d0d7322
1
Parent(s): 0102eee
Add ARCH debug logging to auto_classify
Browse files
api/routers/classification.py
CHANGED
|
@@ -42,8 +42,12 @@ async def auto_classify(req: RunExtractionReq, current_user: dict = Depends(get_
|
|
| 42 |
str_texts = []
|
| 43 |
for b in str_boundaries:
|
| 44 |
pdf_p = os.path.join(project_cache, b.get("pdf_path", ""))
|
| 45 |
-
|
| 46 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
|
| 48 |
# Fallback to DB-stored (truncated) texts if PDFs are missing from disk
|
| 49 |
if not str_texts:
|
|
@@ -53,8 +57,12 @@ async def auto_classify(req: RunExtractionReq, current_user: dict = Depends(get_
|
|
| 53 |
arch_texts = []
|
| 54 |
for b in arch_boundaries:
|
| 55 |
pdf_p = os.path.join(project_cache, b.get("pdf_path", ""))
|
| 56 |
-
|
| 57 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
if not arch_texts:
|
| 60 |
arch_texts = state_data.get("arch_texts") or []
|
|
@@ -90,7 +98,20 @@ async def auto_classify(req: RunExtractionReq, current_user: dict = Depends(get_
|
|
| 90 |
|
| 91 |
return {
|
| 92 |
"classified_pages": classified,
|
| 93 |
-
"next_step": 3
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
}
|
| 95 |
|
| 96 |
def _background_record_memory(user_id: int, pages: list):
|
|
|
|
| 42 |
str_texts = []
|
| 43 |
for b in str_boundaries:
|
| 44 |
pdf_p = os.path.join(project_cache, b.get("pdf_path", ""))
|
| 45 |
+
exists = os.path.exists(pdf_p)
|
| 46 |
+
print(f"[classify/auto] STR PDF path={pdf_p} exists={exists}")
|
| 47 |
+
if exists:
|
| 48 |
+
extracted = extract_page_text(pdf_p)
|
| 49 |
+
print(f"[classify/auto] STR extracted {len(extracted)} pages, first 200 chars: {extracted[0][:200] if extracted else '(none)'}")
|
| 50 |
+
str_texts.extend(extracted)
|
| 51 |
|
| 52 |
# Fallback to DB-stored (truncated) texts if PDFs are missing from disk
|
| 53 |
if not str_texts:
|
|
|
|
| 57 |
arch_texts = []
|
| 58 |
for b in arch_boundaries:
|
| 59 |
pdf_p = os.path.join(project_cache, b.get("pdf_path", ""))
|
| 60 |
+
exists = os.path.exists(pdf_p)
|
| 61 |
+
print(f"[classify/auto] ARCH PDF path={pdf_p} exists={exists}")
|
| 62 |
+
if exists:
|
| 63 |
+
extracted = extract_page_text(pdf_p)
|
| 64 |
+
print(f"[classify/auto] ARCH extracted {len(extracted)} pages, first 200 chars: {extracted[0][:200] if extracted else '(none)'}")
|
| 65 |
+
arch_texts.extend(extracted)
|
| 66 |
|
| 67 |
if not arch_texts:
|
| 68 |
arch_texts = state_data.get("arch_texts") or []
|
|
|
|
| 98 |
|
| 99 |
return {
|
| 100 |
"classified_pages": classified,
|
| 101 |
+
"next_step": 3,
|
| 102 |
+
"_debug": {
|
| 103 |
+
"cache_dir": project_cache,
|
| 104 |
+
"cache_exists": os.path.isdir(project_cache),
|
| 105 |
+
"cache_files": os.listdir(project_cache) if os.path.isdir(project_cache) else [],
|
| 106 |
+
"str_boundaries": str_boundaries,
|
| 107 |
+
"arch_boundaries": arch_boundaries,
|
| 108 |
+
"str_texts_count": len(str_texts),
|
| 109 |
+
"arch_texts_count": len(arch_texts),
|
| 110 |
+
"str_text_sample": str_texts[0][:300] if str_texts else "(empty)",
|
| 111 |
+
"arch_text_sample": arch_texts[0][:300] if arch_texts else "(empty)",
|
| 112 |
+
"unknown_count": sum(1 for p in classified if p.get("detected_type") == "unknown"),
|
| 113 |
+
"total_classified": len(classified),
|
| 114 |
+
}
|
| 115 |
}
|
| 116 |
|
| 117 |
def _background_record_memory(user_id: int, pages: list):
|