gmakstutis/ragflow
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

update

Browse source
This commit is contained in:
buua436 2025-11-20 17:47:41 +08:00
parent ff40d37251
commit 41bdf792ae
1 changed files with 1 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
deepdoc/parser/pdf_parser.py
View file

@ -379,7 +379,7 @@ class RAGFlowPdfParser:
x0s = [] x0s = []
for x in x0s_raw: for x in x0s_raw:
if abs(x - min_x0) < INDENT_TOL: if abs(x - min_x0) < INDENT_TOL:
x0s.append([min_x0]) # 缩进视为同一列 x0s.append([min_x0])
else: else:
x0s.append([x]) x0s.append([x])
x0s = np.array(x0s, dtype=float) x0s = np.array(x0s, dtype=float)
@ -437,13 +437,6 @@ class RAGFlowPdfParser:
for b in bxs: for b in bxs:
grouped[b["col_id"]].append(b) grouped[b["col_id"]].append(b)
logging.info(f"\n======= Page {pg} column grouping =======")
for col_id, group in grouped.items():
logging.info(f"Column {col_id}: {len(group)} boxes")
for item in group:
logging.info(f" -> box: x0={item['x0']}, x1={item['x1']}, y0={item['top']}, y1={item['bottom']}")
logging.info("======================================\n")
return boxes return boxes
def _text_merge(self, zoomin=3): def _text_merge(self, zoomin=3):