obsidian-plugin/lllin000_PaperForge
1
0
Fork 0
mirror of https://github.com/lllin000/PaperForge.git synced 2026-07-22 06:50:53 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix: use block bbox as natural search bound in prefix recovery, replace 100px threshold

Browse source 
The y-distance limit was a hardcoded 100px. Two improvements:
1. Use the OCR caption block's y-top as the natural upper bound —
   the FIGURE N heading must be above the detected caption bbox.
2. Fall back to page_height * 0.08 (relative, like _cluster_page_assets)
   only when the block has no bbox data.
This commit is contained in:
LLLin000 2026-07-03 02:11:40 +08:00
parent f5f82941df
commit 5e1040b10a
1 changed files with 11 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
paperforge/worker/ocr_figures.py
View file

@ -6085,19 +6085,24 @@ def _recover_figure_heading_prefix(
if len(line_text) > 40:
continue # too long for a heading — probably in-text reference
# Scan lines below the heading (by y-order) for a caption text match
# Scan PDF lines below the heading for a caption text match
heading_y = (line.get("bbox") or [0, 0, 0, 0])[1]
found = False
for j in range(i + 1, min(i + 10, len(sorted_lines))):
# Use OCR block's y-top as natural search bound (Figure N is above the caption bbox)
block_bbox = block.get("bbox") or block.get("block_bbox") or [0, 0, 0, 0]
block_y_top = block_bbox[1]
# Fallback: relative to page height (8% ≈ 3-5 lines at 200dpi)
page_height = float(block.get("page_height") or 1700)
search_limit = heading_y + max(page_height * 0.08, 40.0) if block_y_top == 0 else block_y_top
for j in range(i + 1, min(i + 20, len(sorted_lines))):
nxt = sorted_lines[j]
nxt_y = (nxt.get("bbox") or [0, 0, 0, 0])[1]
if nxt_y - heading_y > 100:
break # too far down — not the same caption block
if nxt_y > search_limit:
break
next_text = str(nxt.get("text", "") or "").strip()
if not next_text or len(next_text) < 10:
continue
# Common-prefix match: at least 15 chars case-insensitive
# Common-prefix match: ≥15 chars or substring match (handles OCR typos)
common = 0
for a, b in zip(next_text.lower(), block_text.lower()):
if a == b: