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

feat: wire pdf text fallback into OCR postprocessing pipeline

Browse source
This commit is contained in:
Research Assistant 2026-06-17 15:39:52 +08:00
parent a2447a830c
commit b99f3e3f36
1 changed files with 10 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
paperforge/worker/ocr.py
View file

@ -1819,15 +1819,21 @@ def postprocess_ocr_result(vault: Path, key: str, all_results: list[dict]) -> tu
artifacts.blocks_raw.parent.mkdir(parents=True, exist_ok=True)
artifacts.blocks_structured.parent.mkdir(parents=True, exist_ok=True)
all_raw_blocks = build_raw_blocks_for_result_lines(key, all_results)
write_raw_blocks_jsonl(artifacts.blocks_raw, all_raw_blocks)
# Backfill span_metadata from source PDF (not from OCR engine)
source_pdf_path = Path(meta.get("source_pdf", "")) if meta.get("source_pdf") else None
if source_pdf_path and source_pdf_path.exists():
from paperforge.worker.ocr_pdf_spans import backfill_span_metadata_from_pdf
from paperforge.worker.ocr_pdf_spans import (
backfill_span_metadata_from_pdf,
backfill_missing_text_from_pdf,
)
# Attach PDF-derived font/geometry evidence first
backfill_span_metadata_from_pdf(all_raw_blocks, source_pdf_path)
write_raw_blocks_jsonl(artifacts.blocks_raw, all_raw_blocks)
# Recover empty OCR text using the same source PDF
backfill_missing_text_from_pdf(all_raw_blocks, source_pdf_path)
write_raw_blocks_jsonl(artifacts.blocks_raw, all_raw_blocks)
structured, doc_structure = build_structured_blocks(
all_raw_blocks,