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: bump PaddleOCR model from VL-1.5 to VL-1.6

Browse source
This commit is contained in:
Research Assistant 2026-06-01 00:32:46 +08:00
parent dde9c903fe
commit c1c5b9b8ba
3 changed files with 11 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
paperforge/ocr_diagnostics.py
View file

@ -61,7 +61,7 @@ def ocr_doctor(config: dict[str, str] | None, live: bool = False) -> dict:
resp = requests.post(
job_url,
headers={"Authorization": f"bearer {token}"},
json={"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.5")},
json={"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.6")},
timeout=30,
)
if resp.status_code == 401:
@ -98,7 +98,7 @@ def ocr_doctor(config: dict[str, str] | None, live: bool = False) -> dict:
# L3 — API response structure
try:
minimal_payload = {
"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.5"),
"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.6"),
"optionalPayload": json.dumps({"useDocOrientationClassify": False}),
}
resp = requests.post(
@ -159,7 +159,7 @@ def ocr_doctor(config: dict[str, str] | None, live: bool = False) -> dict:
resp = requests.post(
job_url,
headers={"Authorization": f"bearer {token}"},
data={"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.5")},
data={"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.6")},
files={"file": f},
timeout=120,
)

2
paperforge/setup_wizard.py
View file

@ -718,7 +718,7 @@ def headless_setup(
{
"PADDLEOCR_API_TOKEN": raw_key,
"PADDLEOCR_JOB_URL": paddleocr_url,
"PADDLEOCR_MODEL": "PaddleOCR-VL-1.5",
"PADDLEOCR_MODEL": "PaddleOCR-VL-1.6",
"ZOTERO_DATA_DIR": zotero_data or "",
},
)

9
paperforge/worker/ocr.py
View file

@ -66,7 +66,7 @@ def ensure_ocr_meta(vault: Path, row: dict) -> dict:
meta = {}
meta.setdefault("zotero_key", key)
meta.setdefault("source_pdf", row.get("pdf_path", ""))
meta.setdefault("ocr_provider", "PaddleOCR-VL-1.5")
meta.setdefault("ocr_provider", "PaddleOCR-VL-1.6")
meta.setdefault("mode", "async")
meta.setdefault("ocr_status", "pending")
meta.setdefault("ocr_job_id", "")
@ -1029,6 +1029,8 @@ def is_embedded_figure_text_block(block: dict, blocks: list[dict], page_width: i
label = block.get("block_label", "")
if label not in {"text", "paragraph_title"}:
return False
if label == "paragraph_title":
return False
text = clean_block_text(block.get("block_content", ""))
if not text:
return False
@ -1681,6 +1683,9 @@ def run_ocr(vault: Path, verbose: bool = False, no_progress: bool = False) -> in
meta["retry_count"] = 0
write_json(paths["ocr"] / key / "meta.json", meta)
ocr_queue = sync_ocr_queue(paths, target_rows)
print(f"DEBUG: target_keys count={len(target_keys)}, target_rows count={len(target_rows)}, queue count={len(ocr_queue)}", flush=True)
if ocr_queue:
print(f"DEBUG: first item={ocr_queue[0].get('zotero_key')} status={ocr_queue[0].get('queue_status')}", flush=True)
max_items_raw = os.environ.get("PADDLEOCR_MAX_ITEMS", "").strip()
max_items = 3
if max_items_raw:
@ -1703,7 +1708,7 @@ def run_ocr(vault: Path, verbose: bool = False, no_progress: bool = False) -> in
# Fallback: parse vault-root .env
token = _read_dotenv(vault, "PADDLEOCR_API_TOKEN")
job_url = os.environ.get("PADDLEOCR_JOB_URL", "https://paddleocr.aistudio-app.com/api/v2/ocr/jobs").strip()
model = os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.5").strip()
model = os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.6").strip()
optional_payload = {"useDocOrientationClassify": False, "useDocUnwarping": False, "useChartRecognition": False}
changed = 0
active_submitted = 0