Externalize dictionaries and add anonymization review corpus

This commit is contained in:
2026-04-21 10:32:57 +02:00
parent 012445755a
commit 500ebc28c2
99 changed files with 1805 additions and 805 deletions

View File

@@ -10,6 +10,7 @@ from pathlib import Path
import json
sys.path.insert(0, str(Path(__file__).parent.parent))
from config_defaults import RUNTIME_DICTIONARIES_CONFIG_PATH
from anonymizer_core_refactored_onnx import process_pdf
# 5 documents du corpus production (OGC 008)
@@ -58,7 +59,7 @@ for pdf_path in test_docs[:5]:
out_dir=out_dir,
make_vector_redaction=False,
also_make_raster_burn=False,
config_path=Path("config/dictionnaires.yml"),
config_path=RUNTIME_DICTIONARIES_CONFIG_PATH,
use_hf=False,
ner_manager=None,
vlm_manager=None,