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

@@ -16,6 +16,7 @@ import re
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
def validate_corpus_sample():
@@ -94,7 +95,7 @@ def validate_corpus_sample():
output_dir,
make_vector_redaction=False,
also_make_raster_burn=False, # Pas de PDF pour aller plus vite
config_path=Path("config/dictionnaires.yml")
config_path=RUNTIME_DICTIONARIES_CONFIG_PATH
)
doc_time = time.time() - doc_start