nacim/Presidio
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Actualiser app.py

Browse Source
This commit is contained in:
nacim
2025-08-05 09:00:08 +00:00
parent 63cc87b8df
commit 5f8c512099
1 changed files with 13 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
app.py
View File

@@ -11,7 +11,8 @@ logger = logging.getLogger(__name__)
app = Flask(__name__) app = Flask(__name__)
# Chargement du moteur Presidio via Provider # Chargement du moteur
analyzer = None analyzer = None
try: try:
logger.info("--- Presidio Analyzer Service Starting ---") logger.info("--- Presidio Analyzer Service Starting ---")
@@ -23,16 +24,19 @@ except Exception as e:
logger.exception("Error during AnalyzerEngine initialization.") logger.exception("Error during AnalyzerEngine initialization.")
analyzer = None analyzer = None
# Regex strict pour IBAN belge format attendu
# Test Temporaire pour les Regex via du Python directement
IBAN_REGEX = re.compile(r"\b[A-Z]{2}[0-9]{2}(?:\s[0-9]{4}){3}\b", re.IGNORECASE) IBAN_REGEX = re.compile(r"\b[A-Z]{2}[0-9]{2}(?:\s[0-9]{4}){3}\b", re.IGNORECASE)
# Regex IPv4
IPV4_REGEX = re.compile( IPV4_REGEX = re.compile(
r"\b(?:(?:25[0-5]|2[0-4][0-9]|1\d{2}|[1-9]?\d)\.){3}" r"\b(?:(?:25[0-5]|2[0-4][0-9]|1\d{2}|[1-9]?\d)\.){3}"
r"(?:25[0-5]|2[0-4][0-9]|1\d{2}|[1-9]?\d)\b" r"(?:25[0-5]|2[0-4][0-9]|1\d{2}|[1-9]?\d)\b"
) )
# Liste des labels/phrases à exclure d’anonymisation (en minuscules) # Liste Temporaire en surcouche des labels/phrases à exclure d’anonymisation
IGNORE_LABELS = { IGNORE_LABELS = {
"témoins", "témoins",
"témoins clés", "témoins clés",
@@ -74,7 +78,8 @@ def analyze_text():
logger.debug(f"Skipping anonymization of label: '{ent_text}'") logger.debug(f"Skipping anonymization of label: '{ent_text}'")
continue continue
# Recadrage IBAN strict # Recadrage IBAN
if res.entity_type == "IBAN": if res.entity_type == "IBAN":
match = IBAN_REGEX.search(ent_text) match = IBAN_REGEX.search(ent_text)
if match: if match:
@@ -91,7 +96,8 @@ def analyze_text():
logger.warning(f"Invalid IBAN detected, skipping: '{ent_text}'") logger.warning(f"Invalid IBAN detected, skipping: '{ent_text}'")
continue continue
# Recadrage IP_ADDRESS strict IPv4 (wildcard possible pour IPv6 si besoin) # Recadrage IP_ADDRESS
if res.entity_type == "IP_ADDRESS": if res.entity_type == "IP_ADDRESS":
match = IPV4_REGEX.search(ent_text) match = IPV4_REGEX.search(ent_text)
if match: if match:
@@ -110,7 +116,7 @@ def analyze_text():
filtered_results.append(res) filtered_results.append(res)
# Retourner le résultat nettoyé # Résultat nettoyé
response_data = [res.to_dict() for res in filtered_results] response_data = [res.to_dict() for res in filtered_results]
return make_response(jsonify(response_data), 200) return make_response(jsonify(response_data), 200)