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
2025-07-28 18:26:47 +00:00
parent 802703e88e
commit 7b13018726
1 changed files with 41 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

72
app.py
View File

@@ -5,9 +5,8 @@ from flask import Flask, request, jsonify, make_response
from presidio_analyzer import AnalyzerEngineProvider from presidio_analyzer import AnalyzerEngineProvider
# Config du logging
logging.basicConfig(level=logging.INFO, logging.basicConfig(level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s') format="%(asctime)s - %(name)s - %(levelname)s - %(message)s")
logger = logging.getLogger(__name__) logger = logging.getLogger(__name__)
app = Flask(__name__) app = Flask(__name__)
@@ -25,11 +24,11 @@ except Exception as e:
analyzer = None analyzer = None
# Regex IBAN strict : 2 lettres, 2 chiffres, puis 4 groupes de 4 alphanum (minimum), # Regex pour recadrage strict
# avec optional whitespace between groups IBAN_REGEX = re.compile(r"\b[A-Z]{2}[0-9]{2}(?:\s[A-Z0-9]{4}){4,7}\b", re.IGNORECASE)
IBAN_REGEX = re.compile(r"\b[A-Z]{2}[0-9]{2}(?:\s?[A-Z0-9]{4}){4,7}\b", re.IGNORECASE) IPV4_REGEX = re.compile(r"\b(?:25[0-5]|2[0-4]\d|1\d{2}|[1-9]?\d)(?:\.(?:25[0-5]|2[0-4]\d|1\d{2}|[1-9]?\d)){3}\b")
# Labels / titres à exclure de l'anonymisation (en minuscules pour normalisation) # Liste des labels/phrases à exclure de l'anonymisation (en minuscules)
IGNORE_LABELS = { IGNORE_LABELS = {
"témoins", "témoins",
"témoins clés", "témoins clés",
@@ -41,15 +40,13 @@ IGNORE_LABELS = {
"montrent", "montrent",
"montrent des", "montrent des",
"montrent des irrégularités", "montrent des irrégularités",
"bénéficiaire" "bénéficiaire",
# ajouter d'autres labels ici si besoin
} }
def normalize_label(txt): def normalize_label(text: str) -> str:
return txt.strip().lower() return text.strip().lower()
@app.route("/analyze", methods=["POST"])
@app.route('/analyze', methods=['POST'])
def analyze_text(): def analyze_text():
if not analyzer: if not analyzer:
return jsonify({"error": "Analyzer engine is not available. Check startup logs."}), 500 return jsonify({"error": "Analyzer engine is not available. Check startup logs."}), 500
@@ -62,47 +59,60 @@ def analyze_text():
if not text_to_analyze: if not text_to_analyze:
return jsonify({"error": "text field is missing or empty"}), 400 return jsonify({"error": "text field is missing or empty"}), 400
results = analyzer.analyze( results = analyzer.analyze(text=text_to_analyze, language=language)
text=text_to_analyze,
language=language
)
filtered_results = [] filtered_results = []
for res in results: for res in results:
ent_text = text_to_analyze[res.start:res.end].strip() ent_text = text_to_analyze[res.start:res.end].strip()
ent_text_norm = normalize_label(ent_text) ent_text_norm = normalize_label(ent_text)
# 1. Ignorer les entités correspondant exactement aux labels/titres à préserver # Skip anonymization for labels to keep
if ent_text_norm in IGNORE_LABELS: if ent_text_norm in IGNORE_LABELS:
logger.debug(f"Skip anonymizing label: '{ent_text}'") logger.debug(f"Skipping anonymization of label: '{ent_text}'")
continue continue
# 2. Si entité de type IBAN, resserrer la sélection strictement au format IBAN # Recadrage stricte 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:
true_iban = match.group(0) true_iban = match.group(0)
# Recalcule start/end dans le texte original
start_offset = ent_text.find(true_iban) start_offset = ent_text.find(true_iban)
if start_offset != -1: if start_offset != -1:
old_start = res.start old_start, old_end = res.start, res.end
old_end = res.end
res.start += start_offset res.start += start_offset
res.end = res.start + len(true_iban) res.end = res.start + len(true_iban)
ent_text = true_iban ent_text = true_iban
logger.debug(f"Correct IBAN span from ({old_start}-{old_end}) to ({res.start}-{res.end}): '{ent_text}'") logger.debug(f"Adjusted IBAN span from ({old_start}-{old_end}) to ({res.start}-{res.end}): '{ent_text}'")
else: else:
# Aucun start trouvé (peu probable), garder tel quel logger.warning(f"Cannot find exact IBAN substring inside entity: '{ent_text}'")
logger.warning(f"Could not find exact IBAN substring in entity text: '{ent_text}'")
else: else:
# Pas de correspondance valide au regex IBAN, on peut choisir d'ignorer ou garder tel quel # Pas un IBAN valide, ignorer cette entité
logger.warning(f"Entity IBAN does not match IBAN regex: '{ent_text}'") logger.warning(f"Entity IBAN does not match strict IBAN regex: '{ent_text}'")
continue
# Recadrage stricte IP_ADDRESS (IPv4 uniquement ici)
if res.entity_type == "IP_ADDRESS":
match = IPV4_REGEX.search(ent_text)
if match:
true_ip = match.group(0)
start_offset = ent_text.find(true_ip)
if start_offset != -1:
old_start, old_end = res.start, res.end
res.start += start_offset
res.end = res.start + len(true_ip)
ent_text = true_ip
logger.debug(f"Adjusted IP_ADDRESS span from ({old_start}-{old_end}) to ({res.start}-{res.end}): '{ent_text}'")
else:
logger.warning(f"Cannot find exact IP substring inside entity: '{ent_text}'")
else:
logger.warning(f"Entity IP_ADDRESS does not match IPv4 regex: '{ent_text}'")
continue
filtered_results.append(res) filtered_results.append(res)
response_data = [res.to_dict() for res in filtered_results] # Optionnel: ici tu peux aussi filtrer les chevauchements si tu veux
# **Important :** retourne uniquement les entités à anonymiser (sans labels exclus) response_data = [res.to_dict() for res in filtered_results]
return make_response(jsonify(response_data), 200) return make_response(jsonify(response_data), 200)
except Exception as e: except Exception as e:
@@ -112,5 +122,5 @@ def analyze_text():
return jsonify({"error": str(e)}), 500 return jsonify({"error": str(e)}), 500
if __name__ == '__main__': if __name__ == "__main__":
app.run(host='0.0.0.0', port=5001) app.run(host="0.0.0.0", port=5001)