Externalize dictionaries and add anonymization review corpus

Pseudonymisation_Gui_Models_V4.py

@@ -48,33 +48,16 @@ try:
 except Exception:
     yaml = None
 
-APP_TITLE = "Pseudonymisation de PDF"
-DEFAULT_CFG = Path("config/dictionnaires.yml")
+from config_defaults import (
+    RUNTIME_DICTIONARIES_CONFIG_PATH,
+    read_default_dictionaries_text,
+    read_runtime_dictionaries_overlay_text,
+)
 
-DEFAULTS_CFG_TEXT = r"""
-# dictionnaires.yml – valeurs par défaut (bloc littéral pour les regex)
-version: 1
-encoding: "utf-8"
-normalization: "NFKC"
-whitelist:
-  sections_titres: [DIM, GHM, GHS, RUM, COMPTE, RENDU, DIAGNOSTIC]
-  noms_maj_excepts: ["Médecin DIM", "Praticien conseil"]
-  org_gpe_keep: true
-blacklist:
-  force_mask_terms: []
-  force_mask_regex: []
-kv_labels_preserve: [FINESS, IPP, "N° OGC", Etablissement]
-regex_overrides:
-  - name: OGC_court
-    pattern: |-
-      \b(?:N°\s*)?OGC\s*[:\-]?\s*([A-Za-z0-9\-]{1,3})\b
-    placeholder: '[OGC]'
-    flags: [IGNORECASE]
-flags:
-  case_insensitive: true
-  unicode_word_boundaries: true
-  regex_engine: "python"
-"""
+APP_TITLE = "Pseudonymisation de PDF"
+DEFAULT_CFG = RUNTIME_DICTIONARIES_CONFIG_PATH
+DEFAULTS_CFG_TEXT = read_default_dictionaries_text()
+RUNTIME_CFG_TEXT = read_runtime_dictionaries_overlay_text()
 
 
 class ToolTip:
@@ -208,7 +191,7 @@ class App:
     # YAML helpers
     def _ensure_cfg_exists(self):
         p = Path(self.cfg_path.get()); p.parent.mkdir(parents=True, exist_ok=True)
-        if not p.exists(): p.write_text(DEFAULTS_CFG_TEXT, encoding="utf-8")
+        if not p.exists(): p.write_text(RUNTIME_CFG_TEXT, encoding="utf-8")
     def _cfg_browse(self):
         d = filedialog.asksaveasfilename(defaultextension=".yml", filetypes=[("YAML","*.yml *.yaml"), ("Tous","*.*")])
         if d: self.cfg_path.set(d)
@@ -225,14 +208,14 @@ class App:
         if yaml is None:
             messagebox.showerror("PyYAML manquant", "Installez PyYAML (pip install pyyaml)."); return
         try:
-            Path(self.cfg_path.get()).write_text(yaml.safe_dump(self.cfg_data or yaml.safe_load(DEFAULTS_CFG_TEXT), allow_unicode=True, sort_keys=False), encoding="utf-8")
+            Path(self.cfg_path.get()).write_text(yaml.safe_dump(self.cfg_data or {}, allow_unicode=True, sort_keys=False), encoding="utf-8")
             self._log("Règles sauvegardées.")
         except Exception as e:
             messagebox.showerror("Erreur", f"Impossible d'écrire le YAML: {e}")
     def _reload_cfg(self): self._load_cfg(); self._log("Règles rechargées.")
     def _restore_defaults(self):
         try:
-            Path(self.cfg_path.get()).write_text(DEFAULTS_CFG_TEXT, encoding="utf-8"); self._log("CFG par défaut écrit."); self._load_cfg()
+            Path(self.cfg_path.get()).write_text(RUNTIME_CFG_TEXT, encoding="utf-8"); self._log("Surcharge locale réinitialisée."); self._load_cfg()
         except Exception as e:
             messagebox.showerror("Erreur", f"Impossible d'écrire le YAML par défaut: {e}")