Spaces:
Sleeping
Sleeping
| # ============================================================ | |
| # 🧠 ENTITY DICTIONARY — MAURITIUS CONTEXT (FINAL) | |
| # ============================================================ | |
| ENTITY_DICTIONARY = { | |
| # ========================== | |
| # 1️⃣ DISTRICTS | |
| # ========================== | |
| "DISTRICT": [ | |
| "Port Louis", "Pamplemousses", "Riviere du Rempart", | |
| "Flacq", "Grand Port", "Savanne", "Moka", | |
| "Plaines Wilhems", "Riviere Noire", "Black River" | |
| ], | |
| # ========================== | |
| # 2️⃣ CITIES / TOWNS | |
| # ========================== | |
| "CITY": [ | |
| "Curepipe", "Vacoas", "Quatre Bornes", | |
| "Beau Bassin", "Rose Hill", "Ebene", "Phoenix", | |
| "Goodlands", "Grand Baie", "Mahebourg", | |
| "Flic en Flac", "Tamarin", "Bel Air", | |
| "Roche Bois", "Piton", "Triolet" | |
| ], | |
| # ========================== | |
| # 3️⃣ VILLAGES (ONLY villages) | |
| # ========================== | |
| "VILLAGE": [ | |
| "Albion", "Amaury", "Amitie-Gokhoola", "Arsenal", | |
| "Baie du Tombeau", "Baie du Cap", "Bambous", | |
| "Bambous Virieux", "Bananes", "Beau Vallon", | |
| "Belle Vue Maurel", "Benares", "Bois Cheri", | |
| "Bois des Amourettes", "Bon Accueil", "Bramsthan", | |
| "Britannia", "Calebasses", "Camp Diable", | |
| "Camp Ithier", "Cap Malheureux", "Case Noyale", | |
| "Chemin Grenier", "Cluny", "Cottage", | |
| "Creve Coeur", "Dagotier", "Dubreuil", | |
| "Ecroignard", "Esperance", "Fond du Sac", | |
| "Grand Bois", "Grand Gaube", "Gros Cailloux", | |
| "Lallmatie", "Laventure", "Le Hochet", | |
| "Long Mountain", "Mapou", "Melrose", | |
| "Midlands", "Montagne Blanche", | |
| "New Grove", "Notre Dame", "Nouvelle France", | |
| "Old Grand Port", "Plaine Magnien", | |
| "Poste de Flacq", "Providence", | |
| "Quartier Militaire", "Richelieu", | |
| "Rose Belle", "Sebastopol", | |
| "Souillac", "St Pierre", "Surinam", | |
| "Terre Rouge", "The Vale", "Tyack", | |
| "Union Park", "Verdun" | |
| ], | |
| # ========================== | |
| # 4️⃣ LANDMARKS / REGIONS | |
| # ========================== | |
| "LANDMARK": [ | |
| "Bagatelle", "Caudan Waterfront", "Le Caudan", | |
| "Ebene Cybercity", "Mont Choisy", | |
| "Chamarel", "Le Morne", "Blue Bay", | |
| "La Cambuse", "Ile aux Cerfs", | |
| "SSR Airport", "Plaisance", "Medine", | |
| "Cascavelle" | |
| ], | |
| # ========================== | |
| # 5️⃣ PERSON NAMES | |
| # ========================== | |
| "PERSON": [ | |
| "Amit", "Vikash", "Rishi", "Kiran", "Anil", | |
| "Sanjay", "Rakesh", "Nikhil", "Kavita", | |
| "Aisha", "Fatima", "Imran", "Yusuf", | |
| "Jean", "Marie", "Patrick", "Kevin", | |
| "Ramesh", "Champa", "Babita", | |
| "Sidharth", "Sameer", "Karuna" | |
| ], | |
| # ========================== | |
| # 6️⃣ PRODUCTS | |
| # ========================== | |
| "PRODUCT": [ | |
| "Honda Fit", "Honda Vezel", "Toyota Aqua", | |
| "Toyota Vitz", "Suzuki Swift", | |
| "iPhone", "Samsung", "OnePlus", "Pixel" | |
| ], | |
| # ========================== | |
| # 7️⃣ REGEX PATTERNS (ORDER CRITICAL) | |
| # ========================== | |
| "REGEX_PATTERNS": { | |
| # 💰 MONEY (atomic) | |
| "MONEY": r"\b(?:Rs|rs|RS)\s?\d+(?:,\d{3})*(?:\.\d+)?\b", | |
| # ⏰ TIME | |
| "TIME": r"\b\d{1,2}\s?(?:am|pm)\b|\b\d{1,2}:\d{2}\b", | |
| # 📅 DATE | |
| "DATE": r"\b\d{1,2}[\/\-.]\d{1,2}[\/\-.]\d{2,4}\b", | |
| # 📞 PHONE | |
| "PHONE": r"\b\d{8}\b", | |
| # 🔢 NUMBER (LOWEST) | |
| "NUMBER": r"\b\d+(?:,\d{3})*(?:\.\d+)?\b" | |
| } | |
| } | |