fon-nlu / contact_resolver.py
Ronaldodev's picture
fon-nlu 1.0.0 : intentions + extraction de montants (fon, français, mélange)
28f510f verified
Raw History Blame Contribute Delete
8.7 kB
import re
import unicodedata
from difflib import SequenceMatcher
def normalize_text(text: str) -> str:
"""Normalise un texte (minuscules, sans accents, sans ponctuations superflues)."""
text = unicodedata.normalize('NFKD', text)
text = ''.join(c for c in text if not unicodedata.combining(c))
text = text.replace('ɔ', 'o').replace('ɛ', 'e').replace('ɖ', 'd')
return re.sub(r'[^a-zA-Z0-9\s]', ' ', text).lower().strip()
# Dictionnaire des équivalences de parenté / surnoms en Fongbé et Français
# Permet de mapper par exemple "nɔ ce" (ma mère en fon) vers le contact enregistré sous "Maman"
RELATIONSHIP_ALIASES = {
# Mère / Maman
"maman": ["maman", "mama", "mere", "no ce", "noxó", "no"],
"no ce": ["maman", "mama", "mere", "no ce"],
"mama": ["maman", "mama", "mere", "no ce"],
"mere": ["maman", "mama", "mere"],
# Père / Papa
"papa": ["papa", "pere", "to ce", "dada", "to"],
"to ce": ["papa", "pere", "to ce"],
"pere": ["papa", "pere", "to ce"],
# Frères et Sœurs
"fofo": ["fofo", "grand frere", "frere", "novi ce"],
"fofo ce": ["fofo", "grand frere", "frere"],
"grand frere": ["fofo", "grand frere", "frere"],
"mon grand frere": ["fofo", "grand frere", "frere"],
"dada": ["dada", "grande soeur", "soeur", "novi ce"],
"dada ce": ["dada", "grande soeur", "soeur"],
"grande soeur": ["dada", "grande soeur", "soeur"],
"novi ce": ["frere", "soeur", "fofo", "dada", "novi"],
"mon frere": ["frere", "fofo", "novi ce"],
"ma soeur": ["soeur", "dada", "novi ce"],
# Conjoints
"asi ce": ["femme", "epouse", "cherie", "madame", "asi"],
"ma femme": ["femme", "epouse", "cherie", "asi ce"],
"asu ce": ["mari", "epoux", "cheri", "asu"],
"mon mari": ["mari", "epoux", "cheri", "asu ce"],
}
class ContactResolver:
"""
Résout un nom oral (en Fongbé ou Français) vers un contact du carnet d'adresses,
en gérant les liens de parenté (Maman, Papa, nɔ ce, tɔ́ ce...), le fuzzy matching
et les homonymes.
"""
def __init__(self, contacts: list[dict]):
self.contacts = contacts
def resolve(self, spoken_name: str, threshold: float = 0.65) -> dict:
norm_spoken = normalize_text(spoken_name)
if not norm_spoken:
return {"status": "NOT_FOUND", "match": None, "candidates": []}
# 1. Recherche d'alias (ex: l'usager dit 'nɔ ce', on cherche 'maman', 'mère', etc.)
target_search_terms = [norm_spoken]
for key, aliases in RELATIONSHIP_ALIASES.items():
if norm_spoken == key or norm_spoken in aliases:
target_search_terms.extend(aliases)
target_search_terms.append(key)
target_search_terms = list(dict.fromkeys(target_search_terms))
exact_matches = []
fuzzy_matches = []
for c in self.contacts:
norm_name = normalize_text(c["name"])
# Vérification directe ou via alias
matched = False
for term in target_search_terms:
if term == norm_name:
exact_matches.append((1.0, c))
matched = True
break
elif term in norm_name.split() or norm_name in term.split():
exact_matches.append((0.92, c))
matched = True
break
if not matched:
# Calcul de similarité phonétique / fuzzy Levenshtein
best_sim = max(SequenceMatcher(None, term, norm_name).ratio() for term in target_search_terms)
if best_sim >= threshold:
fuzzy_matches.append((best_sim, c))
all_matches = sorted(exact_matches + fuzzy_matches, key=lambda x: x[0], reverse=True)
# Déduplication par id de contact
seen_ids = set()
dedup_matches = []
for score, contact in all_matches:
if contact["id"] not in seen_ids:
seen_ids.add(contact["id"])
dedup_matches.append((score, contact))
if not dedup_matches:
return {
"status": "NOT_FOUND",
"spoken_name": spoken_name,
"message_fon": f"Un mɔ « {spoken_name} » ɖò xóxlá towe mɛ ǎ. Nambá tɔn tɛ́ wɛ?",
"message_fr": f"Je n'ai pas trouvé « {spoken_name} » dans tes contacts. Quel est son numéro ?"
}
top_score, top_contact = dedup_matches[0]
close_candidates = [c for score, c in dedup_matches if score >= 0.85]
if len(close_candidates) == 1:
return {
"status": "MATCH_UNIQUE",
"match": top_contact,
"confidence": top_score,
"confirmation_fon": f"A jló na sɛ́ akwɛ ɔ dó {top_contact['name']} ({top_contact['phone']}) à?",
"confirmation_fr": f"Veux-tu envoyer l'argent à {top_contact['name']} ({top_contact['phone']}) ?"
}
return {
"status": "MULTIPLE_MATCHES",
"candidates": close_candidates,
"spoken_name": spoken_name,
"disambiguation_fon": f"Mɛ wè wɛ ɖó nyǐkɔ « {spoken_name} » : " + " kpo ".join(c['name'] for c in close_candidates) + ". Mɛ̌ɖé jló na sɛ́ dó?",
"disambiguation_fr": f"Tu as plusieurs contacts sous « {spoken_name} » : " + ", ".join(c['name'] for c in close_candidates) + ". Lequel choisis-tu ?"
}
# CONTACTS DE DÉMONSTRATION UNIQUEMENT (numéros factices). Ne jamais utiliser en production : passez toujours la vraie liste de contacts de l'utilisateur.
DEFAULT_CONTACTS = [
{"id": "1", "name": "Maman", "phone": "+229 00 00 00 01"},
{"id": "2", "name": "Papa", "phone": "+229 00 00 00 02"},
{"id": "3", "name": "Fofó", "phone": "+229 00 00 00 03"},
{"id": "4", "name": "Femme", "phone": "+229 00 00 00 04"},
{"id": "5", "name": "Koffi", "phone": "+229 00 00 00 05"},
{"id": "6", "name": "Rachid", "phone": "+229 00 00 00 06"},
{"id": "7", "name": "Mariam", "phone": "+229 00 00 00 07"},
{"id": "8", "name": "Bio", "phone": "+229 00 00 00 08"},
{"id": "9", "name": "Sèna", "phone": "+229 00 00 00 09"},
{"id": "10", "name": "Femi", "phone": "+229 00 00 00 10"},
{"id": "11", "name": "Ablawa", "phone": "+229 00 00 00 11"},
{"id": "12", "name": "Sègla", "phone": "+229 00 00 00 12"},
{"id": "13", "name": "Modeste", "phone": "+229 00 00 00 13"},
{"id": "14", "name": "Gisèle", "phone": "+229 00 00 00 14"},
{"id": "15", "name": "Tonton", "phone": "+229 00 00 00 15"},
{"id": "16", "name": "Tantie", "phone": "+229 00 00 00 16"}
]
_default_resolver = None
def resolve_contact(query: str, contacts: list[dict] = None) -> dict | None:
"""
Fonction utilitaire pour résoudre un nom de contact ou lien de parenté.
Retourne {'nom': str, 'phone': str, 'match_info': dict} ou None.
"""
global _default_resolver
if contacts is not None:
resolver = ContactResolver(contacts)
else:
if _default_resolver is None:
_default_resolver = ContactResolver(DEFAULT_CONTACTS)
resolver = _default_resolver
res = resolver.resolve(query)
if res["status"] == "MATCH_UNIQUE":
return {
"nom": res["match"]["name"],
"phone": res["match"]["phone"],
"match_info": res
}
elif res["status"] == "MULTIPLE_MATCHES" and res.get("candidates"):
return {
"nom": res["candidates"][0]["name"],
"phone": res["candidates"][0]["phone"],
"match_info": res
}
return None
if __name__ == "__main__":
contacts = [
{"id": "1", "name": "Maman", "phone": "+229 00 00 00 17"},
{"id": "2", "name": "Papa", "phone": "+229 00 00 00 18"},
{"id": "3", "name": "Fofó Mensah", "phone": "+229 00 00 00 19"},
{"id": "4", "name": "Chérie / Femme", "phone": "+229 00 00 00 20"},
{"id": "5", "name": "Kofi", "phone": "+229 00 00 00 21"},
{"id": "6", "name": "Kofi Plombier", "phone": "+229 00 00 00 22"}
]
resolver = ContactResolver(contacts)
print("1. 'nɔ ce' (ma mère en fon) ->", resolver.resolve("nɔ ce")['match']['name'])
print("2. 'tɔ́ ce' (mon père en fon) ->", resolver.resolve("tɔ́ ce")['match']['name'])
print("3. 'maman' (en français) ->", resolver.resolve("maman")['match']['name'])
print("4. 'asì ce' (mon épouse) ->", resolver.resolve("asì ce")['match']['name'])
print("5. 'fofó' (grand frère) ->", resolver.resolve("fofó")['match']['name'])