Primer commit

scripts/audit_phone_collisions.py

@@ -0,0 +1,223 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""audit_phone_collisions.py
+
+Auditoria read-only y PARALELA de los casos de colision de telefono que
+hacen fallar a sync_missing_opps_to_brand.py.
+
+Para cada caso (opp en sucursal sin contraparte en Marca cuyo contacto
+colisiona por telefono con un contacto de Marca de nombre distinto), trae
+EN PARALELO:
+  - Detalle completo del contacto de sucursal (+ sus opps).
+  - Detalle completo del contacto de Marca que colisiona (+ sus opps).
+  - Similitud de nombre via common.match_contacts.
+  - Conteo de cuantos contactos de Marca comparten ese telefono.
+
+Clasifica cada caso:
+  - SAME_PERSON: muy probable el mismo (mismo email / similitud alta) ->
+    deberia asociarse la opp al contacto de Marca existente.
+  - DISTINCT_PERSON: pareja/familia con mismo numero -> skip + revision manual.
+  - UNCERTAIN: requiere ojo humano.
+
+Read-only. No escribe nada.
+"""
+
+import json
+import os
+import sys
+from concurrent.futures import ThreadPoolExecutor, as_completed
+
+ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+if ROOT_DIR not in sys.path:
+    sys.path.insert(0, ROOT_DIR)
+SCRIPTS_DIR = os.path.dirname(os.path.abspath(__file__))
+if SCRIPTS_DIR not in sys.path:
+    sys.path.insert(0, SCRIPTS_DIR)
+
+import requests
+from common import match_contacts, normalize_phone
+
+BASE_URL = "https://services.leadconnectorhq.com"
+BRAND_LOC = "GbKkBpCmKu2QmloKFHy3"
+BRAND_TOKEN = "pit-4e4266f8-97ac-4150-a971-cc9158809640"
+
+# Tokens por location (de la mesa de control)
+TOKENS = {
+    "GbKkBpCmKu2QmloKFHy3": "pit-4e4266f8-97ac-4150-a971-cc9158809640",  # Marca
+    "uZnMH5bO6MXTHcgHeyq9": "pit-dd42c1ce-2ab7-4bf9-8bc0-c0087a83b2e5",  # Pilares
+}
+
+
+def _load_token(loc):
+    if loc in TOKENS:
+        return TOKENS[loc]
+    import csv
+    path = os.path.join(ROOT_DIR, "Bucéfalo - Mesa de control - API Tokens - MP.csv")
+    with open(path, encoding="utf-8-sig") as f:
+        for row in csv.DictReader(f):
+            if (row.get("Location_ID") or "").strip() == loc:
+                return (row.get("API_token") or "").strip()
+    raise RuntimeError(f"sin token para {loc}")
+
+
+def _headers(token):
+    return {"Authorization": f"Bearer {token}", "Version": "2021-07-28", "Accept": "application/json", "Content-Type": "application/json"}
+
+
+def get_contact(loc, contact_id):
+    token = _load_token(loc)
+    r = requests.get(f"{BASE_URL}/contacts/{contact_id}", headers=_headers(token), timeout=30)
+    if r.status_code != 200:
+        return {"error": f"HTTP {r.status_code}: {r.text[:160]}"}
+    return r.json().get("contact", {})
+
+
+def get_opps_for_contact(loc, contact_id):
+    token = _load_token(loc)
+    r = requests.get(f"{BASE_URL}/opportunities/search", headers=_headers(token),
+                     params={"location_id": loc, "contact_id": contact_id, "limit": 50}, timeout=30)
+    if r.status_code != 200:
+        return []
+    return r.json().get("opportunities", []) or []
+
+
+def count_brand_contacts_sharing_phone(phone):
+    """Cuantos contactos de Marca comparten el telefono (normalizado)."""
+    token = BRAND_TOKEN
+    body = {"locationId": BRAND_LOC, "pageLimit": 20, "page": 1,
+            "filters": [{"field": "phone", "operator": "contains", "value": phone}]}
+    r = requests.post(f"{BASE_URL}/contacts/search", headers=_headers(token), json=body, timeout=30)
+    if r.status_code != 200:
+        return None, []
+    contacts = r.json().get("contacts", []) or []
+    np = normalize_phone(phone)
+    matches = [c for c in contacts if normalize_phone(c.get("phone")) == np]
+    return len(matches), [
+        {"id": c.get("id"), "name": f"{c.get('firstName') or ''} {c.get('lastName') or ''}".strip(),
+         "phone": c.get("phone"), "email": c.get("email")}
+        for c in matches
+    ]
+
+
+def audit_one(case):
+    """Procesa un caso. Hace las 5 consultas concurrentemente."""
+    branch_loc = case["branch_location_id"]
+    branch_cid = case["branch_contact_id"]
+    brand_cid = case["colliding_brand_contact_id"]
+    phone = case["phone"]
+
+    out = {"opp_id": case["opp_id"], "branch_name": case["branch_name"],
+           "opp_name": case["opp_name"], "phone": phone}
+
+    with ThreadPoolExecutor(max_workers=5) as ex:
+        futs = {
+            "branch_contact": ex.submit(get_contact, branch_loc, branch_cid),
+            "brand_contact": ex.submit(get_contact, BRAND_LOC, brand_cid),
+            "branch_opps": ex.submit(get_opps_for_contact, branch_loc, branch_cid),
+            "brand_opps": ex.submit(get_opps_for_contact, BRAND_LOC, brand_cid),
+            "phone_share": ex.submit(count_brand_contacts_sharing_phone, phone),
+        }
+        res = {k: f.result() for k, f in futs.items()}
+
+    bc = res["branch_contact"]
+    mc = res["brand_contact"]
+    out["branch_contact"] = {
+        "id": bc.get("id"), "name": f"{bc.get('firstName') or ''} {bc.get('lastName') or ''}".strip(),
+        "email": bc.get("email"), "phone": bc.get("phone"),
+        "dateAdded": bc.get("dateAdded"), "tags": bc.get("tags"),
+    }
+    out["brand_collider"] = {
+        "id": mc.get("id"), "name": f"{mc.get('firstName') or ''} {mc.get('lastName') or ''}".strip(),
+        "email": mc.get("email"), "phone": mc.get("phone"),
+        "dateAdded": mc.get("dateAdded"), "tags": mc.get("tags"),
+    }
+    out["branch_opps_count"] = len(res["branch_opps"])
+    out["brand_opps_count"] = len(res["brand_opps"])
+    out["brand_opps"] = [{"name": o.get("name"), "status": o.get("status"),
+                          "value": o.get("monetaryValue"), "pipelineId": o.get("pipelineId")}
+                         for o in res["brand_opps"][:10]]
+    share_count, share_list = res["phone_share"]
+    out["brand_contacts_sharing_phone"] = share_count
+    out["brand_contacts_sharing_phone_list"] = share_list
+
+    # Similitud de nombre (sin requerir phone/email, solo nombre via match_contacts)
+    src = {"first_name": bc.get("firstName"), "last_name": bc.get("lastName"),
+           "phone": bc.get("phone"), "email": bc.get("email")}
+    dst = {"first_name": mc.get("firstName"), "last_name": mc.get("lastName"),
+           "phone": mc.get("phone"), "email": mc.get("email")}
+    try:
+        m = match_contacts(src, dst, threshold=0.80)
+        out["match_level"] = m.get("level")
+        out["match_name_similarity"] = round(m.get("name_score", 0) or 0, 3)
+    except Exception as e:
+        out["match_level"] = f"error:{e}"
+
+    # Email igual?
+    be = (bc.get("email") or "").strip().lower()
+    me = (mc.get("email") or "").strip().lower()
+    placeholder = {"sincorreo@gmail.com", "noemail@gmail.com", ""}
+    out["same_email"] = bool(be and be == me and be not in placeholder)
+
+    # Clasificacion heuristica
+    sim = out.get("match_name_similarity", 0) or 0
+    if out["same_email"] or sim >= 0.80:
+        verdict = "SAME_PERSON"
+    elif sim >= 0.55:
+        verdict = "UNCERTAIN"
+    else:
+        verdict = "DISTINCT_PERSON"
+    out["verdict"] = verdict
+    return out
+
+
+def main():
+    clean = os.path.join(ROOT_DIR, "generated", "reports", "missing_opps_clean.json")
+    data = json.load(open(clean, encoding="utf-8"))
+    cases = []
+    for it in data["items"]:
+        coll = next((a for a in it.get("actions", []) if a.get("action") == "phone_collision_unresolved"), None)
+        if not coll:
+            continue
+        bc = it.get("branch_contact", {})
+        cases.append({
+            "opp_id": it["opp_id"],
+            "branch_name": it.get("branch_name"),
+            "branch_location_id": it.get("branch_location_id"),
+            "branch_contact_id": bc.get("id"),
+            "opp_name": it.get("opp_name"),
+            "phone": bc.get("phone"),
+            "colliding_brand_contact_id": coll.get("colliding_brand_contact_id"),
+        })
+
+    print(f"Auditando {len(cases)} casos de colision EN PARALELO...\n")
+    results = []
+    with ThreadPoolExecutor(max_workers=len(cases) or 1) as ex:
+        futs = [ex.submit(audit_one, c) for c in cases]
+        for f in as_completed(futs):
+            results.append(f.result())
+
+    results.sort(key=lambda r: r["opp_name"])
+    for r in results:
+        print("=" * 78)
+        print(f"OPP {r['opp_id']} | {r['branch_name']} | tel {r['phone']}")
+        bc = r["branch_contact"]; mc = r["brand_collider"]
+        print(f"  SUCURSAL : {bc['name']!r}  email={bc['email']}  alta={bc['dateAdded']}")
+        print(f"  MARCA(col): {mc['name']!r}  email={mc['email']}  alta={mc['dateAdded']}")
+        print(f"  similitud_nombre={r.get('match_name_similarity')}  match_level={r.get('match_level')}  same_email={r['same_email']}")
+        print(f"  contactos en Marca con ese telefono: {r['brand_contacts_sharing_phone']} -> {[x['name'] for x in r['brand_contacts_sharing_phone_list']]}")
+        print(f"  opps del colider en Marca: {r['brand_opps_count']} -> {[(o['name'], o['status'], o['value']) for o in r['brand_opps']]}")
+        print(f"  >>> VEREDICTO: {r['verdict']}")
+        print()
+
+    out_path = os.path.join(ROOT_DIR, "generated", "reports", "phone_collisions_audit.json")
+    with open(out_path, "w", encoding="utf-8") as f:
+        json.dump(results, f, ensure_ascii=False, indent=2)
+    print(f"Reporte: {out_path}")
+    # Resumen
+    from collections import Counter
+    verdicts = Counter(r["verdict"] for r in results)
+    print(f"Resumen veredictos: {dict(verdicts)}")
+
+
+if __name__ == "__main__":
+    main()