initial

honeypot_api.py

@@ -95,26 +95,75 @@ def detect_scam(text):
 
 def extract_intelligence(text):
 
-    patterns = {
-        "phoneNumbers": r"\b\+?\d{1,3}[- ]?\d{10}\b",
-        "bankAccounts": r"\b\d{12,18}\b",
-        "upiIds": r"[a-zA-Z0-9.\-_+]+@[a-zA-Z]+",
-        "phishingLinks": r"https?://[^\s]+",
-        "emailAddresses": r"[a-zA-Z0-9.\-_+]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]+",
-        "caseIds": r"\b(?:CASE|REF|ID|TICKET)[- ]?[A-Z0-9]{4,}\b",
-        "policyNumbers": r"\b(?:POLICY|POL|INS)[- ]?[A-Z0-9]{4,}\b",
-        "orderNumbers": r"\b(?:ORDER|ORD)[- ]?[A-Z0-9]{4,}\b",
+    extracted = {
+        "phoneNumbers": [],
+        "bankAccounts": [],
+        "upiIds": [],
+        "phishingLinks": [],
+        "emailAddresses": [],
+        "caseIds": [],
+        "policyNumbers": [],
+        "orderNumbers": [],
     }
 
-    extracted = {k: [] for k in patterns}
-
-    for key, pattern in patterns.items():
-        matches = re.findall(pattern, text)
-        if matches:
-            extracted[key] = list(set(matches))
+    # Phone Numbers
+    phone_matches = re.findall(r"\+?\d{1,3}[- ]?\d{10}", text)
+    extracted["phoneNumbers"] = list(set(phone_matches))
+
+    # Bank Accounts (12–18 digits only)
+    bank_matches = re.findall(r"\b\d{12,18}\b", text)
+    extracted["bankAccounts"] = list(set(bank_matches))
+
+    # Email Addresses (strict)
+    email_matches = re.findall(
+        r"[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}",
+        text
+    )
+    extracted["emailAddresses"] = list(set(email_matches))
+
+    # UPI IDs (exclude emails with dot-domain)
+    upi_matches = re.findall(
+        r"\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9]+(?:\.[a-zA-Z0-9]+)?\b",
+        text
+    )
+
+    clean_upi = []
+    for u in upi_matches:
+        if u not in extracted["emailAddresses"] and "." not in u.split("@")[1]:
+            clean_upi.append(u)
+
+    extracted["upiIds"] = list(set(clean_upi))
+
+    # Phishing Links (remove trailing punctuation)
+    link_matches = re.findall(r"https?://[^\s]+", text)
+    clean_links = [l.rstrip(".,)") for l in link_matches]
+    extracted["phishingLinks"] = list(set(clean_links))
+
+    # Case IDs (full capture including suffix)
+    case_matches = re.findall(
+        r"\b(?:CASE|REF|ID|TICKET)[- ]?\d+(?:-\d+)*\b",
+        text,
+        flags=re.IGNORECASE
+    )
+    extracted["caseIds"] = list(set(case_matches))
+
+    # Policy Numbers (full capture)
+    policy_matches = re.findall(
+        r"\b(?:POLICY|POL|INS)[- ]?\d+(?:-\d+)*\b",
+        text,
+        flags=re.IGNORECASE
+    )
+    extracted["policyNumbers"] = list(set(policy_matches))
+
+    # Order / Transaction IDs
+    order_matches = re.findall(
+        r"\b(?:TXN|ORDER|ORD)[- ]?\d+(?:-\d+)*\b",
+        text,
+        flags=re.IGNORECASE
+    )
+    extracted["orderNumbers"] = list(set(order_matches))
 
     return extracted
-
 # ======================================================
 # HUMAN-LIKE CONVERSATION ENGINE
 # ======================================================