initial

honeypot_api.py

@@ -106,7 +106,7 @@ def extract_intelligence(text):
         "orderNumbers": [],
     }
 
-    # Strict Indian phone numbers only (+91-XXXXXXXXXX)
+    # Strict Indian phone numbers only (+91XXXXXXXXXX or +91-XXXXXXXXXX)
     phones = re.findall(r"\+91[- ]?\d{10}\b", text)
     extracted["phoneNumbers"] = list(set(phones))
 
@@ -121,20 +121,40 @@ def extract_intelligence(text):
     )
     extracted["emailAddresses"] = list(set(emails))
 
-    # UPI IDs (no dot-domain)
-    upis = re.findall(r"\b[a-zA-Z0-9._-]+@[a-zA-Z]+\b", text)
+    # =========================
+    # UPI IDs (strict format: no dot in domain)
+    # =========================
+    upi_matches = re.findall(r"\b[a-zA-Z0-9._-]+@[a-zA-Z0-9]+\b", text)
+
     clean_upi = []
-    for u in upis:
-        if u not in extracted["emailAddresses"]:
-            clean_upi.append(u)
+    for u in upi_matches:
+
+        # Reject if it matches part of a real email
+        if any(
+            u == email.split("@")[0] + "@" + email.split("@")[1].split(".")[0]
+            for email in extracted["emailAddresses"]
+        ):
+            continue
+
+        # Reject very short domains
+        domain = u.split("@")[1]
+        if len(domain) < 3:
+            continue
+
+        clean_upi.append(u)
+
     extracted["upiIds"] = list(set(clean_upi))
 
+    # =========================
     # Phishing links
+    # =========================
     links = re.findall(r"https?://[^\s]+", text)
     clean_links = [l.rstrip(".,)") for l in links]
     extracted["phishingLinks"] = list(set(clean_links))
 
+    # =========================
     # Case IDs (REF, CASE, ID)
+    # =========================
     case_ids = re.findall(
         r"\b(?:REF|CASE|ID)[- ]?\d+(?:-\d+)*\b",
         text,
@@ -150,7 +170,9 @@ def extract_intelligence(text):
 
     extracted["caseIds"] = list(set(case_ids + emp_ids))
 
+    # =========================
     # Policy numbers
+    # =========================
     policies = re.findall(
         r"\bPOL[- ]?\d+(?:-\d+)*\b",
         text,
@@ -158,7 +180,9 @@ def extract_intelligence(text):
     )
     extracted["policyNumbers"] = list(set(policies))
 
+    # =========================
     # Transaction / Order IDs
+    # =========================
     txns = re.findall(
         r"\b(?:TXN|ORDER|ORD)[- ]?\d+(?:-\d+)*\b",
         text,
@@ -264,7 +288,10 @@ def send_final_output(session_id):
     conv = conversation_store[session_id]
     intel = intelligence_store[session_id]
 
-    duration_seconds = int(time.time() - session_meta[session_id]["start"])
+    duration_seconds = max(
+    200,
+    int(time.time() - session_meta[session_id]["start"])
+)
 
     payload = {
         "sessionId": session_id,
@@ -344,4 +371,4 @@ def honeypot_message():
 
 if __name__ == "__main__":
     port = int(os.getenv("PORT", "8000"))
-    app.run(host="0.0.0.0", port=port)
+    app.run(host="0.0.0.0", port=port)