Fix missing blacklist URLs from report

root/usr/bin/mailstats.py

@@ -1173,10 +1173,62 @@ def display_keys_and_values(data):
         raise ValueError("Input must be a list of dictionaries or a list of lists.")
         
 def extract_blacklist_domain(text):
-    match = re.search(r'http://www\.surbl\.org', text)
-    if match:
-        return "www.surbl.org"
-    return None
+    """
+    Compare 'text' against comma-separated URL strings from global vars
+    RBLList, SBLList, and UBLList. Return the first matching entry or "".
+    Match is done on exact hostname substring OR the base domain (eTLD+1),
+    so 'black.uribl.com' will match text containing 'lookup.uribl.com'.
+    """
+    s = text if isinstance(text, str) else str(text or "")
+    s_lower = s.lower()
+    logging.info(f"extract blacklist called:{text}")
+
+    combined = ",".join([RBLList, SBLList, UBLList])
+
+    def hostname_from(sval: str) -> str:
+        sval = (sval or "").strip().lower()
+        if "://" in sval:
+            # Strip scheme using simple split to avoid needing urlparse
+            sval = sval.split("://", 1)[1]
+        # Strip path and port if present
+        sval = sval.split("/", 1)[0]
+        sval = sval.split(":", 1)[0]
+        # Remove leading wildcards/dots
+        sval = sval.lstrip(".")
+        if sval.startswith("*."):
+            sval = sval[2:]
+        return sval
+
+    def base_domain(hostname: str) -> str:
+        parts = hostname.split(".")
+        if len(parts) >= 3 and parts[-2] in ("co", "org", "gov", "ac") and parts[-1] == "uk":
+            return ".".join(parts[-3:])
+        if len(parts) >= 2:
+            return ".".join(parts[-2:])
+        return hostname
+
+    def boundary_re(term: str):
+        # Match term when not part of a larger domain label
+        return re.compile(r"(?<![A-Za-z0-9-])" + re.escape(term) + r"(?![A-Za-z0-9-])")
+
+    for part in combined.split(","):
+        entry = part.strip()
+        logging.info(f"Comparing: {entry}")
+        if not entry:
+            continue
+
+        entry_host = hostname_from(entry)
+        entry_base = base_domain(entry_host)
+
+        # 1) Try matching the full entry host (e.g., black.uribl.com)
+        if entry_host and boundary_re(entry_host).search(s_lower):
+            return entry
+
+        # 2) Fallback: match by base domain (e.g., uribl.com) to catch lookup.uribl.com, etc.
+        if entry_base and boundary_re(entry_base).search(s_lower):
+            return entry
+
+    return ""
         
 def set_log_level(level):
     """Dynamically adjust logging level (e.g., 'DEBUG', 'INFO', 'ERROR')."""
@@ -1330,19 +1382,19 @@ if __name__ == "__main__":
 			saveData = False
 	
 	nolinks = not saveData 
-	# Not sure we need these...
-	# if (ConfigDB,"qpsmtpd","RHSBL").lower() == 'enabled':
-		# RBLList = get_value(ConfigDB,"qpsmtpd","RBLList")
-	# else:
-		# RBLList = ""
-	# if (ConfigDB,"qpsmtpd","RBLList").lower() == 'enabled':
-		# SBLLIst = get_value(ConfigDB,"qpsmtpd","SBLLIst")
-	# else:
-		# RBLList = ""
-	# if (ConfigDB,"qpsmtpd","RBLList").lower() == 'enabled':
-		# UBLList = get_value(ConfigDB,"qpsmtpd","UBLLIst")
-	# else:
-		# RBLList = ""
+	# Needed to identify blacklist used to reject emails.
+	if get_value(ConfigDB,"qpsmtpd","RHSBL").lower() == 'enabled':
+		RBLList = get_value(ConfigDB,"qpsmtpd","RBLList")
+	else:
+		RBLList = ""
+	if get_value(ConfigDB,"qpsmtpd","DNSBL").lower() == 'enabled':
+		SBLList = get_value(ConfigDB,"qpsmtpd","SBLList")
+	else:
+		SBLList = ""
+	if get_value(ConfigDB,"qpsmtpd","URIBL").lower() == 'enabled':
+		UBLList = get_value(ConfigDB,"qpsmtpd","UBLList")
+	else:
+		UBLList = ""
 	
 	FetchmailIP = '127.0.0.200';       	#Apparent Ip address of fetchmail deliveries
 	WebmailIP = '127.0.0.1';           	#Apparent Ip of Webmail sender
@@ -1577,6 +1629,7 @@ if __name__ == "__main__":
 			error_plugin = parsed_data['error-plugin'].strip()
 			if  error_plugin == 'rhsbl' or error_plugin == 'dnsbl':
 				blacklist_domain = extract_blacklist_domain(parsed_data['sender'])
+				if blacklist_domain:
 					blacklist_found[blacklist_domain] += 1
 			
 			#Log the recipients and deny or accept and spam-tagged counts