fix: ipv4 address regex

CHANGELOG.md

@@ -1,5 +1,9 @@
 ## 0.16.11-dev1
 
+### Fixes
+
+- Fix ipv4 regex to correctly include up to three digit octets.
+
 ### Enhancements
 
 - **Enhance quote standardization tests** with additional Unicode scenarios

test_unstructured/cleaners/test_extract.py

@@ -5,7 +5,7 @@
 from unstructured.cleaners import extract
 
 EMAIL_META_DATA_INPUT = """from ABC.DEF.local ([ba23::58b5:2236:45g2:88h2]) by
-    \n ABC.DEF.local ([ba23::58b5:2236:45g2:88h2%25]) with mapi id\
+    \n ABC.DEF.local ([68.183.71.12]) with mapi id\
     n 32.88.5467.123; Fri, 26 Mar 2021 11:04:09 +1200"""
 
 
@@ -37,7 +37,7 @@ def test_extract_email_address():
 def test_extract_ip_address():
     assert extract.extract_ip_address(EMAIL_META_DATA_INPUT) == [
         "ba23::58b5:2236:45g2:88h2",
-        "ba23::58b5:2236:45g2:88h2%25",
+        "68.183.71.12",
     ]
 
 

unstructured/nlp/patterns.py

@@ -92,9 +92,9 @@
 ONE_LINE_BREAK_PARAGRAPH_PATTERN = r"^(?:(?!\.\s*$).)*$"
 ONE_LINE_BREAK_PARAGRAPH_PATTERN_RE = re.compile(ONE_LINE_BREAK_PARAGRAPH_PATTERN)
 
-# IP Address examples: ba23::58b5:2236:45g2:88h2 or 10.0.2.01
+# IP Address examples: ba23::58b5:2236:45g2:88h2, 10.0.2.01 or 68.183.71.12
 IP_ADDRESS_PATTERN = (
-    r"[0-9]{1,2}\.[0-9]{1,2}\.[0-9]{1,2}\.[0-9]{1,2}",
+    r"(?:25[0-5]|2[0-4]\d|1\d{2}|[1-9]?\d)(?:\.(?:25[0-5]|2[0-4]\d|1\d{2}|[1-9]?\d)){3}",
     "[a-z0-9]{4}::[a-z0-9]{4}:[a-z0-9]{4}:[a-z0-9]{4}:[a-z0-9]{4}%?[0-9]*",
 )
 IP_ADDRESS_PATTERN_RE = re.compile(f"({'|'.join(IP_ADDRESS_PATTERN)})")