fix field parse regex

logtools/log/sources/parse/chronicles_raw_source.py

@@ -14,9 +14,9 @@ _LOG_LINE = re.compile(
     r'count=(?P<count>\d+)$'
 )
 
-_TOPICS = re.compile(r'((\w+=("[\w\s]+"|\S+) )+)?\w+=("[\w\s]+"|\S+)$')
+_TOPICS = re.compile(r'((\w+=("[^"]+"|\S+) )+)?\w+=("[^"]+"|\S+)$')
 
-_TOPICS_KV = re.compile(r'(?P<key>\w+)=(?P<value>"[\w\s]+"|\S+)')
+_TOPICS_KV = re.compile(r'(?P<key>\w+)=(?P<value>"[^"]+"|\S+)')
 
 
 class LogLevel(Enum):

logtools/log/sources/parse/tests/test_chronicles_raw_source.py

@@ -50,3 +50,21 @@ def test_should_parse_chronicles_fields():
         'type': 'WantBlock',
         'items': '1',
     }
+
+
+def test_should_parse_topics_with_non_alphanumeric_characters():
+    source = ChroniclesRawSource(
+        StringLogSource(
+            lines='WRN 2024-02-02 20:38:47.316+00:00 a message topics="codex pendingblocks" address="cid: zDx*QP4zx9" '
+                  'count=10641'
+        )
+    ).__iter__()
+
+    line = next(source)
+
+    assert line.message == "a message"
+    assert line.count == 10641
+    assert line.fields == {
+        'topics': '"codex pendingblocks"',
+        'address': '"cid: zDx*QP4zx9"',
+    }