fix: Handle None message fields safely in get_message_content()

- Fixed AttributeError when msg['message'] is None
- Added get_message_content() to public API for safe extraction
- Added black box tests to prevent regression
- Handles all None scenarios defensively

Fixes bug reported by claude-explorer where msg.get('message', {}).get('content')
would fail when message field exists but is None.

🤖 Generated with Claude Code (https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: alicoding

claude_parser/__init__.py

@@ -16,6 +16,7 @@
 from .session import SessionManager
 from .export import export_for_llamaindex
 from .filtering import filter_messages_by_type, filter_messages_by_tool, search_messages_by_content, exclude_tool_operations
+from .messages.utils import get_message_content, get_text
 
 # Version info
 __version__ = "2.1.0"

claude_parser/messages/utils.py

@@ -87,4 +87,27 @@ def is_hook_message(msg: Dict[str, Any]) -> bool:
 
 def is_tool_operation(msg: Dict[str, Any]) -> bool:
     """Check if message is a tool operation"""
-    return bool(msg.get('tool_use_id') or msg.get('tool_result'))
+    return bool(msg.get('tool_use_id') or msg.get('tool_result'))
+
+
+def get_message_content(msg: Dict[str, Any]) -> str:
+    """Safely extract content from message, handling None values.
+
+    @FRAMEWORK_FIRST: Handles the bug where msg['message'] = None
+    """
+    # Direct content field
+    if 'content' in msg:
+        content = msg['content']
+        if isinstance(content, str):
+            return content
+        elif isinstance(content, list):
+            # Handle content blocks
+            return get_text(msg)
+
+    # Nested message field - MUST check for None explicitly
+    if 'message' in msg and msg['message'] is not None:
+        nested = msg['message']
+        if isinstance(nested, dict):
+            return nested.get('content', '')
+
+    return ''

tests/test_message_none_bug.py

@@ -0,0 +1,51 @@
+#!/usr/bin/env python3
+"""
+Black Box Test: Message Content Extraction with None Values
+Tests that get_message_content() handles None message fields correctly
+"""
+
+import pytest
+from claude_parser import get_message_content, get_text
+
+
+def test_message_none_bug_fixed():
+    """Test that msg['message'] = None doesn't cause AttributeError
+
+    This was the reported bug from claude-explorer where
+    msg.get('message', {}).get('content', '') would fail
+    """
+    # The bug case: message key exists but is None
+    msg = {'message': None}
+
+    # Should not raise AttributeError
+    content = get_message_content(msg)
+    assert content == ''
+
+    # Also test with get_text
+    text = get_text(msg)
+    assert text == ''
+
+
+def test_message_content_safe_extraction():
+    """Test safe extraction with various None scenarios"""
+    test_cases = [
+        ({'message': None}, ''),  # Bug case
+        ({'content': 'direct'}, 'direct'),  # Direct content
+        ({'message': {'content': 'nested'}}, 'nested'),  # Nested
+        ({}, ''),  # Empty
+        ({'content': None}, ''),  # Content is None
+        ({'message': {}}, ''),  # Empty message dict
+    ]
+
+    for msg, expected in test_cases:
+        assert get_message_content(msg) == expected
+
+
+def test_content_block_extraction():
+    """Test extraction from content blocks"""
+    msg = {'content': [{'type': 'text', 'text': 'Block text'}]}
+    assert get_message_content(msg) == 'Block text'
+
+    # Empty blocks
+    msg2 = {'content': []}
+    assert get_message_content(msg2) == ''