Fix regex pattern in agent.py:set_knowledge to support Chinese characters (#2454)

Co-Authored-By: Joe Moura <joao@crewai.com>

src/crewai/agent.py

@@ -140,7 +140,7 @@ class Agent(BaseAgent):
                 self.embedder = crew_embedder
 
             if self.knowledge_sources:
-                full_pattern = re.compile(r"[^a-zA-Z0-9\-_\r\n]|(\.\.)")
+                full_pattern = re.compile(r"[^a-zA-Z0-9\u4e00-\u9fa5\-_\r\n]|(\.\.)")
                 knowledge_agent_name = f"{re.sub(full_pattern, '_', self.role)}"
                 if isinstance(self.knowledge_sources, list) and all(
                     isinstance(k, BaseKnowledgeSource) for k in self.knowledge_sources

tests/test_chinese_support/test_chinese_agent.py

@@ -0,0 +1,36 @@
+import re
+import pytest
+from unittest.mock import patch, MagicMock
+from crewai.agent import Agent
+from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource
+
+def test_agent_with_chinese_role_name():
+    """Test that an agent with a Chinese role name works correctly with the updated regex pattern."""
+    # Create a knowledge source with some content
+    content = "This is some test content."
+    string_source = StringKnowledgeSource(content=content)
+    
+    # Mock the Knowledge class to avoid actual initialization
+    with patch("crewai.agent.Knowledge") as MockKnowledge:
+        mock_knowledge_instance = MockKnowledge.return_value
+        
+        # Create an agent with a Chinese role name
+        agent = Agent(
+            role="中文角色",  # Chinese role name
+            goal="Test Chinese character support",
+            backstory="Testing Chinese character support in agent role names.",
+            knowledge_sources=[string_source],
+        )
+        
+        # Call set_knowledge to trigger the regex pattern
+        agent.set_knowledge()
+        
+        # Check that Knowledge was called with the correct collection_name
+        calls = MockKnowledge.call_args_list
+        for call in calls:
+            args, kwargs = call
+            if 'collection_name' in kwargs:
+                collection_name = kwargs['collection_name']
+                print(f"Collection name: {collection_name}")
+                # The collection name should contain the Chinese characters
+                assert "中文角色" == collection_name

tests/test_chinese_support/test_chinese_agent_name.py

@@ -0,0 +1,28 @@
+import pytest
+import re
+from crewai.agent import Agent
+from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource
+
+def test_agent_with_chinese_role_name():
+    """Test that an agent with a Chinese role name works correctly."""
+    # Create a knowledge source with some content
+    content = "This is some test content."
+    string_source = StringKnowledgeSource(content=content)
+    
+    # Create an agent with a Chinese role name
+    agent = Agent(
+        role="中文角色",  # Chinese role name
+        goal="Test Chinese character support",
+        backstory="Testing Chinese character support in agent role names.",
+        knowledge_sources=[string_source],
+    )
+    
+    # Test that the regex pattern in agent.py correctly preserves Chinese characters
+    full_pattern = re.compile(r"[^a-zA-Z0-9\u4e00-\u9fa5\-_\r\n]|(\.\.)")
+    knowledge_agent_name = f"{re.sub(full_pattern, '_', agent.role)}"
+    
+    # Verify that the agent was created successfully
+    assert agent.role == "中文角色"
+    
+    # Verify that the Chinese characters are preserved in the knowledge_agent_name
+    assert knowledge_agent_name == "中文角色"