Add parameterized tests for diverse Chinese character scenarios

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

src/crewai/agent.py

@@ -140,7 +140,10 @@ class Agent(BaseAgent):
                 self.embedder = crew_embedder
 
             if self.knowledge_sources:
-                full_pattern = re.compile(r"[^a-zA-Z0-9\-_\r\n]|(\.\.)")
+                # Unicode ranges for CJK characters:
+                # \u4e00-\u9fff: Common Chinese characters
+                # \u3400-\u4dbf: Extended CJK characters
+                full_pattern = re.compile(r"[^\w\u4e00-\u9fff\u3400-\u4dbf\-_\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 +1,39 @@
+import re
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+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 +1,31 @@
+import re
+
+import pytest
+
+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"

tests/test_chinese_support/test_mixed_character_support.py

@@ -0,0 +1,48 @@
+import re
+
+import pytest
+
+from crewai.agent import Agent
+from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource
+
+
+@pytest.mark.parametrize(
+    "role_name,expected_output",
+    [
+        ("中文角色", "中文角色"),
+        ("中文角色123", "中文角色123"),
+        ("中文_角色", "中文_角色"),
+        ("测试-角色", "测试-角色"),
+        ("漢字データ", "漢字データ"),
+        ("ABC中文123", "ABC中文123"),
+        ("测试_Test-123", "测试_Test-123"),
+        ("中文 Test Space", "中文_Test_Space"),
+        ("中文角色@#$", "中文角色___"),
+    ],
+)
+def test_mixed_character_support(role_name, expected_output):
+    """Test that various mixed character scenarios work as expected."""
+    # Create a knowledge source with some content
+    content = "This is some test content."
+    string_source = StringKnowledgeSource(content=content)
+
+    # Create an agent with the test role name
+    agent = Agent(
+        role=role_name,
+        goal="Test mixed character support",
+        backstory="Testing mixed character support in agent role names.",
+        knowledge_sources=[string_source],
+    )
+
+    # Test that the regex pattern in agent.py correctly handles the role name
+    # Unicode ranges for CJK characters:
+    # \u4e00-\u9fff: Common Chinese characters
+    # \u3400-\u4dbf: Extended CJK characters
+    full_pattern = re.compile(r"[^\w\u4e00-\u9fff\u3400-\u4dbf\-_\r\n]|(\.\.)")
+    knowledge_agent_name = f"{re.sub(full_pattern, '_', agent.role)}"
+
+    # Verify that the agent was created successfully
+    assert agent.role == role_name
+
+    # Verify that the role name is processed correctly
+    assert knowledge_agent_name == expected_output