Fix #2351: Sanitize collection names to meet ChromaDB requirements

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

src/crewai/agent.py

@@ -142,8 +142,8 @@ class Agent(BaseAgent):
                 self.embedder = crew_embedder
 
             if self.knowledge_sources:
-                full_pattern = re.compile(r"[^a-zA-Z0-9\-_\r\n]|(\.\.)")
-                knowledge_agent_name = f"{re.sub(full_pattern, '_', self.role)}"
+                from crewai.utilities import sanitize_collection_name
+                knowledge_agent_name = sanitize_collection_name(self.role)
                 if isinstance(self.knowledge_sources, list) and all(
                     isinstance(k, BaseKnowledgeSource) for k in self.knowledge_sources
                 ):

src/crewai/utilities/__init__.py

@@ -7,6 +7,7 @@ from .parser import YamlParser
 from .printer import Printer
 from .prompts import Prompts
 from .rpm_controller import RPMController
+from .string_utils import sanitize_collection_name
 from .exceptions.context_window_exceeding_exception import (
     LLMContextLengthExceededException,
 )
@@ -25,4 +26,5 @@ __all__ = [
     "YamlParser",
     "LLMContextLengthExceededException",
     "EmbeddingConfigurator",
+    "sanitize_collection_name",
 ]

src/crewai/utilities/string_utils.py

@@ -80,3 +80,48 @@ def interpolate_only(
             result = result.replace(placeholder, value)
 
     return result
+
+
+from typing import Optional
+
+
+def sanitize_collection_name(name: Optional[str]) -> str:
+    """
+    Sanitize a collection name to meet ChromaDB requirements:
+    1. 3-63 characters long
+    2. Starts and ends with alphanumeric character
+    3. Contains only alphanumeric characters, underscores, or hyphens
+    4. No consecutive periods
+    5. Not a valid IPv4 address
+
+    Args:
+        name: The original collection name to sanitize
+
+    Returns:
+        A sanitized collection name that meets ChromaDB requirements
+    """
+    if not name:
+        return "default_collection"
+
+    # Replace spaces and invalid characters with underscores
+    sanitized = re.sub(r"[^a-zA-Z0-9_-]", "_", name)
+
+    # Ensure it starts with alphanumeric
+    if not sanitized[0].isalnum():
+        sanitized = "a" + sanitized
+
+    # Ensure it ends with alphanumeric
+    if not sanitized[-1].isalnum():
+        sanitized = sanitized[:-1] + "z"
+
+    # Ensure length is between 3-63 characters
+    if len(sanitized) < 3:
+        # Add padding with alphanumeric character at the end
+        sanitized = sanitized + "x" * (3 - len(sanitized))
+    if len(sanitized) > 63:
+        sanitized = sanitized[:63]
+        # Ensure it still ends with alphanumeric after truncation
+        if not sanitized[-1].isalnum():
+            sanitized = sanitized[:-1] + "z"
+
+    return sanitized