fix: delegate collection name sanitization to knowledge store

src/crewai/agent.py

@@ -142,20 +142,13 @@ class Agent(BaseAgent):
                 self.embedder = crew_embedder
 
             if self.knowledge_sources:
-                try:
-                    from crewai.utilities import sanitize_collection_name
-                    knowledge_agent_name = sanitize_collection_name(self.role)
-                except Exception as e:
-                    self._logger.warning(f"Error sanitizing collection name: {e}")
-                    knowledge_agent_name = "default_agent"
-                    
                 if isinstance(self.knowledge_sources, list) and all(
                     isinstance(k, BaseKnowledgeSource) for k in self.knowledge_sources
                 ):
                     self.knowledge = Knowledge(
                         sources=self.knowledge_sources,
                         embedder=self.embedder,
-                        collection_name=knowledge_agent_name,
+                        collection_name=self.role,
                         storage=self.knowledge_storage or None,
                     )
         except (TypeError, ValueError) as e:

src/crewai/knowledge/storage/knowledge_storage.py

@@ -98,8 +98,11 @@ class KnowledgeStorage(BaseKnowledgeStorage):
                 else "knowledge"
             )
             if self.app:
+                from crewai.utilities.chromadb import sanitize_collection_name
+
                 self.collection = self.app.get_or_create_collection(
-                    name=collection_name, embedding_function=self.embedder
+                    name=sanitize_collection_name(collection_name),
+                    embedding_function=self.embedder,
                 )
             else:
                 raise Exception("Vector Database Client not initialized")

src/crewai/utilities/__init__.py

@@ -7,7 +7,6 @@ from .parser import YamlParser
 from .printer import Printer
 from .prompts import Prompts
 from .rpm_controller import RPMController
-from .string_utils import sanitize_collection_name, is_ipv4_pattern
 from .exceptions.context_window_exceeding_exception import (
     LLMContextLengthExceededException,
 )
@@ -26,6 +25,4 @@ __all__ = [
     "YamlParser",
     "LLMContextLengthExceededException",
     "EmbeddingConfigurator",
-    "sanitize_collection_name",
-    "is_ipv4_pattern",
 ]

src/crewai/utilities/chromadb.py

@@ -0,0 +1,62 @@
+import re
+from typing import Optional
+
+MIN_COLLECTION_LENGTH = 3
+MAX_COLLECTION_LENGTH = 63
+DEFAULT_COLLECTION = "default_collection"
+
+# Compiled regex patterns for better performance
+INVALID_CHARS_PATTERN = re.compile(r"[^a-zA-Z0-9_-]")
+IPV4_PATTERN = re.compile(r"^(\d{1,3}\.){3}\d{1,3}$")
+
+
+def is_ipv4_pattern(name: str) -> bool:
+    """
+    Check if a string matches an IPv4 address pattern.
+
+    Args:
+        name: The string to check
+
+    Returns:
+        True if the string matches an IPv4 pattern, False otherwise
+    """
+    return bool(IPV4_PATTERN.match(name))
+
+
+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
+
+    if is_ipv4_pattern(name):
+        name = f"ip_{name}"
+
+    sanitized = INVALID_CHARS_PATTERN.sub("_", name)
+
+    if not sanitized[0].isalnum():
+        sanitized = "a" + sanitized
+
+    if not sanitized[-1].isalnum():
+        sanitized = sanitized[:-1] + "z"
+
+    if len(sanitized) < MIN_COLLECTION_LENGTH:
+        sanitized = sanitized + "x" * (MIN_COLLECTION_LENGTH - len(sanitized))
+    if len(sanitized) > MAX_COLLECTION_LENGTH:
+        sanitized = sanitized[:MAX_COLLECTION_LENGTH]
+        if not sanitized[-1].isalnum():
+            sanitized = sanitized[:-1] + "z"
+
+    return sanitized

src/crewai/utilities/string_utils.py

@@ -80,74 +80,3 @@ def interpolate_only(
             result = result.replace(placeholder, value)
 
     return result
-
-
-from typing import Optional
-
-# Constants for ChromaDB collection name requirements
-MIN_LENGTH = 3
-MAX_LENGTH = 63
-DEFAULT_COLLECTION = "default_collection"
-
-# Compiled regex patterns for better performance
-INVALID_CHARS_PATTERN = re.compile(r"[^a-zA-Z0-9_-]")
-IPV4_PATTERN = re.compile(r"^(\d{1,3}\.){3}\d{1,3}$")
-
-
-def is_ipv4_pattern(name: str) -> bool:
-    """
-    Check if a string matches an IPv4 address pattern.
-
-    Args:
-        name: The string to check
-
-    Returns:
-        True if the string matches an IPv4 pattern, False otherwise
-    """
-    return bool(IPV4_PATTERN.match(name))
-
-
-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
-
-    # Handle IPv4 pattern
-    if is_ipv4_pattern(name):
-        name = f"ip_{name}"
-
-    # Replace spaces and invalid characters with underscores
-    sanitized = INVALID_CHARS_PATTERN.sub("_", 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 MIN_LENGTH-MAX_LENGTH characters
-    if len(sanitized) < MIN_LENGTH:
-        # Add padding with alphanumeric character at the end
-        sanitized = sanitized + "x" * (MIN_LENGTH - len(sanitized))
-    if len(sanitized) > MAX_LENGTH:
-        sanitized = sanitized[:MAX_LENGTH]
-        # Ensure it still ends with alphanumeric after truncation
-        if not sanitized[-1].isalnum():
-            sanitized = sanitized[:-1] + "z"
-
-    return sanitized