Self-hosted Docker Fails to Run Properly with Python SDK in retrieve_related_memory_items
I'm encountering issues when using the Python SDK with a self-hosted Docker deployment (minimal mode). The problem occurs specifically during memory operations.
Environment:
Deployment Mode: Minimal Code Version: b01b3ba1bee164a81c38dac47743b39f474cc47f Issues Details:
Error in memorize -> cluster_memories: When creating a memory task via memorize, the process fails during the cluster_memories method iteration with the error: Failed to parse function arguments niter: line 1 column 2382 (char 2381)
Inconsistent Behavior: Direct API call vs. Python SDK:
Direct API call via curl: The request succeeds (HTTP 200), and the server returns a normal response. However, the created_at and updated_at fields in the response are null, while happened_at contains a valid value. Python SDK call: The server also returns HTTP 200, but the Python SDK throws a Pydantic validation error before returning the result to my code. The error indicates that the related_memories list contains items where the memory.created_at field (and likely updated_at) is None, which does not comply with the expected datetime type in the RelatedMemoryItemsResponse model. The specific validation errors are:
Request validation failed for RelatedMemoryItemsResponse: 4 validation errors: related_memories.0.memory.created_at Input should be a valid datetime [type=datetime_type, input_value=None, input_type=NoneType] ... (similar errors for other indices and likely for updated_at) Summary: The core issue seems to be a data format mismatch. The server's API is returning null for certain datetime fields (created_at, updated_at), which the Python SDK's Pydantic models strictly require to be valid datetime objects. This causes the SDK to fail when processing an otherwise successful (200) response from the server.