Fixes ollama and updates openai dependency (#3452)

mem0/llms/ollama.py

@@ -91,6 +91,14 @@ class OllamaLLM(LLMBase):
             "messages": messages,
         }
 
+        # Handle JSON response format by modifying the system prompt
+        if response_format and response_format.get("type") == "json_object":
+            # Add JSON format instruction to the last message or create a system message
+            if messages and messages[-1]["role"] == "user":
+                messages[-1]["content"] += "\n\nPlease respond with valid JSON only."
+            else:
+                messages.append({"role": "user", "content": "Please respond with valid JSON only."})
+
         # Add options for Ollama (temperature, num_predict, top_p)
         options = {
             "temperature": self.config.temperature,

mem0/memory/main.py

@@ -7,6 +7,7 @@ import logging
 import os
 import uuid
 import warnings
+
 from copy import deepcopy
 from datetime import datetime
 from typing import Any, Dict, Optional
@@ -38,6 +39,9 @@ from mem0.utils.factory import (
     VectorStoreFactory,
 )
 
+# Suppress SWIG deprecation warnings globally
+warnings.filterwarnings("ignore", category=DeprecationWarning, message=".*SwigPy.*")
+warnings.filterwarnings("ignore", category=DeprecationWarning, message=".*swigvarlink.*")
 
 def _build_filters_and_metadata(
     *,  # Enforce keyword-only arguments
@@ -386,8 +390,12 @@ class Memory(MemoryBase):
                 response = ""
 
             try:
-                response = remove_code_blocks(response)
+                if not response or not response.strip():
-                new_memories_with_actions = json.loads(response)
+                    logger.warning("Empty response from LLM, no memories to extract")
+                    new_memories_with_actions = {}
+                else:
+                    response = remove_code_blocks(response)
+                    new_memories_with_actions = json.loads(response)
             except Exception as e:
                 logger.error(f"Invalid JSON response: {e}")
                 new_memories_with_actions = {}
@@ -1229,8 +1237,12 @@ class AsyncMemory(MemoryBase):
                 logger.error(f"Error in new memory actions response: {e}")
                 response = ""
             try:
-                response = remove_code_blocks(response)
+                if not response or not response.strip():
-                new_memories_with_actions = json.loads(response)
+                    logger.warning("Empty response from LLM, no memories to extract")
+                    new_memories_with_actions = {}
+                else:
+                    response = remove_code_blocks(response)
+                    new_memories_with_actions = json.loads(response)
             except Exception as e:
                 logger.error(f"Invalid JSON response: {e}")
                 new_memories_with_actions = {}

mem0/vector_stores/faiss.py

@@ -8,7 +8,13 @@ from typing import Dict, List, Optional
 import numpy as np
 from pydantic import BaseModel
 
+import warnings
+
 try:
+    # Suppress SWIG deprecation warnings from FAISS
+    warnings.filterwarnings("ignore", category=DeprecationWarning, message=".*SwigPy.*")
+    warnings.filterwarnings("ignore", category=DeprecationWarning, message=".*swigvarlink.*")
+    
     logging.getLogger("faiss").setLevel(logging.WARNING)
     logging.getLogger("faiss.loader").setLevel(logging.WARNING)
 

pyproject.toml

@@ -14,7 +14,7 @@ requires-python = ">=3.9,<4.0"
 dependencies = [
     "qdrant-client>=1.9.1",
     "pydantic>=2.7.3",
-    "openai>=1.90.0,<1.100.0",
+    "openai>=1.90.0,<1.110.0",
     "posthog>=3.5.0",
     "pytz>=2024.1",
     "sqlalchemy>=2.0.31",
@@ -54,7 +54,7 @@ llms = [
     "groq>=0.3.0",
     "together>=0.2.10",
     "litellm>=1.74.0",
-    "openai>=1.90.0,<1.100.0",
+    "openai>=1.90.0,<1.110.0",
     "ollama>=0.1.0",
     "vertexai>=0.1.0",
     "google-generativeai>=0.3.0",

tests/memory/test_main.py

@@ -66,7 +66,7 @@ class TestAddToVectorStoreErrors:
         mock_memory.llm.generate_response.side_effect = ['{"facts": ["test fact"]}', ""]
 
         # Execute
-        with caplog.at_level(logging.ERROR):
+        with caplog.at_level(logging.WARNING):
             result = mock_memory._add_to_vector_store(
                 messages=[{"role": "user", "content": "test"}], metadata={}, filters={}, infer=True
             )
@@ -74,7 +74,7 @@ class TestAddToVectorStoreErrors:
         # Verify
         assert mock_memory.llm.generate_response.call_count == 2
         assert result == []  # Should return empty list when no memories processed
-        assert "Invalid JSON response" in caplog.text
+        assert "Empty response from LLM, no memories to extract" in caplog.text
 
 
 @pytest.mark.asyncio
@@ -117,11 +117,11 @@ class TestAsyncAddToVectorStoreErrors:
         mock_capture_event = mocker.MagicMock()
         mocker.patch("mem0.memory.main.capture_event", mock_capture_event)
 
-        with caplog.at_level(logging.ERROR):
+        with caplog.at_level(logging.WARNING):
             result = await mock_async_memory._add_to_vector_store(
                 messages=[{"role": "user", "content": "test"}], metadata={}, effective_filters={}, infer=True
             )
 
         assert result == []
-        assert "Invalid JSON response" in caplog.text
+        assert "Empty response from LLM, no memories to extract" in caplog.text
         assert mock_capture_event.call_count == 1