Merge remote-tracking branch 'upstream/dev-20260302-v2.0.8' into feat/transfer_pref_to_polar_db

Author: yuan.wang

src/memos/configs/mem_scheduler.py

@@ -250,8 +250,12 @@ def validate_partial_initialization(self) -> "AuthConfig":
                 "All configuration components are None. This may indicate missing environment variables or configuration files."
             )
         elif failed_components:
-            logger.warning(
-                f"Failed to initialize components: {', '.join(failed_components)}. Successfully initialized: {', '.join(initialized_components)}"
+            # Use info level: individual from_local_env() methods already log
+            # warnings for actual initialization failures. Components that are
+            # simply not configured (no env vars) are not errors.
+            logger.info(
+                f"Components not configured: {', '.join(failed_components)}. "
+                f"Successfully initialized: {', '.join(initialized_components)}"
             )
 
         return self

src/memos/embedders/universal_api.py

@@ -14,6 +14,21 @@
 logger = get_logger(__name__)
 
 
+def _sanitize_unicode(text: str) -> str:
+    """
+    Remove Unicode surrogates and other problematic characters.
+    Surrogates (U+D800-U+DFFF) cause UnicodeEncodeError with some APIs.
+    """
+    try:
+        # Encode with 'surrogatepass' then decode, replacing invalid chars
+        cleaned = text.encode("utf-8", errors="surrogatepass").decode("utf-8", errors="replace")
+        # Replace replacement char with empty string for cleaner output
+        return cleaned.replace("\ufffd", "")
+    except Exception:
+        # Fallback: remove all non-BMP characters
+        return "".join(c for c in text if ord(c) < 0x10000)
+
+
 class UniversalAPIEmbedder(BaseEmbedder):
     def __init__(self, config: UniversalAPIEmbedderConfig):
         self.provider = config.provider
@@ -54,6 +69,8 @@ def __init__(self, config: UniversalAPIEmbedderConfig):
     def embed(self, texts: list[str]) -> list[list[float]]:
         if isinstance(texts, str):
             texts = [texts]
+        # Sanitize Unicode to prevent encoding errors with emoji/surrogates
+        texts = [_sanitize_unicode(t) for t in texts]
         # Truncate texts if max_tokens is configured
         texts = self._truncate_texts(texts)
         logger.info(f"Embeddings request with input: {texts}")

src/memos/mem_os/utils/default_config.py

@@ -3,12 +3,15 @@
 Provides simplified configuration generation for users.
 """
 
+import logging
 from typing import Literal
 
 from memos.configs.mem_cube import GeneralMemCubeConfig
 from memos.configs.mem_os import MOSConfig
 from memos.mem_cube.general import GeneralMemCube
 
+logger = logging.getLogger(__name__)
+
 
 def get_default_config(
     openai_api_key: str,
@@ -116,20 +119,9 @@ def get_default_config(
             },
         }
 
-    # Add activation memory if enabled
-    if config_dict.get("enable_activation_memory", False):
-        config_dict["act_mem"] = {
-            "backend": "kv_cache",
-            "config": {
-                "memory_filename": kwargs.get(
-                    "activation_memory_filename", "activation_memory.pickle"
-                ),
-                "extractor_llm": {
-                    "backend": "openai",
-                    "config": openai_config,
-                },
-            },
-        }
+    # Note: act_mem configuration belongs in MemCube config (get_default_cube_config),
+    # not in MOSConfig which doesn't have an act_mem field (extra="forbid").
+    # The enable_activation_memory flag above is sufficient for MOSConfig.
 
     return MOSConfig(**config_dict)
 
@@ -237,21 +229,33 @@ def get_default_cube_config(
             },
         }
 
-    # Configure activation memory if enabled
+    # Configure activation memory if enabled.
+    # KV cache activation memory requires a local HuggingFace/vLLM model (it
+    # extracts internal attention KV tensors via build_kv_cache), so it cannot
+    # work with remote API backends like OpenAI.
+    # Only create act_mem when activation_memory_backend is explicitly provided.
     act_mem_config = {}
     if kwargs.get("enable_activation_memory", False):
-        act_mem_config = {
-            "backend": "kv_cache",
-            "config": {
-                "memory_filename": kwargs.get(
-                    "activation_memory_filename", "activation_memory.pickle"
-                ),
-                "extractor_llm": {
-                    "backend": "openai",
-                    "config": openai_config,
+        extractor_backend = kwargs.get("activation_memory_backend")
+        if extractor_backend in ("huggingface", "huggingface_singleton", "vllm"):
+            act_mem_config = {
+                "backend": "kv_cache",
+                "config": {
+                    "memory_filename": kwargs.get(
+                        "activation_memory_filename", "activation_memory.pickle"
+                    ),
+                    "extractor_llm": {
+                        "backend": extractor_backend,
+                        "config": kwargs.get("activation_memory_llm_config", {}),
+                    },
                 },
-            },
-        }
+            }
+        else:
+            logger.info(
+                "Activation memory (kv_cache) requires a local model backend "
+                "(huggingface/vllm) via activation_memory_backend kwarg. "
+                "Skipping act_mem in MemCube config."
+            )
 
     # Create MemCube configuration
     cube_config_dict = {