Dev 20260319 v2.0.10

.github/PULL_REQUEST_TEMPLATE.md

@@ -2,7 +2,7 @@
 
 Please include a summary of the change, the problem it solves, the implementation approach, and relevant context. List any dependencies required for this change.
 
-Related Issue (Required):  Fixes @issue_number
+Related Issue (Required):  Fixes #issue_number
 
 ## Type of change
 

.github/workflows/python-tests.yml

@@ -102,7 +102,11 @@ jobs:
       if: ${{ !startsWith(matrix.os, 'macos-13') }}
       run: |
         poetry install --no-interaction --extras all
-    - name: PyTest unit tests
+    - name: PyTest unit tests with coverage
       if: ${{ !startsWith(matrix.os, 'macos-13') }}
+      shell: bash
       run: |
-        poetry run pytest tests -vv --durations=10
+        poetry run pytest tests -vv --durations=10 \
+          --cov=src/memos \
+          --cov-report=term-missing \
+          --cov-fail-under=28

.gitignore

@@ -63,6 +63,8 @@ pip-delete-this-directory.txt
 
 # Unit test / coverage reports
 htmlcov/
+report/
+cov-report/
 .tox/
 .nox/
 .coverage

Makefile

@@ -1,4 +1,4 @@
-.PHONY: test
+.PHONY: test test-report test-cov
 
 install:
 	poetry install --extras all --with dev --with test
@@ -9,10 +9,25 @@ clean:
 	rm -rf .pytest_cache
 	rm -rf .ruff_cache
 	rm -rf tmp
+	rm -rf report cov-report
+	rm -f .coverage .coverage.*
 
 test:
 	poetry run pytest tests
 
+test-report:
+	poetry run pytest tests -vv --durations=10 \
+		--html=report/index.html \
+		--cov=src/memos \
+		--cov-report=term-missing \
+		--cov-report=html:cov-report/src
+
+test-cov:
+	poetry run pytest tests \
+		--cov=src/memos \
+		--cov-report=term-missing \
+		--cov-report=html:cov-report/src
+
 format:
 	poetry run ruff check --fix
 	poetry run ruff format

pyproject.toml

@@ -4,7 +4,7 @@
 ##############################################################################
 
 name = "MemoryOS"
-version = "2.0.9"
+version = "2.0.10"
 description = "Intelligence Begins with Memory"
 license = {text = "Apache-2.0"}
 readme = "README.md"
@@ -170,6 +170,8 @@ optional = true
 [tool.poetry.group.test.dependencies]
 pytest = "^8.3.5"
 pytest-asyncio = "^0.23.5"
+pytest-cov = "^6.1"
+pytest-html = "^4.2"
 ruff = "^0.11.8"
 
 [tool.poetry.group.eval]
@@ -208,6 +210,23 @@ filterwarnings = [
 ]
 
 
+[tool.coverage.run]
+source = ["src/memos"]
+branch = true
+
+[tool.coverage.report]
+show_missing = true
+skip_empty = true
+exclude_lines = [
+    "pragma: no cover",
+    "if TYPE_CHECKING:",
+    "if __name__ == .__main__.",
+]
+
+[tool.coverage.html]
+directory = "cov-report"
+
+
 [tool.ruff]
 ##############################################################################
 # Ruff is a fast Python linter and formatter.

src/memos/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "2.0.9"
+__version__ = "2.0.10"
 
 from memos.configs.mem_cube import GeneralMemCubeConfig
 from memos.configs.mem_os import MOSConfig

src/memos/api/config.py

@@ -321,23 +321,40 @@ def get_activation_config() -> dict[str, Any]:
 
     @staticmethod
     def get_memreader_config() -> dict[str, Any]:
-        """Get MemReader configuration for chat/doc extraction (fine-tuned 0.6B model)."""
-        return {
-            "backend": "openai",
-            "config": {
-                "model_name_or_path": os.getenv("MEMRADER_MODEL", "gpt-4o-mini"),
-                "temperature": 0.6,
-                "max_tokens": int(os.getenv("MEMRADER_MAX_TOKENS", "8000")),
-                "top_p": 0.95,
-                "top_k": 20,
-                "api_key": os.getenv("MEMRADER_API_KEY", "EMPTY"),
-                # Default to OpenAI base URL when env var is not provided to satisfy pydantic
-                # validation requirements during tests/import.
-                "api_base": os.getenv("MEMRADER_API_BASE", "https://api.openai.com/v1"),
-                "remove_think_prefix": True,
-            },
+        """Get MemReader configuration for chat/doc extraction (fine-tuned 0.6B model).
+
+        When MEMREADER_GENERAL_MODEL is configured (i.e. a separate stable LLM exists),
+        the backup client is automatically enabled so that primary failures (self-deployed
+        model) fall back to the general LLM.
+        """
+        config = {
+            "model_name_or_path": os.getenv("MEMRADER_MODEL", "gpt-4o-mini"),
+            "temperature": 0.6,
+            "max_tokens": int(os.getenv("MEMRADER_MAX_TOKENS", "8000")),
+            "top_p": 0.95,
+            "top_k": 20,
+            "api_key": os.getenv("MEMRADER_API_KEY", "EMPTY"),
+            # Default to OpenAI base URL when env var is not provided to satisfy pydantic
+            # validation requirements during tests/import.
+            "api_base": os.getenv("MEMRADER_API_BASE", "https://api.openai.com/v1"),
+            "remove_think_prefix": True,
         }
 
+        general_model = os.getenv("MEMREADER_GENERAL_MODEL")
+        enable_backup = os.getenv("MEMREADER_ENABLE_BACKUP", "false").lower() == "true"
+        if general_model and enable_backup:
+            config["backup_client"] = True
+            config["backup_model_name_or_path"] = general_model
+            config["backup_api_key"] = os.getenv(
+                "MEMREADER_GENERAL_API_KEY", os.getenv("OPENAI_API_KEY", "EMPTY")
+            )
+            config["backup_api_base"] = os.getenv(
+                "MEMREADER_GENERAL_API_BASE",
+                os.getenv("OPENAI_API_BASE", "https://api.openai.com/v1"),
+            )
+
+        return {"backend": "openai", "config": config}
+
     @staticmethod
     def get_memreader_general_llm_config() -> dict[str, Any]:
         """Get general LLM configuration for non-chat/doc tasks.
@@ -837,7 +854,7 @@ def get_scheduler_config() -> dict[str, Any]:
                 ),
                 "context_window_size": int(os.getenv("MOS_SCHEDULER_CONTEXT_WINDOW_SIZE", "5")),
                 "thread_pool_max_workers": int(
-                    os.getenv("MOS_SCHEDULER_THREAD_POOL_MAX_WORKERS", "10000")
+                    os.getenv("MOS_SCHEDULER_THREAD_POOL_MAX_WORKERS", "200")
                 ),
                 "consume_interval_seconds": float(
                     os.getenv("MOS_SCHEDULER_CONSUME_INTERVAL_SECONDS", "0.01")
@@ -850,6 +867,8 @@ def get_scheduler_config() -> dict[str, Any]:
                     "MOS_SCHEDULER_ENABLE_ACTIVATION_MEMORY", "false"
                 ).lower()
                 == "true",
+                "use_redis_queue": os.getenv("MEMSCHEDULER_USE_REDIS_QUEUE", "False").lower()
+                == "true",
             },
         }
 

src/memos/api/handlers/component_init.py

@@ -233,7 +233,7 @@ def init_server() -> dict[str, Any]:
     searcher: Searcher = tree_mem.get_searcher(
         manual_close_internet=os.getenv("ENABLE_INTERNET", "true").lower() == "false",
         moscube=False,
-        process_llm=mem_reader.llm,
+        process_llm=mem_reader.general_llm,
     )
     logger.debug("Searcher created")
 
@@ -255,12 +255,13 @@ def init_server() -> dict[str, Any]:
     # Initialize Scheduler
     scheduler_config_dict = APIConfig.get_scheduler_config()
     scheduler_config = SchedulerConfigFactory(
-        backend="optimized_scheduler", config=scheduler_config_dict
+        backend=scheduler_config_dict["backend"],
+        config=scheduler_config_dict["config"],
     )
     mem_scheduler: OptimizedScheduler = SchedulerFactory.from_config(scheduler_config)
     mem_scheduler.initialize_modules(
         chat_llm=llm,
-        process_llm=mem_reader.llm,
+        process_llm=mem_reader.general_llm,
         db_engine=BaseDBManager.create_default_sqlite_engine(),
         mem_reader=mem_reader,
         redis_client=redis_client,

src/memos/configs/llm.py

@@ -28,6 +28,22 @@ class OpenAILLMConfig(BaseLLMConfig):
         default="https://api.openai.com/v1", description="Base URL for OpenAI API"
     )
     extra_body: Any = Field(default=None, description="extra body")
+    backup_client: bool = Field(
+        default=False,
+        description="Whether to enable backup client for fallback on primary failure",
+    )
+    backup_api_key: str | None = Field(
+        default=None, description="API key for backup OpenAI-compatible endpoint"
+    )
+    backup_api_base: str | None = Field(
+        default=None, description="Base URL for backup OpenAI-compatible endpoint"
+    )
+    backup_model_name_or_path: str | None = Field(
+        default=None, description="Model name for backup endpoint"
+    )
+    backup_headers: dict[str, Any] | None = Field(
+        default=None, description="Default headers for backup client requests"
+    )
 
 
 class OpenAIResponsesLLMConfig(BaseLLMConfig):
@@ -42,22 +58,18 @@ class OpenAIResponsesLLMConfig(BaseLLMConfig):
     )
 
 
-class QwenLLMConfig(BaseLLMConfig):
-    api_key: str = Field(..., description="API key for DashScope (Qwen)")
+class QwenLLMConfig(OpenAILLMConfig):
     api_base: str = Field(
         default="https://dashscope-intl.aliyuncs.com/compatible-mode/v1",
         description="Base URL for Qwen OpenAI-compatible API",
     )
-    extra_body: Any = Field(default=None, description="extra body")
 
 
-class DeepSeekLLMConfig(BaseLLMConfig):
-    api_key: str = Field(..., description="API key for DeepSeek")
+class DeepSeekLLMConfig(OpenAILLMConfig):
     api_base: str = Field(
         default="https://api.deepseek.com",
         description="Base URL for DeepSeek OpenAI-compatible API",
     )
-    extra_body: Any = Field(default=None, description="Extra options for API")
 
 
 class AzureLLMConfig(BaseLLMConfig):

src/memos/configs/mem_scheduler.py

@@ -155,7 +155,10 @@ def validate_backend(cls, backend: str) -> str:
     @model_validator(mode="after")
     def create_config(self) -> "SchedulerConfigFactory":
         config_class = self.backend_to_class[self.backend]
-        self.config = config_class(**self.config)
+        raw = self.config
+        if isinstance(raw, dict) and "config" in raw and "use_redis_queue" not in raw:
+            raw = raw["config"]
+        self.config = config_class(**raw)
         return self
 
 

src/memos/graph_dbs/neo4j_community.py

@@ -721,6 +721,8 @@ def get_by_metadata(
         user_name: str | None = None,
         filter: dict | None = None,
         knowledgebase_ids: list[str] | None = None,
+        user_name_flag: bool = True,
+        status: str | None = None,
     ) -> list[str]:
         """
         Retrieve node IDs that match given metadata filters.
@@ -745,15 +747,20 @@ def get_by_metadata(
             - Can be used for faceted recall or prefiltering before embedding rerank.
         """
         logger.info(
-            f"[get_by_metadata] filters: {filters},user_name: {user_name},filter: {filter},knowledgebase_ids: {knowledgebase_ids}"
+            f"[get_by_metadata] filters: {filters},user_name: {user_name},filter: {filter},knowledgebase_ids: {knowledgebase_ids},status: {status}"
         )
         print(
-            f"[get_by_metadata] filters: {filters},user_name: {user_name},filter: {filter},knowledgebase_ids: {knowledgebase_ids}"
+            f"[get_by_metadata] filters: {filters},user_name: {user_name},filter: {filter},knowledgebase_ids: {knowledgebase_ids},status: {status}"
         )
         user_name = user_name if user_name else self.config.user_name
         where_clauses = []
         params = {}
 
+        # Add status filter if provided
+        if status:
+            where_clauses.append("n.status = $status")
+            params["status"] = status
+
         for i, f in enumerate(filters):
             field = f["field"]
             op = f.get("op", "=")