MemoryOS 0.1.13__tar.gz → 0.2.0__tar.gz

{memoryos-0.1.13 → memoryos-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: MemoryOS
-Version: 0.1.13
+Version: 0.2.0
 Summary: Intelligence Begins with Memory
 License: Apache-2.0
 Keywords: memory,llm,language model,memoryOS,agent
@@ -22,6 +22,7 @@ Requires-Dist: ollama (>=0.4.8,<0.5.0)
 Requires-Dist: openai (>=1.77.0,<2.0.0)
 Requires-Dist: qdrant-client (>=1.14.2,<2.0.0)
 Requires-Dist: redis (>=6.2.0,<7.0.0)
+Requires-Dist: schedule (>=1.2.2,<2.0.0)
 Requires-Dist: sentence-transformers (>=4.1.0,<5.0.0)
 Requires-Dist: sqlalchemy (>=2.0.41,<3.0.0)
 Requires-Dist: tenacity (>=9.1.2,<10.0.0)
@@ -31,13 +32,11 @@ Description-Content-Type: text/markdown
 
 <div align="center">
   <a href="https://memos.openmem.net/">
-    <img src="docs/assets/banner_new.gif" alt="MemOS Banner">
+    <img src="https://statics.memtensor.com.cn/memos/memos-banner.gif" alt="MemOS Banner">
   </a>
 
-
-
 <h1 align="center">
-  <img src="docs/assets/memos_logo.png" alt="MemOS Logo" width="50"/> MemOS 1.0: 星河 (Stellar)  <img src="https://img.shields.io/badge/status-Preview-blue" alt="Preview Badge"/>
+  <img src="https://statics.memtensor.com.cn/logo/memos_color_m.png" alt="MemOS Logo" width="50"/> MemOS 1.0: 星河 (Stellar)  <img src="https://img.shields.io/badge/status-Preview-blue" alt="Preview Badge"/>
 </h1>
 
   <p>
@@ -50,7 +49,7 @@ Description-Content-Type: text/markdown
     <a href="https://pypi.org/project/MemoryOS">
       <img src="https://img.shields.io/pypi/pyversions/MemoryOS.svg" alt="Supported Python versions">
     </a>
-    <a href="https://memos.openmem.net/docs/home">
+    <a href="https://memos-docs.openmem.net/home/overview/">
       <img src="https://img.shields.io/badge/Documentation-view-blue.svg" alt="Documentation">
     </a>
     <a href="https://arxiv.org/abs/2507.03724">
@@ -62,7 +61,7 @@ Description-Content-Type: text/markdown
     <a href="https://discord.gg/Txbx3gebZR">
       <img src="https://img.shields.io/badge/Discord-join%20chat-7289DA.svg?logo=discord" alt="Discord">
     </a>
-    <a href="docs/assets/qr_code.png">
+    <a href="https://statics.memtensor.com.cn/memos/qr-code.png">
       <img src="https://img.shields.io/badge/WeChat-Group-07C160.svg?logo=wechat" alt="WeChat Group">
     </a>
     <a href="https://opensource.org/license/apache-2-0/">
@@ -73,17 +72,14 @@ Description-Content-Type: text/markdown
 
 ---
 
-  <a href="https://memos.openmem.net/">
-    <img src="docs/assets/sota_score.jpg" alt="SOTA SCORE">
-  </a>
-
+<img src="https://statics.memtensor.com.cn/memos/sota_score.jpg" alt="SOTA SCORE">
 
 **MemOS** is an operating system for Large Language Models (LLMs) that enhances them with long-term memory capabilities. It allows LLMs to store, retrieve, and manage information, enabling more context-aware, consistent, and personalized interactions.
 
-- **Website**: <a href="https://memos.openmem.net/" target="_blank">https://memos.openmem.net/</a>
-- **Documentation**: <a href="https://memos.openmem.net/docs/home" target="_blank">https://memos.openmem.net/docs/home</a>
-- **API Reference**: <a href="https://memos.openmem.net/docs/api/info" target="_blank">https://memos.openmem.net/docs/api/info</a>
-- **Source Code**: <a href="https://github.com/MemTensor/MemOS" target="_blank">https://github.com/MemTensor/MemOS</a>
+- **Website**: https://memos.openmem.net/
+- **Documentation**: https://memos-docs.openmem.net/home/overview/
+- **API Reference**: https://memos-docs.openmem.net/docs/api/info/
+- **Source Code**: https://github.com/MemTensor/MemOS
 
 ## 📈 Performance Benchmark
 
@@ -97,19 +93,12 @@ MemOS demonstrates significant improvements over baseline memory solutions in mu
 
 > 💡 **Temporal reasoning accuracy improved by 159% compared to the OpenAI baseline.**
 
-
-
 ### Details of End-to-End Evaluation on LOCOMO
 
 > [!NOTE]
 > Comparison of LLM Judge Scores across five major tasks in the LOCOMO benchmark. Each bar shows the mean evaluation score judged by LLMs for a given method-task pair, with standard deviation as error bars. MemOS-0630 consistently outperforms baseline methods (LangMem, Zep, OpenAI, Mem0) across all task types, especially in multi-hop and temporal reasoning scenarios.
 
-<a href="https://memos.openmem.net/">
-  <img src="docs/assets/score_all_end2end.jpg" alt="END2END SCORE">
-</a>
-
-
-
+<img src="https://statics.memtensor.com.cn/memos/score_all_end2end.jpg" alt="END2END SCORE">
 
 ## ✨ Key Features
 
@@ -227,7 +216,7 @@ Join our community to ask questions, share your projects, and connect with other
 - **Discord**: Join our <a href="https://discord.gg/Txbx3gebZR" target="_blank">Discord Server</a>.
 - **WeChat**: Scan the QR code to join our WeChat group.
 
-<img src="docs/assets/qr_code.png" alt="QR Code" width="600">
+<img src="https://statics.memtensor.com.cn/memos/qr-code.png" alt="QR Code" width="600">
 
 ## 📜 Citation
 
@@ -270,7 +259,7 @@ url = {https://global-sci.com/article/91443/memory3-language-modeling-with-expli
 
 ## 🙌 Contributing
 
-We welcome contributions from the community! Please read our [contribution guidelines](https://memos.openmem.net/docs/contribution/overview) to get started.
+We welcome contributions from the community! Please read our [contribution guidelines](https://memos-docs.openmem.net/contribution/overview) to get started.
 
 ## 📄 License
 

{memoryos-0.1.13 → memoryos-0.2.0}/README.md

@@ -1,12 +1,10 @@
 <div align="center">
   <a href="https://memos.openmem.net/">
-    <img src="docs/assets/banner_new.gif" alt="MemOS Banner">
+    <img src="https://statics.memtensor.com.cn/memos/memos-banner.gif" alt="MemOS Banner">
   </a>
 
-
-
 <h1 align="center">
-  <img src="docs/assets/memos_logo.png" alt="MemOS Logo" width="50"/> MemOS 1.0: 星河 (Stellar)  <img src="https://img.shields.io/badge/status-Preview-blue" alt="Preview Badge"/>
+  <img src="https://statics.memtensor.com.cn/logo/memos_color_m.png" alt="MemOS Logo" width="50"/> MemOS 1.0: 星河 (Stellar)  <img src="https://img.shields.io/badge/status-Preview-blue" alt="Preview Badge"/>
 </h1>
 
   <p>
@@ -19,7 +17,7 @@
     <a href="https://pypi.org/project/MemoryOS">
       <img src="https://img.shields.io/pypi/pyversions/MemoryOS.svg" alt="Supported Python versions">
     </a>
-    <a href="https://memos.openmem.net/docs/home">
+    <a href="https://memos-docs.openmem.net/home/overview/">
       <img src="https://img.shields.io/badge/Documentation-view-blue.svg" alt="Documentation">
     </a>
     <a href="https://arxiv.org/abs/2507.03724">
@@ -31,7 +29,7 @@
     <a href="https://discord.gg/Txbx3gebZR">
       <img src="https://img.shields.io/badge/Discord-join%20chat-7289DA.svg?logo=discord" alt="Discord">
     </a>
-    <a href="docs/assets/qr_code.png">
+    <a href="https://statics.memtensor.com.cn/memos/qr-code.png">
       <img src="https://img.shields.io/badge/WeChat-Group-07C160.svg?logo=wechat" alt="WeChat Group">
     </a>
     <a href="https://opensource.org/license/apache-2-0/">
@@ -42,17 +40,14 @@
 
 ---
 
-  <a href="https://memos.openmem.net/">
-    <img src="docs/assets/sota_score.jpg" alt="SOTA SCORE">
-  </a>
-
+<img src="https://statics.memtensor.com.cn/memos/sota_score.jpg" alt="SOTA SCORE">
 
 **MemOS** is an operating system for Large Language Models (LLMs) that enhances them with long-term memory capabilities. It allows LLMs to store, retrieve, and manage information, enabling more context-aware, consistent, and personalized interactions.
 
-- **Website**: <a href="https://memos.openmem.net/" target="_blank">https://memos.openmem.net/</a>
-- **Documentation**: <a href="https://memos.openmem.net/docs/home" target="_blank">https://memos.openmem.net/docs/home</a>
-- **API Reference**: <a href="https://memos.openmem.net/docs/api/info" target="_blank">https://memos.openmem.net/docs/api/info</a>
-- **Source Code**: <a href="https://github.com/MemTensor/MemOS" target="_blank">https://github.com/MemTensor/MemOS</a>
+- **Website**: https://memos.openmem.net/
+- **Documentation**: https://memos-docs.openmem.net/home/overview/
+- **API Reference**: https://memos-docs.openmem.net/docs/api/info/
+- **Source Code**: https://github.com/MemTensor/MemOS
 
 ## 📈 Performance Benchmark
 
@@ -66,19 +61,12 @@ MemOS demonstrates significant improvements over baseline memory solutions in mu
 
 > 💡 **Temporal reasoning accuracy improved by 159% compared to the OpenAI baseline.**
 
-
-
 ### Details of End-to-End Evaluation on LOCOMO
 
 > [!NOTE]
 > Comparison of LLM Judge Scores across five major tasks in the LOCOMO benchmark. Each bar shows the mean evaluation score judged by LLMs for a given method-task pair, with standard deviation as error bars. MemOS-0630 consistently outperforms baseline methods (LangMem, Zep, OpenAI, Mem0) across all task types, especially in multi-hop and temporal reasoning scenarios.
 
-<a href="https://memos.openmem.net/">
-  <img src="docs/assets/score_all_end2end.jpg" alt="END2END SCORE">
-</a>
-
-
-
+<img src="https://statics.memtensor.com.cn/memos/score_all_end2end.jpg" alt="END2END SCORE">
 
 ## ✨ Key Features
 
@@ -196,7 +184,7 @@ Join our community to ask questions, share your projects, and connect with other
 - **Discord**: Join our <a href="https://discord.gg/Txbx3gebZR" target="_blank">Discord Server</a>.
 - **WeChat**: Scan the QR code to join our WeChat group.
 
-<img src="docs/assets/qr_code.png" alt="QR Code" width="600">
+<img src="https://statics.memtensor.com.cn/memos/qr-code.png" alt="QR Code" width="600">
 
 ## 📜 Citation
 
@@ -239,7 +227,7 @@ url = {https://global-sci.com/article/91443/memory3-language-modeling-with-expli
 
 ## 🙌 Contributing
 
-We welcome contributions from the community! Please read our [contribution guidelines](https://memos.openmem.net/docs/contribution/overview) to get started.
+We welcome contributions from the community! Please read our [contribution guidelines](https://memos-docs.openmem.net/contribution/overview) to get started.
 
 ## 📄 License
 

{memoryos-0.1.13 → memoryos-0.2.0}/pyproject.toml

@@ -2,7 +2,7 @@
 
 [tool.poetry]
 name = "MemoryOS"
-version = "0.1.13"
+version = "0.2.0"
 description = "Intelligence Begins with Memory"
 license = "Apache-2.0"
 authors = ["MemTensor <lizy@memtensor.cn>"]
@@ -26,6 +26,7 @@ fastapi = {extras = ["all"], version = "^0.115.12"}
 sentence-transformers = "^4.1.0"
 sqlalchemy = "^2.0.41"
 redis = "^6.2.0"
+schedule = "^1.2.2"
 
 [tool.poetry.group.dev]
 optional = false

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.13"
+__version__ = "0.2.0"
 
 from memos.configs.mem_cube import GeneralMemCubeConfig
 from memos.configs.mem_os import MOSConfig

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/configs/llm.py

@@ -24,6 +24,7 @@ class OpenAILLMConfig(BaseLLMConfig):
     api_base: str = Field(
         default="https://api.openai.com/v1", description="Base URL for OpenAI API"
     )
+    extra_body: Any = Field(default=None, description="extra body")
 
 
 class OllamaLLMConfig(BaseLLMConfig):

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/configs/mem_reader.py

@@ -22,6 +22,10 @@ class BaseMemReaderConfig(BaseConfig):
     chunker: ChunkerConfigFactory = Field(
         ..., description="Chunker configuration for the MemReader"
     )
+    remove_prompt_example: bool = Field(
+        default=False,
+        description="whether remove example in memory extraction prompt to save token",
+    )
 
 
 class SimpleStructMemReaderConfig(BaseMemReaderConfig):

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/configs/memory.py

@@ -161,6 +161,11 @@ class TreeTextMemoryConfig(BaseTextMemoryConfig):
         description="Internet retriever configuration (optional)",
     )
 
+    reorganize: bool | None = Field(
+        False,
+        description="Optional description for this memory configuration.",
+    )
+
 
 # ─── 3. Global Memory Config Factory ──────────────────────────────────────────
 

memoryos-0.2.0/src/memos/graph_dbs/item.py

@@ -0,0 +1,46 @@
+import uuid
+
+from typing import Any, Literal
+
+from pydantic import BaseModel, ConfigDict, Field, field_validator
+
+from memos.memories.textual.item import TextualMemoryItem
+
+
+class GraphDBNode(TextualMemoryItem):
+    pass
+
+
+class GraphDBEdge(BaseModel):
+    """Represents an edge in a graph database (corresponds to Neo4j relationship)."""
+
+    id: str = Field(
+        default_factory=lambda: str(uuid.uuid4()), description="Unique identifier for the edge"
+    )
+    source: str = Field(..., description="Source node ID")
+    target: str = Field(..., description="Target node ID")
+    type: Literal["RELATED", "PARENT"] = Field(
+        ..., description="Relationship type (must be one of 'RELATED', 'PARENT')"
+    )
+    properties: dict[str, Any] | None = Field(
+        default=None, description="Additional properties for the edge"
+    )
+
+    model_config = ConfigDict(extra="forbid")
+
+    @field_validator("id")
+    @classmethod
+    def validate_id(cls, v):
+        """Validate that ID is a valid UUID."""
+        if not isinstance(v, str) or not uuid.UUID(v, version=4):
+            raise ValueError("ID must be a valid UUID string")
+        return v
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> "GraphDBEdge":
+        """Create GraphDBEdge from dictionary."""
+        return cls(**data)
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert to dictionary format."""
+        return self.model_dump(exclude_none=True)

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/graph_dbs/neo4j.py

@@ -92,6 +92,16 @@ class Neo4jGraphDB(BaseGraphDB):
             result = session.run(query, memory_type=memory_type)
             return result.single()["count"]
 
+    def count_nodes(self, scope: str) -> int:
+        query = """
+        MATCH (n:Memory)
+        WHERE n.memory_type = $scope
+        RETURN count(n) AS count
+        """
+        with self.driver.session(database=self.db_name) as session:
+            result = session.run(query, {"scope": scope}).single()
+            return result["count"]
+
     def remove_oldest_memory(self, memory_type: str, keep_latest: int) -> None:
         """
         Remove all WorkingMemory nodes except the latest `keep_latest` entries.
@@ -336,6 +346,49 @@ class Neo4jGraphDB(BaseGraphDB):
         """
         raise NotImplementedError
 
+    def get_neighbors_by_tag(
+        self,
+        tags: list[str],
+        exclude_ids: list[str],
+        top_k: int = 5,
+        min_overlap: int = 1,
+    ) -> list[dict[str, Any]]:
+        """
+        Find top-K neighbor nodes with maximum tag overlap.
+
+        Args:
+            tags: The list of tags to match.
+            exclude_ids: Node IDs to exclude (e.g., local cluster).
+            top_k: Max number of neighbors to return.
+            min_overlap: Minimum number of overlapping tags required.
+
+        Returns:
+            List of dicts with node details and overlap count.
+        """
+        query = """
+            MATCH (n:Memory)
+            WHERE NOT n.id IN $exclude_ids
+            AND n.status = 'activated'
+            AND n.type <> 'reasoning'
+            AND n.memory_type <> 'WorkingMemory'
+            WITH n, [tag IN n.tags WHERE tag IN $tags] AS overlap_tags
+            WHERE size(overlap_tags) >= $min_overlap
+            RETURN n, size(overlap_tags) AS overlap_count
+            ORDER BY overlap_count DESC
+            LIMIT $top_k
+        """
+
+        params = {
+            "tags": tags,
+            "exclude_ids": exclude_ids,
+            "min_overlap": min_overlap,
+            "top_k": top_k,
+        }
+
+        with self.driver.session(database=self.db_name) as session:
+            result = session.run(query, params)
+            return [_parse_node(dict(record["n"])) for record in result]
+
     def get_children_with_embeddings(self, id: str) -> list[str]:
         query = """
         MATCH (p:Memory)-[:PARENT]->(c:Memory)
@@ -386,14 +439,10 @@ class Neo4jGraphDB(BaseGraphDB):
             record = session.run(query, {"center_id": center_id}).single()
 
             if not record:
-                logger.warning(
-                    f"No active node found for center_id={center_id} with status={center_status}"
-                )
                 return {"core_node": None, "neighbors": [], "edges": []}
 
             centers = record["centers"]
             if not centers or centers[0] is None:
-                logger.warning(f"Center node not found or inactive for id={center_id}")
                 return {"core_node": None, "neighbors": [], "edges": []}
 
             core_node = _parse_node(dict(centers[0]))
@@ -730,6 +779,24 @@ class Neo4jGraphDB(BaseGraphDB):
             results = session.run(query, {"scope": scope})
             return [_parse_node(dict(record["n"])) for record in results]
 
+    def get_structure_optimization_candidates(self, scope: str) -> list[dict]:
+        """
+        Find nodes that are likely candidates for structure optimization:
+        - Isolated nodes, nodes with empty background, or nodes with exactly one child.
+        - Plus: the child of any parent node that has exactly one child.
+        """
+        query = """
+                MATCH (n:Memory)
+                WHERE n.memory_type = $scope
+                  AND n.status = 'activated'
+                  AND NOT ( (n)-[:PARENT]->() OR ()-[:PARENT]->(n) )
+                RETURN n.id AS id, n AS node
+                """
+
+        with self.driver.session(database=self.db_name) as session:
+            results = session.run(query, {"scope": scope})
+            return [_parse_node({"id": record["id"], **dict(record["node"])}) for record in results]
+
     def drop_database(self) -> None:
         """
         Permanently delete the entire database this instance is using.
@@ -744,7 +811,7 @@ class Neo4jGraphDB(BaseGraphDB):
 
     def _ensure_database_exists(self):
         with self.driver.session(database="system") as session:
-            session.run(f"CREATE DATABASE {self.db_name} IF NOT EXISTS")
+            session.run(f"CREATE DATABASE $db_name IF NOT EXISTS", db_name=self.db_name)
 
         # Wait until the database is available
         for _ in range(10):

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/llms/openai.py

@@ -22,6 +22,7 @@ class OpenAILLM(BaseLLM):
         response = self.client.chat.completions.create(
             model=self.config.model_name_or_path,
             messages=messages,
+            extra_body=self.config.extra_body,
             temperature=self.config.temperature,
             max_tokens=self.config.max_tokens,
             top_p=self.config.top_p,

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/mem_reader/simple_struct.py

@@ -1,7 +1,7 @@
 import concurrent.futures
 import copy
 import json
-
+import re
 from abc import ABC
 from typing import Any
 
@@ -17,6 +17,7 @@ from memos.parsers.factory import ParserFactory
 from memos.templates.mem_reader_prompts import (
     SIMPLE_STRUCT_DOC_READER_PROMPT,
     SIMPLE_STRUCT_MEM_READER_PROMPT,
+    SIMPLE_STRUCT_MEM_READER_EXAMPLE,
 )
 
 
@@ -39,11 +40,11 @@ class SimpleStructMemReader(BaseMemReader, ABC):
         self.chunker = ChunkerFactory.from_config(config.chunker)
 
     def _process_chat_data(self, scene_data_info, info):
-        prompt = (
-            SIMPLE_STRUCT_MEM_READER_PROMPT.replace("${user_a}", "user")
-            .replace("${user_b}", "assistant")
-            .replace("${conversation}", "\n".join(scene_data_info))
+        prompt = SIMPLE_STRUCT_MEM_READER_PROMPT.replace(
+            "${conversation}", "\n".join(scene_data_info)
         )
+        if self.config.remove_prompt_example:
+            prompt = prompt.replace(SIMPLE_STRUCT_MEM_READER_EXAMPLE, "")
 
         messages = [{"role": "user", "content": prompt}]
 
@@ -228,7 +229,11 @@ class SimpleStructMemReader(BaseMemReader, ABC):
 
     def parse_json_result(self, response_text):
         try:
-            response_text = response_text.replace("```", "").replace("json", "")
+            json_start = response_text.find("{")
+            response_text = response_text[json_start:]
+            response_text = response_text.replace("```", "").strip()
+            if response_text[-1] != "}":
+                response_text += "}"
             response_json = json.loads(response_text)
             return response_json
         except json.JSONDecodeError as e:

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/memories/textual/item.py

@@ -27,7 +27,9 @@ class TextualMemoryMetadata(BaseModel):
         default="activated",
         description="The status of the memory, e.g., 'activated', 'archived', 'deleted'.",
     )
-    type: Literal["procedure", "fact", "event", "opinion", "topic"] | None = Field(default=None)
+    type: Literal["procedure", "fact", "event", "opinion", "topic", "reasoning"] | None = Field(
+        default=None
+    )
     memory_time: str | None = Field(
         default=None,
         description='The time the memory occurred or refers to. Must be in standard `YYYY-MM-DD` format. Relative expressions such as "yesterday" or "tomorrow" are not allowed.',

{memoryos-0.1.13 → memoryos-0.2.0}/src/memos/memories/textual/tree.py

@@ -35,7 +35,11 @@ class TreeTextMemory(BaseTextMemory):
         self.dispatcher_llm: OpenAILLM | OllamaLLM = LLMFactory.from_config(config.dispatcher_llm)
         self.embedder: OllamaEmbedder = EmbedderFactory.from_config(config.embedder)
         self.graph_store: Neo4jGraphDB = GraphStoreFactory.from_config(config.graph_db)
-        self.memory_manager: MemoryManager = MemoryManager(self.graph_store, self.embedder)
+        self.is_reorganize = config.reorganize
+
+        self.memory_manager: MemoryManager = MemoryManager(
+            self.graph_store, self.embedder, self.extractor_llm, is_reorganize=self.is_reorganize
+        )
 
         # Create internet retriever if configured
         self.internet_retriever = None