smartmemory-core 0.8.0__tar.gz → 0.9.0__tar.gz

{smartmemory_core-0.8.0 → smartmemory_core-0.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: smartmemory-core
-Version: 0.8.0
+Version: 0.9.0
 Summary: Multi-layered AI memory system with graph databases, vector stores, and intelligent processing pipelines
 Author: SmartMemory Team
 License-Expression: AGPL-3.0-or-later
@@ -37,24 +37,24 @@ Requires-Dist: dspy>=3.0.0; extra == "llm"
 Provides-Extra: ml
 Requires-Dist: scikit-learn>=1.0.0; extra == "ml"
 Provides-Extra: server
-Requires-Dist: smartmemory-core[llm,ml]; extra == "server"
+Requires-Dist: smartmemory-core[llm,ml,rerank]; extra == "server"
 Requires-Dist: falkordb>=1.0.0; extra == "server"
 Requires-Dist: redis>=4.0.0; extra == "server"
 Requires-Dist: requests>=2.28.0; extra == "server"
-Provides-Extra: rebel
-Requires-Dist: transformers>=4.0.0; extra == "rebel"
-Provides-Extra: relik
-Requires-Dist: relik>=0.1.0; extra == "relik"
 Provides-Extra: coreference
 Requires-Dist: fastcoref>=2.1.0; extra == "coreference"
+Provides-Extra: rerank
+Requires-Dist: sentence-transformers>=2.2.0; extra == "rerank"
+Provides-Extra: docs
+Requires-Dist: trafilatura>=1.12.0; extra == "docs"
+Requires-Dist: pymupdf>=1.24.0; extra == "docs"
+Requires-Dist: python-docx>=1.1.0; extra == "docs"
 Provides-Extra: slack
 Requires-Dist: slack-sdk>=3.0.0; extra == "slack"
 Provides-Extra: aws
 Requires-Dist: boto3>=1.26.0; extra == "aws"
 Provides-Extra: wikipedia
 Requires-Dist: wikipedia-api>=0.6.0; extra == "wikipedia"
-Provides-Extra: smolagents
-Requires-Dist: smolagents>=0.1.0; extra == "smolagents"
 Provides-Extra: cli
 Requires-Dist: click>=8.0.0; extra == "cli"
 Requires-Dist: rich>=13.0.0; extra == "cli"
@@ -74,19 +74,32 @@ Requires-Dist: mypy>=1.0.0; extra == "dev"
 Requires-Dist: build>=0.10.0; extra == "dev"
 Requires-Dist: twine>=4.0.0; extra == "dev"
 Provides-Extra: all
-Requires-Dist: smartmemory-core[aws,cli,coreference,lite,rebel,relik,server,slack,smolagents,watch,wikipedia]; extra == "all"
+Requires-Dist: smartmemory-core[aws,cli,coreference,lite,rerank,server,slack,watch,wikipedia]; extra == "all"
 Dynamic: license-file
 
-# SmartMemory - Multi-Layered AI Memory System
+# SmartMemory — Typed AI Memory for LLM Agents
 
 [![Docs](https://img.shields.io/badge/docs-smartmemory.ai-blue)](https://docs.smartmemory.ai/smartmemory/intro)
 [![PyPI version](https://badge.fury.io/py/smartmemory-core.svg)](https://pypi.org/project/smartmemory-core/)
 [![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](https://www.gnu.org/licenses/agpl-3.0)
 [![Python 3.11+](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
 
-**[Read the docs](https://docs.smartmemory.ai/smartmemory/intro)** | **[Maya sample app](https://docs.smartmemory.ai/maya)**
+**[Read the docs](https://docs.smartmemory.ai/smartmemory/intro)** · **[Compare vs Mem0 / Zep / Letta / claude-mem](https://www.smartmemory.ai/compare)** · **[Maya sample app](https://docs.smartmemory.ai/maya)**
 
-SmartMemory is a comprehensive AI memory system that provides persistent, multi-layered memory storage and retrieval for AI applications. It combines graph databases, vector stores, and intelligent processing pipelines to create a unified memory architecture.
+SmartMemory is an AI memory system with **five typed memory types** — episodic, semantic, procedural, working, and zettelkasten — instead of a single vector store. It combines a knowledge graph (FalkorDB), vector embeddings, and an observable 11-stage extraction pipeline so LLM agents can remember, reason over relationships, and update structured knowledge across sessions and tools.
+
+### Who this is for
+
+- **Pick SmartMemory** when you need typed memory that drops into LangChain, LlamaIndex, the Anthropic Agent SDK, or any MCP client; when you want bitemporal accuracy (event time + system time); when multi-tenant isolation must work out of the box; when you want to inspect every stage of extraction.
+- **Pick something else** when you only need a personalization shim for chatbot history (Mem0 is simpler), a complete agent framework (Letta), or zero-config session capture for Claude Code (claude-mem). See the [full comparison](https://www.smartmemory.ai/compare).
+
+### What makes it different
+
+- **5 typed memory types** (not a single blob/vector): episodic events, semantic facts, procedural skills, working context, zettelkasten links.
+- **Knowledge graph by default**, with multi-hop and semantic-hop retrieval — query relationships, not just similarity.
+- **Bitemporal time model** preserved at the storage layer — every fact has both an event time and a system time.
+- **Observable 11-stage pipeline**: classify → coreference → simplify → entity_ruler → llm_extract → ontology_constrain → store → link → enrich → ground → evolve. Per-stage timing, tokens, cost.
+- **Native MCP server**, multi-tenant workspaces, and origin-provenance tiers built in — not bolted on.
 
 ## 🚀 Quick Install
 
@@ -143,7 +156,7 @@ SmartMemory implements a multi-layered memory architecture with the following co
 
 ### Memory Types
 
-- **Working Memory**: Short-term context buffer (in-memory, capacity=10)
+- **Pending Memory**: Short-term buffer for items awaiting consolidation (formerly "working"; routed at ingest by the `ConsolidationRouter`)
 - **Semantic Memory**: Facts and concepts with vector embeddings
 - **Episodic Memory**: Personal experiences and learning history
 - **Procedural Memory**: Skills, strategies, and learned patterns
@@ -172,7 +185,7 @@ Each stage implements the `StageCommand` protocol (`execute(state, config) → s
 
 ## Key Features
 
-- **9 Memory Types**: Working, Semantic, Episodic, Procedural, Zettelkasten, Reasoning, Opinion, Observation, Decision
+- **9 Memory Types**: Pending, Semantic, Episodic, Procedural, Zettelkasten, Reasoning, Opinion, Observation, Decision
 - **11-Stage NLP Pipeline**: classify → coreference → simplify → entity_ruler → llm_extract → ontology_constrain → store → link → enrich → ground → evolve
 - **Self-Learning EntityRuler**: Pattern-matching NER that improves with use — LLM discoveries feed back into rules (96.9% entity F1 at 4ms)
 - **Evolver Framework**: Core auto-registered evolvers plus specialist lifecycle evolvers for decay, consolidation, opinion synthesis, retrieval-based strengthening, Hebbian co-retrieval, and stale memory detection
@@ -180,7 +193,7 @@ Each stage implements the `StageCommand` protocol (`execute(state, config) → s
 - **Zero-Infra Lite Mode**: SQLite + usearch backend — `pip install smartmemory-core[lite]` and go
 - **Server Mode**: FalkorDB graph + Redis caching for production-scale deployments
 - **Hybrid Search**: Graph-structured search + BM25/embedding RRF fusion with query decomposition for compound queries
-- **20 Auto-Registered Plugins**: 4 extractors, 5 enrichers, 10 evolvers, and 1 grounder loaded by default, with additional specialist plugins available for opt-in use
+- **18 Auto-Registered Plugins**: 4 extractors, 5 enrichers, 8 evolvers, and 1 grounder loaded by default, with additional specialist plugins available for opt-in use
 - **Plugin Security**: Sandboxing, permissions, and resource limits for safe plugin execution
 - **Flexible Scoping**: Optional `ScopeProvider` for multi-tenancy or unrestricted OSS usage
 
@@ -273,12 +286,12 @@ from smartmemory import SmartMemory, MemoryItem
 # Initialize SmartMemory
 memory = SmartMemory()
 
-# Add working memory (short-term context)
-working_item = MemoryItem(
+# Add pending memory (short-term context awaiting consolidation)
+pending_item = MemoryItem(
     content="Current conversation context",
-    memory_type="working"
+    memory_type="pending"
 )
-memory.add(working_item)
+memory.add(pending_item)
 
 # Add semantic memory (facts and concepts)
 semantic_item = MemoryItem(
@@ -404,8 +417,6 @@ SmartMemory includes built-in evolvers that automatically transform memories. In
 ### Available Evolvers
 
 **Core evolvers** — memory type transitions and lifecycle:
-- **WorkingToEpisodicEvolver**: Converts working memory to episodic when buffer is full
-- **WorkingToProceduralEvolver**: Extracts repeated patterns as procedures
 - **EpisodicToSemanticEvolver**: Promotes stable facts to semantic memory
 - **EpisodicToZettelEvolver**: Converts episodic events to Zettelkasten notes
 - **EpisodicDecayEvolver**: Archives old episodic memories
@@ -422,7 +433,10 @@ SmartMemory includes built-in evolvers that automatically transform memories. In
 - **RetrievalBasedStrengtheningEvolver**: Memories accessed more frequently become harder to forget
 - **HebbianCoRetrievalEvolver**: Reinforces edges between memories retrieved together ("neurons that fire together wire together")
 - **InterferenceBasedConsolidationEvolver**: Similar competing memories interfere, strengthening the dominant one
-- **EnhancedWorkingToEpisodicEvolver**: Context-aware working→episodic transition with richer metadata
+- **EnhancedWorkingToEpisodicEvolver**: Context-aware pending→episodic transition with richer metadata
+
+**Replaced by ConsolidationRouter (CORE-MEMORY-DYNAMICS-1 M1):**
+- The former `WorkingToEpisodicEvolver` and `WorkingToProceduralEvolver` were retired. Routing from the `pending` bucket to `episodic` / `procedural` now happens at ingest time via the `ConsolidationRouter` pipeline stage.
 
 Evolvers run automatically as part of the memory lifecycle. See the [examples](examples/) directory for evolution demonstrations.
 
@@ -437,7 +451,7 @@ SmartMemory includes **20 auto-registered plugins** with additional specialist p
 **Auto-registered by default** (loaded by `PluginManager`):
 - **4 Extractors**: `LLMExtractor`, `LLMSingleExtractor`, `ConversationAwareLLMExtractor`, `SpacyExtractor`
 - **5 Enrichers**: `BasicEnricher`, `SentimentEnricher`, `TemporalEnricher`, `ExtractSkillsToolsEnricher`, `TopicEnricher`
-- **10 Evolvers**: `WorkingToEpisodicEvolver`, `WorkingToProceduralEvolver`, `EpisodicToSemanticEvolver`, `EpisodicToZettelEvolver`, `EpisodicDecayEvolver`, `SemanticDecayEvolver`, `ZettelPruneEvolver`, `ExponentialDecayEvolver`, `InterferenceBasedConsolidationEvolver`, `RetrievalBasedStrengtheningEvolver`
+- **8 Evolvers**: `EpisodicToSemanticEvolver`, `EpisodicToZettelEvolver`, `EpisodicDecayEvolver`, `SemanticDecayEvolver`, `ZettelPruneEvolver`, `ExponentialDecayEvolver`, `InterferenceBasedConsolidationEvolver`, `RetrievalBasedStrengtheningEvolver`
 - **1 Grounder**: `WikipediaGrounder`
 
 **Specialist plugins** (used by specific pipeline stages or opt-in features):
@@ -550,6 +564,35 @@ PYTHONPATH=. python examples/conversational_assistant_example.py
 
 ## API Reference
 
+### Progress Event Bus (PLAT-PROGRESS-1)
+
+Unified progress emission via Redis Streams. Producers call `emit()`; consumers subscribe over SSE via `/memory/progress/stream` (see `smart-memory-service`). Stream key: `sm:progress:<workspace_id>`; Redis DB 1; MAXLEN default 10_000 (override via `SMARTMEMORY_PROGRESS_MAXLEN`).
+
+```python
+from smartmemory.progress import bind, unbind, bound_progress, emit
+
+# One emit anywhere in the process writes a ProgressEvent to the
+# caller's workspace stream. run_id + scope come from contextvars set
+# by bind() / bound_progress(); never passed as kwargs.
+
+# Simple request-scoped pattern
+bind(run_id="abc-123", scope="workspace:team-42")
+try:
+    emit(kind="pipeline.stage", status="progress", stage="extract",
+         payload={"progress": 50, "message": "Extracting entities"})
+finally:
+    unbind()
+
+# Context-managed pattern — preferred for background tasks that run in
+# a fresh asyncio context (e.g. FastAPI BackgroundTasks). bind()
+# bindings in the scheduling handler don't propagate.
+with bound_progress(run_id="abc-123", scope="workspace:team-42"):
+    emit(kind="pipeline.stage", status="ok", stage="extract",
+         payload={"progress": 100, "message": "Completed extraction"})
+```
+
+Status values: `started | progress | ok | warn | error`. Kinds are dotted producer-owned tags (`pipeline.stage`, `evolver.result`, `studio.job`, …); the transport has no registry.
+
 ### SmartMemory Class
 
 Main interface for memory operations:

{smartmemory_core-0.8.0 → smartmemory_core-0.9.0}/README.md

@@ -1,13 +1,26 @@
-# SmartMemory - Multi-Layered AI Memory System
+# SmartMemory — Typed AI Memory for LLM Agents
 
 [![Docs](https://img.shields.io/badge/docs-smartmemory.ai-blue)](https://docs.smartmemory.ai/smartmemory/intro)
 [![PyPI version](https://badge.fury.io/py/smartmemory-core.svg)](https://pypi.org/project/smartmemory-core/)
 [![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](https://www.gnu.org/licenses/agpl-3.0)
 [![Python 3.11+](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
 
-**[Read the docs](https://docs.smartmemory.ai/smartmemory/intro)** | **[Maya sample app](https://docs.smartmemory.ai/maya)**
+**[Read the docs](https://docs.smartmemory.ai/smartmemory/intro)** · **[Compare vs Mem0 / Zep / Letta / claude-mem](https://www.smartmemory.ai/compare)** · **[Maya sample app](https://docs.smartmemory.ai/maya)**
 
-SmartMemory is a comprehensive AI memory system that provides persistent, multi-layered memory storage and retrieval for AI applications. It combines graph databases, vector stores, and intelligent processing pipelines to create a unified memory architecture.
+SmartMemory is an AI memory system with **five typed memory types** — episodic, semantic, procedural, working, and zettelkasten — instead of a single vector store. It combines a knowledge graph (FalkorDB), vector embeddings, and an observable 11-stage extraction pipeline so LLM agents can remember, reason over relationships, and update structured knowledge across sessions and tools.
+
+### Who this is for
+
+- **Pick SmartMemory** when you need typed memory that drops into LangChain, LlamaIndex, the Anthropic Agent SDK, or any MCP client; when you want bitemporal accuracy (event time + system time); when multi-tenant isolation must work out of the box; when you want to inspect every stage of extraction.
+- **Pick something else** when you only need a personalization shim for chatbot history (Mem0 is simpler), a complete agent framework (Letta), or zero-config session capture for Claude Code (claude-mem). See the [full comparison](https://www.smartmemory.ai/compare).
+
+### What makes it different
+
+- **5 typed memory types** (not a single blob/vector): episodic events, semantic facts, procedural skills, working context, zettelkasten links.
+- **Knowledge graph by default**, with multi-hop and semantic-hop retrieval — query relationships, not just similarity.
+- **Bitemporal time model** preserved at the storage layer — every fact has both an event time and a system time.
+- **Observable 11-stage pipeline**: classify → coreference → simplify → entity_ruler → llm_extract → ontology_constrain → store → link → enrich → ground → evolve. Per-stage timing, tokens, cost.
+- **Native MCP server**, multi-tenant workspaces, and origin-provenance tiers built in — not bolted on.
 
 ## 🚀 Quick Install
 
@@ -64,7 +77,7 @@ SmartMemory implements a multi-layered memory architecture with the following co
 
 ### Memory Types
 
-- **Working Memory**: Short-term context buffer (in-memory, capacity=10)
+- **Pending Memory**: Short-term buffer for items awaiting consolidation (formerly "working"; routed at ingest by the `ConsolidationRouter`)
 - **Semantic Memory**: Facts and concepts with vector embeddings
 - **Episodic Memory**: Personal experiences and learning history
 - **Procedural Memory**: Skills, strategies, and learned patterns
@@ -93,7 +106,7 @@ Each stage implements the `StageCommand` protocol (`execute(state, config) → s
 
 ## Key Features
 
-- **9 Memory Types**: Working, Semantic, Episodic, Procedural, Zettelkasten, Reasoning, Opinion, Observation, Decision
+- **9 Memory Types**: Pending, Semantic, Episodic, Procedural, Zettelkasten, Reasoning, Opinion, Observation, Decision
 - **11-Stage NLP Pipeline**: classify → coreference → simplify → entity_ruler → llm_extract → ontology_constrain → store → link → enrich → ground → evolve
 - **Self-Learning EntityRuler**: Pattern-matching NER that improves with use — LLM discoveries feed back into rules (96.9% entity F1 at 4ms)
 - **Evolver Framework**: Core auto-registered evolvers plus specialist lifecycle evolvers for decay, consolidation, opinion synthesis, retrieval-based strengthening, Hebbian co-retrieval, and stale memory detection
@@ -101,7 +114,7 @@ Each stage implements the `StageCommand` protocol (`execute(state, config) → s
 - **Zero-Infra Lite Mode**: SQLite + usearch backend — `pip install smartmemory-core[lite]` and go
 - **Server Mode**: FalkorDB graph + Redis caching for production-scale deployments
 - **Hybrid Search**: Graph-structured search + BM25/embedding RRF fusion with query decomposition for compound queries
-- **20 Auto-Registered Plugins**: 4 extractors, 5 enrichers, 10 evolvers, and 1 grounder loaded by default, with additional specialist plugins available for opt-in use
+- **18 Auto-Registered Plugins**: 4 extractors, 5 enrichers, 8 evolvers, and 1 grounder loaded by default, with additional specialist plugins available for opt-in use
 - **Plugin Security**: Sandboxing, permissions, and resource limits for safe plugin execution
 - **Flexible Scoping**: Optional `ScopeProvider` for multi-tenancy or unrestricted OSS usage
 
@@ -194,12 +207,12 @@ from smartmemory import SmartMemory, MemoryItem
 # Initialize SmartMemory
 memory = SmartMemory()
 
-# Add working memory (short-term context)
-working_item = MemoryItem(
+# Add pending memory (short-term context awaiting consolidation)
+pending_item = MemoryItem(
     content="Current conversation context",
-    memory_type="working"
+    memory_type="pending"
 )
-memory.add(working_item)
+memory.add(pending_item)
 
 # Add semantic memory (facts and concepts)
 semantic_item = MemoryItem(
@@ -325,8 +338,6 @@ SmartMemory includes built-in evolvers that automatically transform memories. In
 ### Available Evolvers
 
 **Core evolvers** — memory type transitions and lifecycle:
-- **WorkingToEpisodicEvolver**: Converts working memory to episodic when buffer is full
-- **WorkingToProceduralEvolver**: Extracts repeated patterns as procedures
 - **EpisodicToSemanticEvolver**: Promotes stable facts to semantic memory
 - **EpisodicToZettelEvolver**: Converts episodic events to Zettelkasten notes
 - **EpisodicDecayEvolver**: Archives old episodic memories
@@ -343,7 +354,10 @@ SmartMemory includes built-in evolvers that automatically transform memories. In
 - **RetrievalBasedStrengtheningEvolver**: Memories accessed more frequently become harder to forget
 - **HebbianCoRetrievalEvolver**: Reinforces edges between memories retrieved together ("neurons that fire together wire together")
 - **InterferenceBasedConsolidationEvolver**: Similar competing memories interfere, strengthening the dominant one
-- **EnhancedWorkingToEpisodicEvolver**: Context-aware working→episodic transition with richer metadata
+- **EnhancedWorkingToEpisodicEvolver**: Context-aware pending→episodic transition with richer metadata
+
+**Replaced by ConsolidationRouter (CORE-MEMORY-DYNAMICS-1 M1):**
+- The former `WorkingToEpisodicEvolver` and `WorkingToProceduralEvolver` were retired. Routing from the `pending` bucket to `episodic` / `procedural` now happens at ingest time via the `ConsolidationRouter` pipeline stage.
 
 Evolvers run automatically as part of the memory lifecycle. See the [examples](examples/) directory for evolution demonstrations.
 
@@ -358,7 +372,7 @@ SmartMemory includes **20 auto-registered plugins** with additional specialist p
 **Auto-registered by default** (loaded by `PluginManager`):
 - **4 Extractors**: `LLMExtractor`, `LLMSingleExtractor`, `ConversationAwareLLMExtractor`, `SpacyExtractor`
 - **5 Enrichers**: `BasicEnricher`, `SentimentEnricher`, `TemporalEnricher`, `ExtractSkillsToolsEnricher`, `TopicEnricher`
-- **10 Evolvers**: `WorkingToEpisodicEvolver`, `WorkingToProceduralEvolver`, `EpisodicToSemanticEvolver`, `EpisodicToZettelEvolver`, `EpisodicDecayEvolver`, `SemanticDecayEvolver`, `ZettelPruneEvolver`, `ExponentialDecayEvolver`, `InterferenceBasedConsolidationEvolver`, `RetrievalBasedStrengtheningEvolver`
+- **8 Evolvers**: `EpisodicToSemanticEvolver`, `EpisodicToZettelEvolver`, `EpisodicDecayEvolver`, `SemanticDecayEvolver`, `ZettelPruneEvolver`, `ExponentialDecayEvolver`, `InterferenceBasedConsolidationEvolver`, `RetrievalBasedStrengtheningEvolver`
 - **1 Grounder**: `WikipediaGrounder`
 
 **Specialist plugins** (used by specific pipeline stages or opt-in features):
@@ -471,6 +485,35 @@ PYTHONPATH=. python examples/conversational_assistant_example.py
 
 ## API Reference
 
+### Progress Event Bus (PLAT-PROGRESS-1)
+
+Unified progress emission via Redis Streams. Producers call `emit()`; consumers subscribe over SSE via `/memory/progress/stream` (see `smart-memory-service`). Stream key: `sm:progress:<workspace_id>`; Redis DB 1; MAXLEN default 10_000 (override via `SMARTMEMORY_PROGRESS_MAXLEN`).
+
+```python
+from smartmemory.progress import bind, unbind, bound_progress, emit
+
+# One emit anywhere in the process writes a ProgressEvent to the
+# caller's workspace stream. run_id + scope come from contextvars set
+# by bind() / bound_progress(); never passed as kwargs.
+
+# Simple request-scoped pattern
+bind(run_id="abc-123", scope="workspace:team-42")
+try:
+    emit(kind="pipeline.stage", status="progress", stage="extract",
+         payload={"progress": 50, "message": "Extracting entities"})
+finally:
+    unbind()
+
+# Context-managed pattern — preferred for background tasks that run in
+# a fresh asyncio context (e.g. FastAPI BackgroundTasks). bind()
+# bindings in the scheduling handler don't propagate.
+with bound_progress(run_id="abc-123", scope="workspace:team-42"):
+    emit(kind="pipeline.stage", status="ok", stage="extract",
+         payload={"progress": 100, "message": "Completed extraction"})
+```
+
+Status values: `started | progress | ok | warn | error`. Kinds are dotted producer-owned tags (`pipeline.stage`, `evolver.result`, `studio.job`, …); the transport has no registry.
+
 ### SmartMemory Class
 
 Main interface for memory operations:

smartmemory_core-0.9.0/VERSION

@@ -0,0 +1 @@
+0.9.0

{smartmemory_core-0.8.0 → smartmemory_core-0.9.0}/config.json

@@ -3,14 +3,16 @@
     "backend_class": "FalkorDBBackend",
     "host": "localhost",
     "port": 9010,
+    "main_graph_name": "smartmemory",
+    "ontology_graph_name": "ontology",
     "graph_name": "ontology"
   },
   "semantic": {},
   "episodic": {},
   "procedural": {},
-  "working": {},
+  "pending": {},
   "zettelkasten": {},
-  "working_memory": {
+  "pending_memory": {
     "persist": true,
     "vectorize": false,
     "retention_days": 14,
@@ -20,7 +22,7 @@
     "default": "falkordb",
     "type": "falkordb",
     "backend": "falkordb",
-    "dimension": 1536,
+    "dimension": 768,
     "metric": "cosine",
     "hnsw_m": 16,
     "hnsw_ef_construction": 200,
@@ -28,8 +30,8 @@
     "host": "localhost",
     "port": 9010,
     "embedding": {
-      "provider": "openai",
-      "model": "text-embedding-ada-002",
+      "provider": "local",
+      "model": "nomic-ai/nomic-embed-text-v1.5",
       "openai_api_key": "",
       "huggingface_api_key": "",
       "huggingface_model": "sentence-transformers/all-MiniLM-L6-v2"
@@ -41,8 +43,8 @@
       "engine": "basic",
       "ontogpt": {
         "enabled": false,
-        "provider": "openai",
-        "model": "gpt-5-mini",
+        "provider": "groq",
+        "model": "llama-3.3-70b-versatile",
         "api_base": "",
         "api_version": "",
         "template": "",
@@ -52,36 +54,21 @@
     }
   },
   "extractor": {
-    "default": "llm",
-    "gliner2": {
-      "model_name": "fastino/gliner2-base-v1",
-      "confidence_threshold": 0.5
-    },
-    "relik": {
-      "model_name": "relik-ie/relik-relation-extraction-small-wikipedia-ner"
-    },
-    "gliner": {
-      "model_name": "urchade/gliner_base"
-    },
+    "default": "llm_single",
     "llm": {
-      "model_name": "gpt-5-mini",
-      "openai_api_key": ""
     },
     "spacy": {
       "model_name": "en_core_web_sm"
     },
-    "rebel": {
-      "model_name": "Babelscape/rebel-large"
-    },
     "gpt4o_triple": {
-      "model_name": "gpt-5-mini",
+      "model_name": "gpt-4o-mini",
       "openai_api_key": ""
     }
   },
   "enricher": {
     "default": "default",
     "temporal": {
-      "model_name": "gpt-5-mini",
+      "model_name": "gpt-4o-mini",
       "openai_api_key": ""
     }
   },

{smartmemory_core-0.8.0 → smartmemory_core-0.9.0}/pyproject.toml

@@ -61,26 +61,32 @@ ml = [
 
 # Full server mode: FalkorDB graph database + Redis cache/observability + LLM + ML
 server = [
-    "smartmemory-core[llm,ml]",
+    "smartmemory-core[llm,ml,rerank]",
     "falkordb>=1.0.0",
     "redis>=4.0.0",
     "requests>=2.28.0",
 ]
 
-# Optional dependencies for specific extractors
-rebel = ["transformers>=4.0.0"]
-relik = ["relik>=0.1.0"]
-
 # Optional dependencies for pipeline stages
 coreference = ["fastcoref>=2.1.0"]
 
+# Local embedding backend used by EmbeddingPlugin when no API embedder is configured.
+# Also satisfies the CORE-RERANK-1 cross-encoder path when wired.
+rerank = ["sentence-transformers>=2.2.0"]
+
+# Document loaders (lazy — not installed by default; auto-degrade with ImportError if missing)
+docs = [
+    "trafilatura>=1.12.0",
+    "pymupdf>=1.24.0",
+    "python-docx>=1.1.0",
+]
+
 # Optional dependencies for integrations
 # NOTE: claude-cli is a private git dep — install locally with:
 #   pip install "claude-cli @ git+ssh://git@github.com/regression-io/claude-cli.git"
 slack = ["slack-sdk>=3.0.0"]
 aws = ["boto3>=1.26.0"]
 wikipedia = ["wikipedia-api>=0.6.0"]
-smolagents = ["smolagents>=0.1.0"]
 
 # CLI tools
 cli = [
@@ -112,7 +118,7 @@ dev = [
 
 # All optional dependencies
 all = [
-    "smartmemory-core[server,lite,cli,rebel,relik,coreference,slack,aws,wikipedia,smolagents,watch]",
+    "smartmemory-core[server,lite,cli,coreference,rerank,slack,aws,wikipedia,watch]",
 ]
 
 [project.urls]

{smartmemory_core-0.8.0 → smartmemory_core-0.9.0}/scripts/migrate_claude_memory.py

@@ -105,7 +105,7 @@ MEMORIES = [
             "- decompose_query=True on SmartMemory.search() splits compound queries\n"
             "- Splits on and/or/,/; into 1-4 sub-queries\n"
             "- Each sub-query runs independently, results merged via cross-query RRF (rrf_k=60)\n"
-            "- Working memory exemption: memory_type='working' ignores decomposition\n"
+            "- Pending memory exemption: memory_type='pending' ignores decomposition\n"
             "- API surface: decompose: bool = False on service MCP, REST, standalone MCP\n"
             "- Key files: smartmemory/search/query_decomposer.py, rrf_merge.py, smart_memory.py\n"
             "- Contract: docs/features/CORE-SEARCH-1/search-contract.json"

smartmemory_core-0.9.0/smartmemory/activation/__init__.py

@@ -0,0 +1,52 @@
+"""Activation primitives for CORE-MEMORY-DYNAMICS-1 Milestone 2a.
+
+This package provides the activation-scoring foundation used by
+``get_working_context`` ranking and by the three M2a evolvers. The nested
+``metadata["activation"]`` dict shape is defined in the activation contract
+(``smart-memory-docs/docs/features/CORE-MEMORY-DYNAMICS-1/activation-contract.json``).
+
+Public surface:
+
+- ``compute_activation_score(item, now=None)`` — decays score from last boost.
+- ``boost_activation(memory, item_id, delta, reason)`` — Redis-first producer
+  with non-raising direct-write fallback.
+- ``run_decay_sweep(memory)`` — batched per-workspace decay pass.
+- ``migrate_workspace(memory)`` — idempotent Phase A seed (Slice B).
+- Constants: ``ACTIVATION_FLOOR``, ``DEFAULT_SCORE``, ``DEFAULT_HALF_LIFE_HOURS``,
+  Redis key templates, boost sizes.
+
+**Full-dict invariant:** when ``metadata["activation"]`` is present, all five
+sub-keys (score, last_boost_at, half_life_hours, boost_count, rot_score) must
+be present. Partial dicts are a contract violation — see
+``_merged_activation_dict`` in ``boost.py``.
+"""
+
+from smartmemory.activation.constants import (
+    ACTIVATION_FLOOR,
+    BOOST_SIZE_CORETRIEVE,
+    BOOST_SIZE_RETRIEVE,
+    BOOST_SIZE_SURFACE,
+    DECAY_INTERVAL_SECONDS,
+    DEFAULT_HALF_LIFE_HOURS,
+    DEFAULT_SCORE,
+    FLUSH_INTERVAL_SECONDS,
+    NAMESPACE,
+    REDIS_DECAY_LOCK_TEMPLATE,
+    REDIS_PENDING_KEY_TEMPLATE,
+)
+from smartmemory.activation.score import compute_activation_score
+
+__all__ = [
+    "NAMESPACE",
+    "ACTIVATION_FLOOR",
+    "DEFAULT_SCORE",
+    "DEFAULT_HALF_LIFE_HOURS",
+    "BOOST_SIZE_RETRIEVE",
+    "BOOST_SIZE_SURFACE",
+    "BOOST_SIZE_CORETRIEVE",
+    "FLUSH_INTERVAL_SECONDS",
+    "DECAY_INTERVAL_SECONDS",
+    "REDIS_PENDING_KEY_TEMPLATE",
+    "REDIS_DECAY_LOCK_TEMPLATE",
+    "compute_activation_score",
+]