smartmemory-core 0.7.2__tar.gz → 0.7.3__tar.gz

{smartmemory_core-0.7.2 → smartmemory_core-0.7.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: smartmemory-core
-Version: 0.7.2
+Version: 0.7.3
 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

smartmemory_core-0.7.3/VERSION

@@ -0,0 +1 @@
+0.7.3

{smartmemory_core-0.7.2 → smartmemory_core-0.7.3}/smartmemory/memory/ingestion/handlers/__init__.py

@@ -1,4 +1,4 @@
-"""Structured ingestion handlers (CORE-STRUCT-1 Phase 1b/1c, CORE-GAPS-2).
+"""Structured ingestion handlers (CORE-STRUCT-1 Phase 1b/1c/2, CORE-GAPS-2).
 
 Each handler implements the StructuredHandler protocol for a specific schema type.
 """
@@ -6,6 +6,7 @@ Each handler implements the StructuredHandler protocol for a specific schema typ
 from smartmemory.memory.ingestion.handlers.code_entity import CodeEntityHandler
 from smartmemory.memory.ingestion.handlers.conversation_turn import ConversationTurnHandler
 from smartmemory.memory.ingestion.handlers.decision import DecisionHandler
+from smartmemory.memory.ingestion.handlers.document import DocumentHandler
 from smartmemory.memory.ingestion.handlers.hook_capture import HookCaptureHandler
 from smartmemory.memory.ingestion.handlers.plan import PlanHandler
 from smartmemory.memory.ingestion.handlers.plan_task import PlanTaskHandler
@@ -16,6 +17,7 @@ __all__ = [
     "CodeEntityHandler",
     "ConversationTurnHandler",
     "DecisionHandler",
+    "DocumentHandler",
     "HookCaptureHandler",
     "PlanHandler",
     "PlanTaskHandler",

smartmemory_core-0.7.3/smartmemory/memory/ingestion/handlers/document.py

@@ -0,0 +1,80 @@
+"""DocumentHandler — structured ingestion for document content (CORE-STRUCT-1 Phase 2).
+
+Strategy: FULL — documents are user-facing knowledge that needs embedding for search.
+Accepts pre-parsed document content with optional metadata (title, author, source,
+file_type, page_count). Callers handle file I/O and parsing; this handler handles
+storage, embedding, and indexing.
+
+For large documents, callers should chunk content before calling ingest_structured()
+multiple times with the same source_url to link chunks. The existing chunking utilities
+in smartmemory/utils/chunking.py can be used for this.
+"""
+
+from __future__ import annotations
+
+from typing import Optional
+
+from smartmemory.models.ingestion_strategy import IngestionStrategy
+from smartmemory.models.memory_item import MemoryItem
+
+
+class DocumentHandler:
+    """Structured handler for document records."""
+
+    strategy = IngestionStrategy.FULL
+    schema_name = "document"
+    embed: Optional[bool] = None  # defer to FULL default (embed)
+
+    def validate(self, data: dict) -> bool:
+        """Require non-empty content."""
+        content = data.get("content")
+        return isinstance(content, str) and len(content.strip()) > 0
+
+    def to_memory_item(self, data: dict) -> MemoryItem:
+        # Pass through caller metadata first, then overlay known fields (known fields win)
+        meta: dict = {}
+        if data.get("metadata"):
+            meta.update(data["metadata"])
+
+        # Document metadata — set after passthrough so known fields are never overwritten
+        if data.get("title"):
+            meta["title"] = data["title"]
+        if data.get("author"):
+            meta["author"] = data["author"]
+        if data.get("source_url"):
+            meta["source_url"] = data["source_url"]
+        if data.get("file_type"):
+            meta["file_type"] = data["file_type"]
+        if data.get("page_count") is not None:
+            meta["page_count"] = data["page_count"]
+        if data.get("section"):
+            meta["section"] = data["section"]
+        if data.get("chunk_index") is not None:
+            meta["chunk_index"] = data["chunk_index"]
+            meta["total_chunks"] = data.get("total_chunks", 1)
+
+        return MemoryItem(
+            content=data["content"],
+            memory_type="document",
+            metadata=meta,
+            origin=data.get("origin", "structured:document"),
+        )
+
+    def get_edges(self, data: dict, item_id: str) -> list[tuple[str, str, str, dict]]:
+        """Create PART_OF edge if this is a chunk of a larger document."""
+        edges = []
+        parent_id = data.get("parent_document_id")
+        if parent_id:
+            edges.append((
+                item_id,
+                parent_id,
+                "PART_OF",
+                {"origin": "structured:document", "confidence": 1.0},
+            ))
+        return edges
+
+    def get_indexes(self) -> list[str]:
+        return [
+            "CREATE INDEX IF NOT EXISTS idx_document_file_type ON nodes(json_extract(properties, '$.file_type'))",
+            "CREATE INDEX IF NOT EXISTS idx_document_source_url ON nodes(json_extract(properties, '$.source_url'))",
+        ]

{smartmemory_core-0.7.2 → smartmemory_core-0.7.3}/smartmemory/memory/ingestion/structured.py

@@ -75,6 +75,7 @@ class StructuredHandlerRegistry:
             ("code_entity", "smartmemory.memory.ingestion.handlers.code_entity", "CodeEntityHandler"),
             ("plan", "smartmemory.memory.ingestion.handlers.plan", "PlanHandler"),
             ("anchor", "smartmemory.memory.ingestion.handlers.anchor", "AnchorHandler"),
+            ("document", "smartmemory.memory.ingestion.handlers.document", "DocumentHandler"),
         ]
         for schema_name, module_path, class_name in _defaults:
             try:

{smartmemory_core-0.7.2 → smartmemory_core-0.7.3}/smartmemory/models/memory_item.py

@@ -28,6 +28,7 @@ MEMORY_TYPES = {
     "hook_capture",  # APPEND: hook event telemetry (CORE-STRUCT-1 Phase 1b)
     "plan",  # INDEXED: plan containers (CORE-GAPS-2 Phase 1)
     "anchor",  # INDEXED: spec anchors for anti-drift (CORE-GAPS-1)
+    "document",  # FULL: document content with metadata (CORE-STRUCT-1 Phase 2)
 }
 
 

{smartmemory_core-0.7.2 → smartmemory_core-0.7.3}/smartmemory_core.egg-info/SOURCES.txt

@@ -197,6 +197,7 @@ smartmemory/memory/ingestion/handlers/anchor.py
 smartmemory/memory/ingestion/handlers/code_entity.py
 smartmemory/memory/ingestion/handlers/conversation_turn.py
 smartmemory/memory/ingestion/handlers/decision.py
+smartmemory/memory/ingestion/handlers/document.py
 smartmemory/memory/ingestion/handlers/hook_capture.py
 smartmemory/memory/ingestion/handlers/plan.py
 smartmemory/memory/ingestion/handlers/plan_task.py

smartmemory_core-0.7.2/VERSION

@@ -1 +0,0 @@
-0.7.2