mcp-sqlite-memory-bank 1.4.3__tar.gz → 1.5.1__tar.gz

{mcp_sqlite_memory_bank-1.4.3/src/mcp_sqlite_memory_bank.egg-info → mcp_sqlite_memory_bank-1.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mcp_sqlite_memory_bank
-Version: 1.4.3
+Version: 1.5.1
 Summary: A dynamic, agent/LLM-friendly SQLite memory bank for MCP servers with semantic search capabilities.
 Author-email: Robert Meisner <robert@catchit.pl>
 License-Expression: MIT

{mcp_sqlite_memory_bank-1.4.3 → mcp_sqlite_memory_bank-1.5.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mcp_sqlite_memory_bank"
-version = "1.4.3"
+version = "1.5.1"
 description = "A dynamic, agent/LLM-friendly SQLite memory bank for MCP servers with semantic search capabilities."
 authors = [
     { name="Robert Meisner", email="robert@catchit.pl" }

{mcp_sqlite_memory_bank-1.4.3 → mcp_sqlite_memory_bank-1.5.1}/src/mcp_sqlite_memory_bank/__init__.py

@@ -25,15 +25,25 @@ from .server import (
     update_rows,
     delete_rows,
     run_select_query,
-    # Search tools
-    search_content,
-    explore_tables,
     # FastMCP app
     app,
     # Constants
     DB_PATH,
 )
 
+# Import search tools from the tools module
+from .tools import (
+    # Search tools
+    search_content,
+    explore_tables,
+    add_embeddings,
+    semantic_search,
+    find_related,
+    smart_search,
+    embedding_stats,
+    auto_semantic_search,
+    auto_smart_search,
+)
 
 from .types import (
     # Response types
@@ -81,6 +91,13 @@ __all__ = [
     # Search tools
     "search_content",
     "explore_tables",
+    "add_embeddings",
+    "semantic_search",
+    "find_related", 
+    "smart_search",
+    "embedding_stats",
+    "auto_semantic_search",
+    "auto_smart_search",
     # FastMCP app
     "app",
     # Constants

mcp_sqlite_memory_bank-1.5.1/src/mcp_sqlite_memory_bank/__main__.py

@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+"""
+Entry point for running the SQLite Memory Bank MCP server.
+
+This module provides a clean entry point for MCP clients to start the server
+without import issues or circular dependencies.
+"""
+
+import logging
+import sys
+import os
+
+# Add the project root to Python path to avoid import issues
+project_root = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+if project_root not in sys.path:
+    sys.path.insert(0, project_root)
+
+# Configure logging before any other imports
+logging.basicConfig(
+    level=logging.INFO, 
+    format="%(asctime)s - %(name)s - %(levelname)s - %(message)s"
+)
+
+def main():
+    """Main entry point for the MCP server."""
+    try:
+        # Import here to avoid circular import issues
+        from .server import app, DB_PATH
+        
+        # Handle help argument
+        if "--help" in sys.argv or "-h" in sys.argv:
+            print("SQLite Memory Bank MCP Server")
+            print("Usage: python -m src.mcp_sqlite_memory_bank")
+            print("")
+            print("This starts the SQLite Memory Bank as an MCP (Model Context Protocol) server.")
+            print("The server communicates via STDIO and provides memory management tools")
+            print("for LLMs and AI agents.")
+            print("")
+            print(f"Database location: {DB_PATH}")
+            print("")
+            print("Environment variables:")
+            print("  DB_PATH: Override the default database path")
+            return
+        
+        # Log startup information
+        logging.info(f"Starting SQLite Memory Bank MCP server with database at {DB_PATH}")
+        
+        # Run the FastMCP app in stdio mode for MCP clients
+        app.run(transport="stdio")
+        
+    except KeyboardInterrupt:
+        logging.info("Server stopped by user")
+        sys.exit(0)
+    except Exception as e:
+        logging.error(f"Failed to start MCP server: {e}")
+        sys.exit(1)
+
+if __name__ == "__main__":
+    main()

{mcp_sqlite_memory_bank-1.4.3 → mcp_sqlite_memory_bank-1.5.1}/src/mcp_sqlite_memory_bank/database.py

@@ -403,27 +403,65 @@ class SQLiteMemoryDatabase:
                     for row in conn.execute(stmt).fetchall():
                         row_dict = dict(row._mapping)
 
-                        # Calculate relevance and matched content
-                        relevance = 0.0
+                        # Enhanced relevance calculation with multiple scoring factors
+                        relevance_scores = []
                         matched_content = []
                         query_lower = query.lower()
+                        query_terms = query_lower.split()
 
                         for col in text_columns:
                             if col.name in row_dict and row_dict[col.name]:
                                 content = str(row_dict[col.name]).lower()
+                                content_length = len(content)
+                                
                                 if query_lower in content:
-                                    frequency = content.count(query_lower)
-                                    relevance += frequency / len(content)
-                                    matched_content.append(f"{col.name}: {row_dict[col.name]}")
-
-                        if relevance > 0:
+                                    # Factor 1: Exact phrase frequency (weighted higher)
+                                    exact_frequency = content.count(query_lower)
+                                    exact_score = (exact_frequency * 2.0) / content_length if content_length > 0 else 0
+                                    
+                                    # Factor 2: Individual term frequency
+                                    term_score = 0.0
+                                    for term in query_terms:
+                                        if term in content:
+                                            term_score += content.count(term) / content_length if content_length > 0 else 0
+                                    
+                                    # Factor 3: Position bonus (early matches score higher)
+                                    position_bonus = 0.0
+                                    first_occurrence = content.find(query_lower)
+                                    if first_occurrence != -1:
+                                        position_bonus = (content_length - first_occurrence) / content_length * 0.1
+                                    
+                                    # Factor 4: Column importance (title/name columns get bonus)
+                                    column_bonus = 0.0
+                                    if any(keyword in col.name.lower() for keyword in ['title', 'name', 'summary', 'description']):
+                                        column_bonus = 0.2
+                                    
+                                    # Combined relevance score
+                                    col_relevance = exact_score + term_score + position_bonus + column_bonus
+                                    relevance_scores.append(col_relevance)
+                                    
+                                    # Enhanced matched content with context
+                                    snippet_start = max(0, first_occurrence - 50)
+                                    snippet_end = min(len(row_dict[col.name]), first_occurrence + len(query) + 50)
+                                    snippet = str(row_dict[col.name])[snippet_start:snippet_end]
+                                    if snippet_start > 0:
+                                        snippet = "..." + snippet
+                                    if snippet_end < len(str(row_dict[col.name])):
+                                        snippet = snippet + "..."
+                                    
+                                    matched_content.append(f"{col.name}: {snippet}")
+
+                        total_relevance = sum(relevance_scores)
+                        if total_relevance > 0:
                             results.append(
                                 {
                                     "table": table_name,
                                     "row_id": row_dict.get("id"),
                                     "row_data": row_dict,
                                     "matched_content": matched_content,
-                                    "relevance": round(relevance, 3),
+                                    "relevance": round(total_relevance, 4),
+                                    "match_quality": "high" if total_relevance > 0.5 else "medium" if total_relevance > 0.1 else "low",
+                                    "match_count": len(relevance_scores)
                                 }
                             )
 
@@ -910,10 +948,20 @@ class SQLiteMemoryDatabase:
                 )
 
             # Enhance with text matching scores
-            semantic_engine = get_semantic_engine(model_name)
-            enhanced_results = semantic_engine.hybrid_search(
-                query, semantic_results, text_columns or [], embedding_column, semantic_weight, text_weight, limit
-            )
+            try:
+                semantic_engine = get_semantic_engine(model_name)
+                
+                # Verify the engine has the required method
+                if not hasattr(semantic_engine, 'hybrid_search') or not callable(getattr(semantic_engine, 'hybrid_search')):
+                    raise DatabaseError("Semantic engine hybrid_search method is not callable")
+                
+                enhanced_results = semantic_engine.hybrid_search(
+                    query, semantic_results, text_columns or [], embedding_column, semantic_weight, text_weight, limit
+                )
+            except Exception as e:
+                # If semantic enhancement fails, return semantic results without text enhancement
+                logging.warning(f"Semantic enhancement failed: {e}")
+                enhanced_results = semantic_results[:limit]
 
             return {
                 "success": True,

mcp_sqlite_memory_bank-1.5.1/src/mcp_sqlite_memory_bank/resources.py

@@ -0,0 +1,432 @@
+"""
+MCP Resources Support for SQLite Memory Bank
+==========================================
+
+This module adds MCP Resources support, allowing the memory bank to expose
+stored content as MCP resources that can be consumed by LLM applications.
+
+Resources provide context and data that can be accessed by AI models through
+the standardized MCP protocol.
+
+Author: Robert Meisner
+"""
+
+from typing import Dict, Any, cast
+from fastmcp import FastMCP
+from .database import get_database
+from .semantic import is_semantic_search_available
+import json
+
+
+class MemoryBankResources:
+    """Manages MCP Resources for the SQLite Memory Bank."""
+    
+    def __init__(self, mcp_app: FastMCP, db_path: str):
+        self.mcp = mcp_app
+        self.db_path = db_path
+        self._register_resources()
+    
+    def _register_resources(self):
+        """Register MCP resources with the FastMCP app."""
+        
+        @self.mcp.resource("memory://tables/list")
+        async def get_tables_list() -> str:
+            """Provide a list of all available tables as an MCP resource."""
+            db = get_database(self.db_path)
+            result = cast(Dict[str, Any], db.list_tables())
+            
+            if not result.get("success"):
+                return json.dumps({"error": "Failed to fetch tables", "details": result})
+            
+            resource_content = {
+                "resource_type": "table_list",
+                "description": "List of all available tables in the memory bank",
+                "tables": result.get("tables", []),
+                "total_count": len(result.get("tables", [])),
+                "last_updated": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://tables/{table_name}/schema")
+        async def get_table_schema(table_name: str) -> str:
+            """Provide table schema information as an MCP resource."""
+            db = get_database(self.db_path)
+            result = cast(Dict[str, Any], db.describe_table(table_name))
+            
+            if not result.get("success"):
+                return json.dumps({"error": f"Failed to fetch schema for table '{table_name}'", "details": result})
+            
+            resource_content = {
+                "resource_type": "table_schema",
+                "table_name": table_name,
+                "description": f"Schema definition for table '{table_name}'",
+                "columns": result.get("columns", []),
+                "column_count": len(result.get("columns", [])),
+                "last_updated": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://tables/{table_name}/data")
+        async def get_table_data(table_name: str) -> str:
+            """Provide table data as an MCP resource."""
+            db = get_database(self.db_path)
+            result = cast(Dict[str, Any], db.read_rows(table_name, {}))
+            
+            if not result.get("success"):
+                return json.dumps({"error": f"Failed to fetch data for table '{table_name}'", "details": result})
+            
+            rows = result.get("rows", [])
+            resource_content = {
+                "resource_type": "table_data",
+                "table_name": table_name,
+                "description": f"All data from table '{table_name}'",
+                "rows": rows,
+                "row_count": len(rows),
+                "last_updated": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://search/{query}")
+        async def search_memory_content(query: str) -> str:
+            """Provide search results as an MCP resource."""
+            db = get_database(self.db_path)
+            result = cast(Dict[str, Any], db.search_content(query, None, 50))  # Search all tables, limit to 50 results
+            
+            if not result.get("success"):
+                return json.dumps({"error": f"Failed to search for '{query}'", "details": result})
+            
+            search_results = result.get("results", [])
+            resource_content = {
+                "resource_type": "search_results",
+                "query": query,
+                "description": f"Search results for query: '{query}'",
+                "results": search_results,
+                "result_count": len(search_results),
+                "last_updated": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://analytics/overview")
+        async def get_memory_overview() -> str:
+            """Provide memory bank overview analytics as an MCP resource."""
+            db = get_database(self.db_path)
+            
+            # Get table list
+            tables_result = cast(Dict[str, Any], db.list_tables())
+            if not tables_result.get("success"):
+                return json.dumps({"error": "Failed to fetch memory overview", "details": tables_result})
+            
+            tables = tables_result.get("tables", [])
+            total_rows = 0
+            table_stats = {}
+            
+            # Get row counts for each table
+            for table in tables:
+                try:
+                    rows_result = cast(Dict[str, Any], db.read_rows(table, {}))
+                    if rows_result.get("success"):
+                        row_count = len(rows_result.get("rows", []))
+                        table_stats[table] = {
+                            "row_count": row_count,
+                            "status": "accessible"
+                        }
+                        total_rows += row_count
+                    else:
+                        table_stats[table] = {
+                            "row_count": 0,
+                            "status": "error"
+                        }
+                except Exception as e:
+                    table_stats[table] = {
+                        "row_count": 0,
+                        "status": f"error: {str(e)}"
+                    }
+            
+            # Find largest table
+            largest_table = None
+            if table_stats:
+                max_rows = 0
+                for table_name, stats in table_stats.items():
+                    row_count_obj = stats.get("row_count", 0)
+                    row_count = int(row_count_obj) if isinstance(row_count_obj, (int, str)) else 0
+                    if row_count > max_rows:
+                        max_rows = row_count
+                        largest_table = table_name
+
+            resource_content = {
+                "resource_type": "memory_overview",
+                "description": "Overview of memory bank contents and usage",
+                "summary": {
+                    "total_tables": len(tables),
+                    "total_rows": total_rows,
+                    "largest_table": largest_table
+                },
+                "table_statistics": table_stats,
+                "last_updated": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://live/recent-activity")
+        async def get_recent_activity() -> str:
+            """Real-time feed of recent memory bank changes and activity."""
+            db = get_database(self.db_path)
+            
+            # Get tables with timestamp columns for activity tracking
+            tables_result = cast(Dict[str, Any], db.list_tables())
+            if not tables_result.get("success"):
+                return json.dumps({"error": "Failed to get tables", "details": tables_result})
+            
+            recent_activity = []
+            tables = tables_result.get("tables", [])
+            
+            for table_name in tables:
+                try:
+                    # Check if table has timestamp column
+                    schema_result = cast(Dict[str, Any], db.describe_table(table_name))
+                    if not schema_result.get("success"):
+                        continue
+                    
+                    columns = schema_result.get("columns", [])
+                    timestamp_cols = [col for col in columns if "timestamp" in col.get("name", "").lower()]
+                    
+                    if timestamp_cols:
+                        # Get recent entries (last 10)
+                        recent_result = cast(Dict[str, Any], db.read_rows(table_name, None, 10))
+                        if recent_result.get("success"):
+                            rows = recent_result.get("rows", [])
+                            for row in rows:
+                                activity_entry = {
+                                    "table": table_name,
+                                    "action": "content_added",
+                                    "timestamp": row.get(timestamp_cols[0]["name"]),
+                                    "content_preview": str(row).replace('"', "'")[:100] + "..." if len(str(row)) > 100 else str(row),
+                                    "row_id": row.get("id")
+                                }
+                                recent_activity.append(activity_entry)
+                
+                except Exception as e:
+                    continue
+            
+            # Sort by timestamp (most recent first)
+            recent_activity.sort(key=lambda x: x.get("timestamp", ""), reverse=True)
+            recent_activity = recent_activity[:20]  # Limit to 20 most recent
+            
+            resource_content = {
+                "resource_type": "recent_activity",
+                "description": "Recent changes and additions to the memory bank",
+                "activities": recent_activity,
+                "activity_count": len(recent_activity),
+                "last_updated": "real-time",
+                "refresh_rate": "dynamic"
+            }
+            
+            return json.dumps(resource_content, indent=2)
+        
+        @self.mcp.resource("memory://live/content-suggestions")
+        async def get_content_suggestions() -> str:
+            """AI-powered suggestions for content improvements and organization."""
+            db = get_database(self.db_path)
+            
+            suggestions = {
+                "organization_suggestions": [],
+                "content_gaps": [],
+                "semantic_opportunities": [],
+                "quality_improvements": []
+            }
+            
+            try:
+                # Get basic analysis
+                tables_result = cast(Dict[str, Any], db.list_tables())
+                if not tables_result.get("success"):
+                    return json.dumps({"error": "Failed to analyze content", "details": tables_result})
+                
+                tables = tables_result.get("tables", [])
+                
+                for table_name in tables:
+                    try:
+                        # Analyze table content
+                        rows_result = cast(Dict[str, Any], db.read_rows(table_name))
+                        if not rows_result.get("success"):
+                            continue
+                        
+                        rows = rows_result.get("rows", [])
+                        
+                        # Check for organization opportunities
+                        if len(rows) > 50:
+                            suggestions["organization_suggestions"].append({
+                                "table": table_name,
+                                "suggestion": "Consider adding categories or tags for better organization",
+                                "reason": f"Large table with {len(rows)} rows could benefit from categorization"
+                            })
+                        
+                        # Check for semantic search opportunities
+                        if is_semantic_search_available():
+                            embedding_stats = cast(Dict[str, Any], db.get_embedding_stats(table_name))
+                            if embedding_stats.get("success") and embedding_stats.get("coverage_percent", 0) == 0:
+                                schema_result = cast(Dict[str, Any], db.describe_table(table_name))
+                                if schema_result.get("success"):
+                                    text_cols = [col for col in schema_result.get("columns", []) 
+                                               if "TEXT" in col.get("type", "").upper()]
+                                    if text_cols and len(rows) > 5:
+                                        suggestions["semantic_opportunities"].append({
+                                            "table": table_name,
+                                            "suggestion": "Set up semantic search for better content discovery",
+                                            "reason": f"Table has {len(text_cols)} text columns and {len(rows)} rows",
+                                            "action": f"Use add_embeddings('{table_name}', {[col['name'] for col in text_cols[:3]]})"
+                                        })
+                        
+                        # Check for content gaps (sparse tables)
+                        if 1 <= len(rows) <= 5:
+                            suggestions["content_gaps"].append({
+                                "table": table_name,
+                                "suggestion": "Consider adding more content or consolidating with other tables",
+                                "reason": f"Table has only {len(rows)} rows - might be underutilized"
+                            })
+                        
+                        # Sample content for quality analysis
+                        if rows:
+                            sample_row = rows[0]
+                            short_values = [k for k, v in sample_row.items() 
+                                          if isinstance(v, str) and 0 < len(v) < 10]
+                            if len(short_values) > 2:
+                                suggestions["quality_improvements"].append({
+                                    "table": table_name,
+                                    "suggestion": "Consider adding more detailed content",
+                                    "reason": f"Several columns have very short values: {short_values[:3]}"
+                                })
+                    
+                    except Exception as e:
+                        continue
+                
+                # Prioritize suggestions
+                priority_order = ["semantic_opportunities", "organization_suggestions", "quality_improvements", "content_gaps"]
+                prioritized = {}
+                for category in priority_order:
+                    if suggestions[category]:
+                        prioritized[category] = suggestions[category]
+                
+                resource_content = {
+                    "resource_type": "content_suggestions",
+                    "description": "AI-powered suggestions for improving your memory bank",
+                    "suggestions": prioritized,
+                    "total_suggestions": sum(len(v) for v in suggestions.values()),
+                    "last_updated": "real-time",
+                    "next_actions": [
+                        "Review semantic opportunities for high-value tables",
+                        "Consider organization improvements for large tables",
+                        "Add more detailed content where suggested"
+                    ]
+                }
+                
+                return json.dumps(resource_content, indent=2)
+                
+            except Exception as e:
+                return json.dumps({
+                    "error": f"Failed to generate content suggestions: {str(e)}",
+                    "suggestions": suggestions
+                })
+        
+        @self.mcp.resource("memory://analytics/insights")
+        async def get_memory_insights() -> str:
+            """Real-time analytics and insights about memory bank usage and patterns."""
+            db = get_database(self.db_path)
+            
+            insights = {
+                "usage_patterns": {},
+                "content_trends": {},
+                "search_recommendations": {},
+                "health_indicators": {}
+            }
+            
+            try:
+                tables_result = cast(Dict[str, Any], db.list_tables())
+                if not tables_result.get("success"):
+                    return json.dumps({"error": "Failed to get insights", "details": tables_result})
+                
+                tables = tables_result.get("tables", [])
+                total_rows = 0
+                content_quality_scores = []
+                
+                for table_name in tables:
+                    rows_result = cast(Dict[str, Any], db.read_rows(table_name))
+                    if rows_result.get("success"):
+                        rows = rows_result.get("rows", [])
+                        row_count = len(rows)
+                        total_rows += row_count
+                        
+                        # Calculate content quality score for this table
+                        if rows:
+                            # Sample content to estimate quality
+                            sample_size = min(5, len(rows))
+                            total_content_length = 0
+                            for row in rows[:sample_size]:
+                                for value in row.values():
+                                    if isinstance(value, str):
+                                        total_content_length += len(value)
+                            
+                            avg_content_length = total_content_length / sample_size if sample_size > 0 else 0
+                            quality_score = min(10, avg_content_length / 50)  # Normalize to 0-10
+                            content_quality_scores.append(quality_score)
+                            
+                            insights["usage_patterns"][table_name] = {
+                                "row_count": row_count,
+                                "avg_content_length": round(avg_content_length),
+                                "quality_score": round(quality_score, 1),
+                                "category": "high_value" if quality_score > 7 else "medium_value" if quality_score > 3 else "low_value"
+                            }
+                
+                # Overall health indicators
+                avg_quality = sum(content_quality_scores) / len(content_quality_scores) if content_quality_scores else 0
+                insights["health_indicators"] = {
+                    "total_tables": len(tables),
+                    "total_content_rows": total_rows,
+                    "average_content_quality": round(avg_quality, 1),
+                    "content_distribution": "balanced" if len(tables) > 0 and total_rows / len(tables) > 10 else "sparse",
+                    "semantic_readiness": "available" if is_semantic_search_available() else "unavailable"
+                }
+                
+                # Search recommendations
+                high_value_tables = [name for name, data in insights["usage_patterns"].items() 
+                                   if data.get("category") == "high_value"]
+                
+                if high_value_tables:
+                    insights["search_recommendations"]["intelligent_search"] = {
+                        "recommended_tables": high_value_tables,
+                        "strategy": "Use intelligent_search() for best results across high-value content"
+                    }
+                
+                if is_semantic_search_available():
+                    insights["search_recommendations"]["semantic_opportunities"] = {
+                        "suggestion": "Consider semantic search for conceptual queries",
+                        "best_for": "Finding related concepts, patterns, and thematic content"
+                    }
+                
+                resource_content = {
+                    "resource_type": "memory_insights",
+                    "description": "Real-time analytics and insights about your memory bank",
+                    "insights": insights,
+                    "last_updated": "real-time",
+                    "recommendations": [
+                        f"Focus on high-value tables: {', '.join(high_value_tables[:3])}" if high_value_tables else "Add more detailed content to improve value",
+                        "Use intelligent_search() for optimal search results",
+                        "Consider semantic search setup for better content discovery" if is_semantic_search_available() else "Install sentence-transformers for semantic search"
+                    ]
+                }
+                
+                return json.dumps(resource_content, indent=2)
+                
+            except Exception as e:
+                return json.dumps({
+                    "error": f"Failed to generate insights: {str(e)}",
+                    "insights": insights
+                })
+
+
+def setup_mcp_resources(mcp_app: FastMCP, db_path: str) -> MemoryBankResources:
+    """Set up MCP Resources for the memory bank."""
+    return MemoryBankResources(mcp_app, db_path)