Skip to main content
Glama
yodakeisuke

Knowledge Graph Memory Server

by yodakeisuke

read_graph

Retrieve the complete knowledge graph to access stored user information and interaction history for persistent memory in chat applications.

Instructions

Read the entire knowledge graph

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault

No arguments

Implementation Reference

  • The core handler function for the 'read_graph' tool, which loads the knowledge graph from the memory file and returns it.
    async readGraph(): Promise<KnowledgeGraph> { return this.loadGraph(); }
  • index.ts:500-508 (registration)
    Registration of the 'read_graph' tool in the ListToolsRequestSchema response, including name, description, and empty input schema.
    { name: "read_graph", description: "Read the entire knowledge graph", inputSchema: { type: "object", properties: {}, }, }, {
  • Dispatch case in the CallToolRequestSchema handler that executes the read_graph tool by calling knowledgeGraphManager.readGraph() and formatting the response.
    case "read_graph": return createResponse(JSON.stringify(await knowledgeGraphManager.readGraph(), null, 2));
  • Type definition for the KnowledgeGraph returned by the read_graph tool.
    interface KnowledgeGraph { entities: Entity[]; relations: Relation[]; }
  • Helper method called by readGraph to parse the JSON lines from the memory file into entities and relations.
    private async loadGraph(): Promise<KnowledgeGraph> { try { // Check if file exists first try { await fs.access(this.memoryFilePath); } catch (error) { console.error(`[Debug] File does not exist, creating empty file`); await fs.writeFile(this.memoryFilePath, '', 'utf-8'); return { entities: [], relations: [] }; } const data = await fs.readFile(this.memoryFilePath, "utf-8"); console.error(`[Debug] Loading graph from: ${this.memoryFilePath}`); console.error(`[Debug] File contents: ${data}`); const lines = data.split("\n").filter(line => line.trim() !== ""); console.error(`[Debug] Found ${lines.length} lines in the file`); const graph: KnowledgeGraph = { entities: [], relations: [] }; for (const line of lines) { try { console.error(`[Debug] Processing line: ${line}`); const item = JSON.parse(line); if (item.type === "entity") { const entity: Entity = { name: item.name, entityType: item.entityType, observations: item.observations, subdomain: item.subdomain }; console.error(`[Debug] Adding entity: ${JSON.stringify(entity, null, 2)}`); graph.entities.push(entity); } else if (item.type === "relation") { const { type, ...relation } = item; graph.relations.push(relation as Relation); } } catch (parseError) { console.error(`[Debug] Error parsing line: ${parseError}`); continue; } } console.error(`[Debug] Loaded ${graph.entities.length} entities and ${graph.relations.length} relations`); return graph; } catch (error) { console.error(`[Debug] Error loading graph:`, error); throw error; } }

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/yodakeisuke/mcp-memory-domain-knowledge'

If you have feedback or need assistance with the MCP directory API, please join our Discord server