scai/dist/pipeline/modules/kgModule.js

> **AI-powered CLI for local code analysis, commit message suggestions, and natural-language queries.** 100% local, private, GDPR-friendly, made in Denmark/EU with ❤️.

import { Config } from '../../config.js';
import { generate } from '../../lib/generate.js';
import path from 'path';
import chalk from 'chalk';
import { cleanupModule } from './cleanupModule.js';
import { normalizeData } from '../../utils/normalizeData.js';
export const kgModule = {
    name: 'knowledge-graph',
    description: 'Generates a knowledge graph of entities, tags, and relationships from file content.',
    run: async (input, content) => {
        const model = Config.getModel();
        const ext = input.filepath ? path.extname(input.filepath).toLowerCase() : '';
        const filename = input.filepath ? path.basename(input.filepath) : '';
        // Build prompt for LLM
        const prompt = `
You are an assistant specialized in building knowledge graphs from code or text.

Your task is to extract structured information from the file content below.

File: ${filename}
Extension: ${ext}

📋 Instructions:
- Identify all entities (functions, classes, modules, or main concepts)
- For each entity, generate tags describing its characteristics, purpose, or category
- Identify relationships between entities (e.g., "uses", "extends", "calls")
- Return output in **valid JSON** format with this structure:

{
  "entities": [
    { "name": "EntityName", "type": "class|function|module|concept", "tags": ["tag1", "tag2"] }
  ],
  "edges": [
    { "from": "EntityName1", "to": "EntityName2", "type": "relationship_type" }
  ]
}

⚠️ Make sure all strings are safely JSON-encoded: escape quotes, backslashes, and newlines.

Do NOT include raw content from the file. Only provide structured JSON output.

--- FILE CONTENT START ---
${content}
--- FILE CONTENT END ---
    `.trim();
        try {
            // === INTERACTION #1: generate.ts (via ModuleIO) ===
            const genInput = {
                query: 'Extract entities and relationships for knowledge graph',
                content: prompt,
            };
            const genOutput = await generate(genInput);
            // === INTERACTION #2: cleanupModule (via ModuleIO) ===
            const cleaned = await cleanupModule.run({
                query: 'Clean up JSON for knowledge graph',
                content: genOutput.data,
            });
            // === Normalize and parse JSON result ===
            const parsed = normalizeData(cleaned.data);
            // Ensure structure validity
            parsed.entities ?? (parsed.entities = []);
            parsed.edges ?? (parsed.edges = []);
            return parsed;
        }
        catch (err) {
            console.log(chalk.yellow(`⚠️ [KG] Garbage or invalid JSON output for ${input.filepath}`));
            return { entities: [], edges: [] };
        }
    },
};