296 lines
9.0 KiB
JavaScript
296 lines
9.0 KiB
JavaScript
|
|
const fs = require('fs');
|
||
|
|
const path = require('path');
|
||
|
|
const http = require('http');
|
||
|
|
const https = require('https');
|
||
|
|
const GraphStore = require('./graph.js');
|
||
|
|
const { semanticDiff, formatSummary } = require('./semantic-diff.js');
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Developer Intelligence Pipeline v2 - Phase 6: LLM Doc Generation
|
||
|
|
* Uses semantic diff context to generate targeted, intelligent documentation.
|
||
|
|
* Supports Ollama and OpenAI-compatible APIs.
|
||
|
|
*/
|
||
|
|
|
||
|
|
const LLM_URL = process.env.LLM_URL || 'http://192.168.86.172:11434';
|
||
|
|
const LLM_MODEL = process.env.LLM_MODEL || 'qwen2.5:7b';
|
||
|
|
const LLM_BACKEND = process.env.LLM_BACKEND || 'ollama'; // 'ollama' or 'openai'
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Call LLM API (Ollama or OpenAI-compatible).
|
||
|
|
*/
|
||
|
|
function callLLM(prompt, maxTokens = 1024) {
|
||
|
|
return new Promise((resolve, reject) => {
|
||
|
|
let url, body, headers;
|
||
|
|
|
||
|
|
if (LLM_BACKEND === 'ollama') {
|
||
|
|
url = new URL('/api/generate', LLM_URL);
|
||
|
|
body = JSON.stringify({
|
||
|
|
model: LLM_MODEL,
|
||
|
|
prompt,
|
||
|
|
stream: false,
|
||
|
|
options: { num_predict: maxTokens, temperature: 0.3 },
|
||
|
|
});
|
||
|
|
headers = { 'Content-Type': 'application/json' };
|
||
|
|
} else {
|
||
|
|
url = new URL('/v1/chat/completions', LLM_URL);
|
||
|
|
body = JSON.stringify({
|
||
|
|
model: LLM_MODEL,
|
||
|
|
messages: [{ role: 'user', content: prompt }],
|
||
|
|
max_tokens: maxTokens,
|
||
|
|
temperature: 0.3,
|
||
|
|
});
|
||
|
|
headers = {
|
||
|
|
'Content-Type': 'application/json',
|
||
|
|
'Authorization': `Bearer ${process.env.OPENAI_API_KEY || 'not-needed'}`,
|
||
|
|
};
|
||
|
|
}
|
||
|
|
|
||
|
|
const client = url.protocol === 'https:' ? https : http;
|
||
|
|
const req = client.request(url, { method: 'POST', headers }, (res) => {
|
||
|
|
let data = '';
|
||
|
|
res.on('data', (chunk) => data += chunk);
|
||
|
|
res.on('end', () => {
|
||
|
|
try {
|
||
|
|
const parsed = JSON.parse(data);
|
||
|
|
if (LLM_BACKEND === 'ollama') {
|
||
|
|
resolve(parsed.response || '');
|
||
|
|
} else {
|
||
|
|
resolve(parsed.choices?.[0]?.message?.content || '');
|
||
|
|
}
|
||
|
|
} catch (e) {
|
||
|
|
reject(new Error(`LLM parse error: ${e.message}`));
|
||
|
|
}
|
||
|
|
});
|
||
|
|
});
|
||
|
|
req.on('error', reject);
|
||
|
|
req.setTimeout(60000, () => { req.destroy(); reject(new Error('LLM timeout')); });
|
||
|
|
req.write(body);
|
||
|
|
req.end();
|
||
|
|
});
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Generate documentation for a single entity using graph context.
|
||
|
|
*/
|
||
|
|
async function generateEntityDoc(entityId, graph, sourceCode) {
|
||
|
|
const entity = graph.nodes.get(entityId);
|
||
|
|
if (!entity) return null;
|
||
|
|
|
||
|
|
// Gather context: callers, callees, container
|
||
|
|
const outgoing = graph.edges.filter(e => e.source === entityId);
|
||
|
|
const incoming = graph.edges.filter(e => e.target === entityId);
|
||
|
|
|
||
|
|
const calls = outgoing.filter(e => e.type === 'CALLS').map(e => e.target);
|
||
|
|
const calledBy = incoming.filter(e => e.type === 'CALLS').map(e => e.source);
|
||
|
|
const containedBy = incoming.filter(e => e.type === 'CONTAINS').map(e => e.source);
|
||
|
|
|
||
|
|
// Extract source snippet if available
|
||
|
|
let snippet = '';
|
||
|
|
if (sourceCode && entity.line_range) {
|
||
|
|
const lines = sourceCode.split('\n');
|
||
|
|
const [start, end] = entity.line_range;
|
||
|
|
snippet = lines.slice(start - 1, Math.min(end, start + 50)).join('\n');
|
||
|
|
if (end - start > 50) snippet += '\n// ... truncated';
|
||
|
|
}
|
||
|
|
|
||
|
|
const prompt = `You are a senior engineer writing concise documentation.
|
||
|
|
|
||
|
|
Describe what this ${entity.type.toLowerCase()} does in 2-3 sentences. Be specific about domain logic, not syntax.
|
||
|
|
|
||
|
|
Entity: ${entity.name} (${entity.type}, ${entity.kind})
|
||
|
|
Visibility: ${entity.visibility}
|
||
|
|
Location: ${entityId}
|
||
|
|
${containedBy.length > 0 ? `Part of: ${containedBy.join(', ')}` : ''}
|
||
|
|
${calls.length > 0 ? `Calls: ${calls.slice(0, 10).join(', ')}${calls.length > 10 ? ` (+${calls.length - 10} more)` : ''}` : ''}
|
||
|
|
${calledBy.length > 0 ? `Called by: ${calledBy.slice(0, 10).join(', ')}${calledBy.length > 10 ? ` (+${calledBy.length - 10} more)` : ''}` : ''}
|
||
|
|
|
||
|
|
${snippet ? `Source:\n\`\`\`\n${snippet}\n\`\`\`` : ''}
|
||
|
|
|
||
|
|
Documentation:`;
|
||
|
|
|
||
|
|
return callLLM(prompt);
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Generate a change summary using semantic diff context.
|
||
|
|
*/
|
||
|
|
async function generateDiffDoc(diff) {
|
||
|
|
const summary = formatSummary(diff);
|
||
|
|
|
||
|
|
const prompt = `You are a senior engineer writing a changelog entry for a code review.
|
||
|
|
|
||
|
|
Given this semantic diff, write a concise 3-5 sentence summary suitable for a PR description or release note. Focus on:
|
||
|
|
- What changed and why it matters
|
||
|
|
- Breaking changes that need attention
|
||
|
|
- Impact on downstream consumers
|
||
|
|
|
||
|
|
Semantic Diff:
|
||
|
|
${summary}
|
||
|
|
|
||
|
|
Changelog entry:`;
|
||
|
|
|
||
|
|
return callLLM(prompt, 512);
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Generate docs for all public entities in a file.
|
||
|
|
*/
|
||
|
|
async function generateFileDocs(filePath, graph, repoRoot) {
|
||
|
|
const entityIds = graph.fileIndex.get(filePath);
|
||
|
|
if (!entityIds) return [];
|
||
|
|
|
||
|
|
let sourceCode = '';
|
||
|
|
try {
|
||
|
|
sourceCode = fs.readFileSync(filePath, 'utf8');
|
||
|
|
} catch {}
|
||
|
|
|
||
|
|
const docs = [];
|
||
|
|
for (const id of entityIds) {
|
||
|
|
const entity = graph.nodes.get(id);
|
||
|
|
if (!entity || entity.visibility !== 'public' || entity.type === 'Dependency') continue;
|
||
|
|
|
||
|
|
try {
|
||
|
|
const doc = await generateEntityDoc(id, graph, sourceCode);
|
||
|
|
if (doc) {
|
||
|
|
docs.push({ entityId: id, name: entity.name, type: entity.type, doc });
|
||
|
|
}
|
||
|
|
} catch (err) {
|
||
|
|
console.error(` Failed to generate doc for ${id}: ${err.message}`);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
return docs;
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Batch generate docs for changed entities in a diff.
|
||
|
|
*/
|
||
|
|
async function generateDiffDocs(diff, oldGraph, newGraph, repoRoot) {
|
||
|
|
const results = { changeSummary: '', entityDocs: [] };
|
||
|
|
|
||
|
|
// Generate overall change summary
|
||
|
|
try {
|
||
|
|
results.changeSummary = await generateDiffDoc(diff);
|
||
|
|
} catch (err) {
|
||
|
|
console.error(`Failed to generate change summary: ${err.message}`);
|
||
|
|
}
|
||
|
|
|
||
|
|
// Generate docs for new/modified public entities
|
||
|
|
const entitiesToDoc = [];
|
||
|
|
for (const item of diff.categorized.significant) {
|
||
|
|
if (item.entity && item.entity.visibility === 'public') {
|
||
|
|
entitiesToDoc.push(item.entity.id);
|
||
|
|
}
|
||
|
|
if (item.new && item.new.visibility === 'public') {
|
||
|
|
entitiesToDoc.push(item.new.id);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
for (const id of entitiesToDoc) {
|
||
|
|
const entity = newGraph.nodes.get(id);
|
||
|
|
if (!entity || entity.type === 'Dependency') continue;
|
||
|
|
|
||
|
|
let sourceCode = '';
|
||
|
|
if (entity._file) {
|
||
|
|
try { sourceCode = fs.readFileSync(entity._file, 'utf8'); } catch {}
|
||
|
|
}
|
||
|
|
|
||
|
|
try {
|
||
|
|
const doc = await generateEntityDoc(id, newGraph, sourceCode);
|
||
|
|
if (doc) {
|
||
|
|
results.entityDocs.push({ entityId: id, name: entity.name, type: entity.type, doc });
|
||
|
|
}
|
||
|
|
} catch (err) {
|
||
|
|
console.error(` Failed to generate doc for ${id}: ${err.message}`);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
return results;
|
||
|
|
}
|
||
|
|
|
||
|
|
// --- CLI ---
|
||
|
|
if (require.main === module) {
|
||
|
|
const args = process.argv.slice(2);
|
||
|
|
const command = args[0];
|
||
|
|
|
||
|
|
if (command === 'entity') {
|
||
|
|
const snapshotPath = args[1];
|
||
|
|
const entityId = args[2];
|
||
|
|
const filePath = args[3]; // optional source file
|
||
|
|
|
||
|
|
if (!snapshotPath || !entityId) {
|
||
|
|
console.error('Usage: node docgen.js entity <snapshot.json> <entityId> [source-file]');
|
||
|
|
process.exit(1);
|
||
|
|
}
|
||
|
|
|
||
|
|
const graph = GraphStore.loadSnapshot(snapshotPath);
|
||
|
|
let source = '';
|
||
|
|
if (filePath) {
|
||
|
|
try { source = fs.readFileSync(filePath, 'utf8'); } catch {}
|
||
|
|
}
|
||
|
|
|
||
|
|
generateEntityDoc(entityId, graph, source).then(doc => {
|
||
|
|
console.log(doc);
|
||
|
|
}).catch(err => {
|
||
|
|
console.error(err.message);
|
||
|
|
process.exit(1);
|
||
|
|
});
|
||
|
|
|
||
|
|
} else if (command === 'diff') {
|
||
|
|
const oldPath = args[1];
|
||
|
|
const newPath = args[2];
|
||
|
|
|
||
|
|
if (!oldPath || !newPath) {
|
||
|
|
console.error('Usage: node docgen.js diff <old-snapshot.json> <new-snapshot.json>');
|
||
|
|
process.exit(1);
|
||
|
|
}
|
||
|
|
|
||
|
|
const oldGraph = GraphStore.loadSnapshot(oldPath);
|
||
|
|
const newGraph = GraphStore.loadSnapshot(newPath);
|
||
|
|
const diff = semanticDiff(oldGraph, newGraph);
|
||
|
|
|
||
|
|
generateDiffDocs(diff, oldGraph, newGraph).then(results => {
|
||
|
|
console.log('=== Change Summary ===');
|
||
|
|
console.log(results.changeSummary);
|
||
|
|
console.log('');
|
||
|
|
if (results.entityDocs.length > 0) {
|
||
|
|
console.log('=== Entity Documentation ===');
|
||
|
|
for (const d of results.entityDocs) {
|
||
|
|
console.log(`\n[${d.type}] ${d.name} (${d.entityId})`);
|
||
|
|
console.log(d.doc);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}).catch(err => {
|
||
|
|
console.error(err.message);
|
||
|
|
process.exit(1);
|
||
|
|
});
|
||
|
|
|
||
|
|
} else if (command === 'file') {
|
||
|
|
const snapshotPath = args[1];
|
||
|
|
const filePath = args[2];
|
||
|
|
|
||
|
|
if (!snapshotPath || !filePath) {
|
||
|
|
console.error('Usage: node docgen.js file <snapshot.json> <source-file>');
|
||
|
|
process.exit(1);
|
||
|
|
}
|
||
|
|
|
||
|
|
const graph = GraphStore.loadSnapshot(snapshotPath);
|
||
|
|
generateFileDocs(filePath, graph).then(docs => {
|
||
|
|
for (const d of docs) {
|
||
|
|
console.log(`\n[${d.type}] ${d.name} (${d.entityId})`);
|
||
|
|
console.log(d.doc);
|
||
|
|
}
|
||
|
|
if (docs.length === 0) console.log('No public entities found in file.');
|
||
|
|
}).catch(err => {
|
||
|
|
console.error(err.message);
|
||
|
|
process.exit(1);
|
||
|
|
});
|
||
|
|
|
||
|
|
} else {
|
||
|
|
console.error('Unknown command. Available: entity, diff, file');
|
||
|
|
process.exit(1);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
module.exports = { generateEntityDoc, generateDiffDoc, generateFileDocs, generateDiffDocs, callLLM };
|