Vibe Coding Documentation MCP (MUSE)

import { CodeBlock, CodeContext } from '../types/index.js'; import { generateId, getCurrentTimestamp, extractCodeBlocks } from '../utils/markdown.js'; import { hashCode } from '../core/cache.js'; export interface CollectCodeContextInput { codeBlocks?: CodeBlock[]; rawText?: string; conversationSummary: string; tags?: string[]; autoDetectLanguage?: boolean; removeDuplicates?: boolean; includeStats?: boolean; } export interface CodeStats { totalBlocks: number; totalLines: number; languageBreakdown: Record<string, number>; averageLinesPerBlock: number; } export interface CollectCodeContextOutput { context: CodeContext; stats?: CodeStats; warnings?: string[]; } // 언어 감지를 위한 패턴 const languagePatterns: Record<string, RegExp[]> = { typescript: [ /^import\s+.*\s+from\s+['"].*['"];?$/m, /:\s*(string|number|boolean|any|void|never)\b/, /interface\s+\w+/, /type\s+\w+\s*=/, /<\w+>/ ], javascript: [ /^const\s+\w+\s*=/m, /^let\s+\w+\s*=/m, /^function\s+\w+\s*\(/m, /=>\s*{/, /require\s*\(/ ], python: [ /^def\s+\w+\s*\(/m, /^class\s+\w+.*:/m, /^import\s+\w+$/m, /^from\s+\w+\s+import/m, /:\s*$/m ], java: [ /public\s+(class|interface|enum)/, /private\s+\w+\s+\w+;/, /System\.out\.println/, /@Override/ ], go: [ /^package\s+\w+$/m, /^func\s+\w+\s*\(/m, /^import\s+\(/m, /:=\s*/ ], rust: [ /^fn\s+\w+/m, /^use\s+\w+/m, /^let\s+mut\s+/m, /impl\s+\w+/, /->.*{$/m ], sql: [ /^SELECT\s+/im, /^INSERT\s+INTO/im, /^CREATE\s+TABLE/im, /^UPDATE\s+\w+\s+SET/im ], html: [ /^<!DOCTYPE\s+html>/i, /<html.*>/i, /<div.*>/i, /<\/\w+>/ ], css: [ /^\s*\.\w+\s*{/m, /^\s*#\w+\s*{/m, /:\s*(flex|grid|block|none);/, /@media\s+/ ], json: [ /^\s*{\s*".*":/m, /^\s*\[\s*{/m ], yaml: [ /^\w+:\s*$/m, /^\s+-\s+\w+:/m ], bash: [ /^#!/, /^\s*echo\s+/m, /\$\{?\w+\}?/, /^\s*if\s+\[/m ], markdown: [ /^#\s+/m, /^\*\*.*\*\*/m, /^\[.*\]\(.*\)/m ] }; // 언어 자동 감지 function detectLanguage(code: string): string { const scores: Record<string, number> = {}; for (const [lang, patterns] of Object.entries(languagePatterns)) { scores[lang] = 0; for (const pattern of patterns) { if (pattern.test(code)) { scores[lang]++; } } } const maxScore = Math.max(...Object.values(scores)); if (maxScore === 0) return 'text'; const detectedLang = Object.entries(scores).find(([, score]) => score === maxScore); return detectedLang ? detectedLang[0] : 'text'; } // 파일명 추론 function inferFilename(code: string, language: string): string | undefined { const extensions: Record<string, string> = { typescript: '.ts', javascript: '.js', python: '.py', java: '.java', go: '.go', rust: '.rs', sql: '.sql', html: '.html', css: '.css', json: '.json', yaml: '.yaml', bash: '.sh', markdown: '.md' }; // 클래스나 함수명에서 파일명 추론 const classMatch = code.match(/(?:class|interface)\s+(\w+)/); if (classMatch) { return classMatch[1] + (extensions[language] || '.txt'); } const funcMatch = code.match(/(?:function|def|fn|func)\s+(\w+)/); if (funcMatch) { return funcMatch[1] + (extensions[language] || '.txt'); } return undefined; } // 중복 제거 function removeDuplicateBlocks(blocks: CodeBlock[]): CodeBlock[] { const seen = new Set<string>(); return blocks.filter(block => { const hash = hashCode(block.code); if (seen.has(hash)) return false; seen.add(hash); return true; }); } // 통계 계산 function calculateStats(blocks: CodeBlock[]): CodeStats { const languageBreakdown: Record<string, number> = {}; let totalLines = 0; for (const block of blocks) { const lines = block.code.split('\n').length; totalLines += lines; languageBreakdown[block.language] = (languageBreakdown[block.language] || 0) + 1; } return { totalBlocks: blocks.length, totalLines, languageBreakdown, averageLinesPerBlock: blocks.length > 0 ? Math.round(totalLines / blocks.length) : 0 }; } // 입력 검증 function validateInput(input: CollectCodeContextInput): string[] { const warnings: string[] = []; if (!input.conversationSummary || input.conversationSummary.trim().length === 0) { warnings.push('Conversation summary is empty'); } if (!input.codeBlocks && !input.rawText) { warnings.push('No code blocks or raw text provided'); } if (input.codeBlocks) { for (let i = 0; i < input.codeBlocks.length; i++) { const block = input.codeBlocks[i]; if (!block.code || block.code.trim().length === 0) { warnings.push(`Code block ${i + 1} is empty`); } } } return warnings; } export function collectCodeContext(input: CollectCodeContextInput): CollectCodeContextOutput { const warnings = validateInput(input); let blocks: CodeBlock[] = []; // 직접 제공된 코드 블록이 있으면 사용 if (input.codeBlocks && input.codeBlocks.length > 0) { blocks = input.codeBlocks.map(block => ({ ...block, language: block.language || (input.autoDetectLanguage !== false ? detectLanguage(block.code) : 'text'), filename: block.filename || inferFilename(block.code, block.language || 'text') })); } // rawText가 있으면 코드 블록 추출 else if (input.rawText) { const extracted = extractCodeBlocks(input.rawText); blocks = extracted.map(block => ({ ...block, language: block.language || (input.autoDetectLanguage !== false ? detectLanguage(block.code) : 'text'), filename: block.filename || inferFilename(block.code, block.language || 'text') })); } // 중복 제거 if (input.removeDuplicates !== false && blocks.length > 0) { const originalCount = blocks.length; blocks = removeDuplicateBlocks(blocks); if (blocks.length < originalCount) { warnings.push(`Removed ${originalCount - blocks.length} duplicate code block(s)`); } } // 태그 자동 추출 (언어 기반) const autoTags = new Set<string>(input.tags || []); for (const block of blocks) { if (block.language && block.language !== 'text') { autoTags.add(block.language); } } const context: CodeContext = { sessionId: generateId(), timestamp: getCurrentTimestamp(), codeBlocks: blocks, conversationSummary: input.conversationSummary?.trim() || '', tags: Array.from(autoTags) }; const result: CollectCodeContextOutput = { context }; // 통계 포함 if (input.includeStats !== false && blocks.length > 0) { result.stats = calculateStats(blocks); } if (warnings.length > 0) { result.warnings = warnings; } return result; } export const collectCodeContextSchema = { name: 'muse_collect_code_context', description: 'Collects code blocks and conversation summaries into a structured context for documentation. Supports automatic language detection, duplicate removal, and statistics.', inputSchema: { type: 'object', properties: { codeBlocks: { type: 'array', description: 'Array of code blocks with language and code content', items: { type: 'object', properties: { language: { type: 'string', description: 'Programming language (auto-detected if not provided)' }, code: { type: 'string', description: 'Code content' }, filename: { type: 'string', description: 'Optional filename (auto-inferred if not provided)' }, description: { type: 'string', description: 'Optional description of the code' } }, required: ['code'] } }, rawText: { type: 'string', description: 'Raw text containing code blocks to extract (alternative to codeBlocks)' }, conversationSummary: { type: 'string', description: 'Summary of the conversation or context' }, tags: { type: 'array', items: { type: 'string' }, description: 'Optional tags for categorization (language tags auto-added)' }, autoDetectLanguage: { type: 'boolean', description: 'Automatically detect programming language (default: true)' }, removeDuplicates: { type: 'boolean', description: 'Remove duplicate code blocks (default: true)' }, includeStats: { type: 'boolean', description: 'Include code statistics in output (default: true)' } }, required: ['conversationSummary'] } };