Lspace MCP Server

import { RepositoryManager } from '../core/repositoryManager'; import { Repository } from '../core/repository'; import path from 'path'; import fs from 'fs'; interface SearchResult { repositoryId: string; path: string; title?: string; excerpt?: string; score: number; lastModified?: Date; } interface SearchOptions { repositoryId?: string; limit?: number; offset?: number; } /** * SearchService provides functionality for searching across repositories */ export class SearchService { private repositoryManager: RepositoryManager; private indexCache: Map<string, IndexEntry[]> = new Map(); constructor(repositoryManager: RepositoryManager) { this.repositoryManager = repositoryManager; } /** * Initialize the search index for all repositories */ async initializeIndex(): Promise<void> { const repositories = this.repositoryManager.listRepositories(); for (const repo of repositories) { await this.updateIndex(repo.id); } } /** * Update the search index for a specific repository * @param repositoryId Repository ID */ async updateIndex(repositoryId: string): Promise<void> { const repository = this.repositoryManager.getRepository(repositoryId); // Get all files in the repository const files = await repository.listFiles(); // Clear existing index for this repository this.indexCache.delete(repositoryId); // Only index markdown and text files const indexableFiles = files.filter(file => file.type === 'file' && (file.path.endsWith('.md') || file.path.endsWith('.txt') || file.path.endsWith('.json')) ); // Build index entries const indexEntries: IndexEntry[] = []; for (const file of indexableFiles) { try { const content = await repository.readFile(file.path); // Extract title (first heading for markdown) let title = path.basename(file.path); if (file.path.endsWith('.md')) { const headingMatch = content.match(/^#\s+(.+)$/m); if (headingMatch && headingMatch[1]) { title = headingMatch[1].trim(); } } // Create index entry const entry: IndexEntry = { path: file.path, title, content, lastModified: file.lastModified }; indexEntries.push(entry); } catch (error) { // Skip files that can't be read console.warn(`Error indexing file ${file.path}: ${error}`); } } // Store the index this.indexCache.set(repositoryId, indexEntries); } /** * Perform a keyword search across all repositories * @param query Search query * @param options Search options * @returns Search results */ async keywordSearch(query: string, options: SearchOptions = {}): Promise<SearchResult[]> { // Normalize query const normalizedQuery = query.toLowerCase().trim(); const queryTerms = normalizedQuery.split(/\s+/).filter(Boolean); if (queryTerms.length === 0) { return []; } // Get repositories to search const repoIds = options.repositoryId ? [options.repositoryId] : this.repositoryManager.listRepositories().map(r => r.id); const results: SearchResult[] = []; // Search each repository for (const repoId of repoIds) { // Get index entries for this repository let entries = this.indexCache.get(repoId); // If no entries cached, update the index if (!entries || entries.length === 0) { await this.updateIndex(repoId); entries = this.indexCache.get(repoId) || []; } // Search the entries for (const entry of entries) { const score = this.calculateScore(entry, queryTerms); if (score > 0) { // Create excerpt const excerpt = this.createExcerpt(entry.content, queryTerms[0]); results.push({ repositoryId: repoId, path: entry.path, title: entry.title, excerpt, score, lastModified: entry.lastModified }); } } } // Sort by score descending results.sort((a, b) => b.score - a.score); // Apply limit and offset const offset = options.offset || 0; const limit = options.limit || results.length; return results.slice(offset, offset + limit); } /** * Find files related to a given file * @param repositoryId Repository ID * @param filePath File path * @param limit Maximum number of results * @returns Related files */ async findRelatedFiles( repositoryId: string, filePath: string, limit: number = 5 ): Promise<SearchResult[]> { try { // Get the repository const repository = this.repositoryManager.getRepository(repositoryId); // Get the content of the file const content = await repository.readFile(filePath); // Extract keywords from content const keywords = this.extractKeywords(content); // Perform a search with these keywords const results = await this.keywordSearch(keywords.join(' '), { limit: limit + 1 // +1 because the file itself will be in the results }); // Filter out the original file return results.filter(result => !(result.repositoryId === repositoryId && result.path === filePath) ); } catch (error) { console.error(`Error finding related files: ${error}`); return []; } } /** * Calculate a relevance score for an entry against query terms * @param entry Index entry * @param queryTerms Query terms * @returns Score (0 if no match) */ private calculateScore(entry: IndexEntry, queryTerms: string[]): number { const content = entry.content.toLowerCase(); const title = entry.title.toLowerCase(); let score = 0; // Check each query term for (const term of queryTerms) { // Title matches are worth more if (title.includes(term)) { score += 10; } // Content matches if (content.includes(term)) { // Count occurrences const occurrences = (content.match(new RegExp(term, 'g')) || []).length; score += occurrences; } } return score; } /** * Create a text excerpt highlighting the query match * @param content Full content * @param query Query term to highlight * @returns Excerpt with match context */ private createExcerpt(content: string, query: string): string { const lowerContent = content.toLowerCase(); const index = lowerContent.indexOf(query); if (index === -1) { // Just return the beginning of the content return content.substring(0, 100) + '...'; } // Get context around the match const start = Math.max(0, index - 40); const end = Math.min(content.length, index + query.length + 40); return (start > 0 ? '...' : '') + content.substring(start, end) + (end < content.length ? '...' : ''); } /** * Extract significant keywords from content * @param content Content to analyze * @returns Array of keywords */ private extractKeywords(content: string): string[] { // Split content into words const words = content.toLowerCase() .replace(/[^\w\s]/g, ' ') .split(/\s+/) .filter(Boolean); // Count word frequencies const wordCounts = new Map<string, number>(); for (const word of words) { // Skip stop words if (this.isStopWord(word)) continue; // Skip short words if (word.length < 3) continue; // Count word wordCounts.set(word, (wordCounts.get(word) || 0) + 1); } // Convert to array and sort by frequency const sortedWords = [...wordCounts.entries()] .sort((a, b) => b[1] - a[1]) .map(([word]) => word); // Return top keywords (up to 10) return sortedWords.slice(0, 10); } /** * Check if a word is a common stop word * @param word Word to check * @returns True if it's a stop word */ private isStopWord(word: string): boolean { const stopWords = new Set([ 'a', 'an', 'the', 'and', 'or', 'but', 'if', 'then', 'else', 'when', 'at', 'from', 'by', 'on', 'off', 'for', 'in', 'out', 'over', 'to', 'is', 'am', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'do', 'does', 'did', 'will', 'would', 'shall', 'should', 'can', 'could', 'may', 'might', 'must', 'this', 'that', 'these', 'those', 'i', 'you', 'he', 'she', 'it', 'we', 'they' ]); return stopWords.has(word); } } /** * Index entry for a document */ interface IndexEntry { path: string; title: string; content: string; lastModified?: Date; }