@afterxleep/doc-bot
Version:
Generic MCP server for intelligent documentation access in any project
349 lines (293 loc) • 10 kB
JavaScript
import fs from 'fs-extra';
import path from 'path';
import { createWriteStream } from 'fs';
import axios from 'axios';
import * as tar from 'tar';
import AdmZip from 'adm-zip';
import plist from 'plist';
import crypto from 'crypto';
import { pipeline } from 'stream/promises';
import os from 'os';
/**
* Manages docset storage, downloading, and metadata for doc-bot
* Adapted from scout server's DocsetManager
*/
export class DocsetService {
constructor(storagePath) {
this.storagePath = storagePath || path.join(os.homedir(), 'Developer', 'DocSets');
this.metadataPath = path.join(this.storagePath, 'docsets.json');
this.docsets = new Map();
this.downloadProgress = new Map();
}
async initialize() {
// Create storage directory if it doesn't exist
await fs.ensureDir(this.storagePath);
// Load existing docsets metadata
await this.loadMetadata();
}
async loadMetadata() {
try {
if (await fs.pathExists(this.metadataPath)) {
const data = await fs.readJson(this.metadataPath);
for (const docset of data) {
// Verify docset still exists
const docsetPath = path.join(this.storagePath, docset.id);
if (await fs.pathExists(docsetPath)) {
this.docsets.set(docset.id, {
...docset,
downloadedAt: new Date(docset.downloadedAt)
});
}
}
}
} catch (error) {
// If metadata is corrupted, start fresh
this.docsets.clear();
}
}
async saveMetadata() {
const docsets = Array.from(this.docsets.values());
await fs.writeJson(this.metadataPath, docsets, { spaces: 2 });
}
async addDocset(source) {
// Determine if source is URL or local path
const isUrl = source.startsWith('http://') || source.startsWith('https://');
const isLocalPath = await this.isLocalPath(source);
if (!isUrl && !isLocalPath) {
throw new Error(`Invalid source: ${source}. Must be a valid URL or local file path.`);
}
// Generate unique ID for the docset
const docsetId = crypto.createHash('md5').update(source).digest('hex').substring(0, 8);
// Check if already exists
if (this.docsets.has(docsetId)) {
throw new Error(`Docset from ${source} is already installed`);
}
if (isUrl) {
return await this.addDocsetFromUrl(source, docsetId);
} else {
return await this.addDocsetFromLocal(source, docsetId);
}
}
async isLocalPath(source) {
try {
const stats = await fs.stat(source);
return stats.isFile() || stats.isDirectory();
} catch {
return false;
}
}
async addDocsetFromLocal(localPath, docsetId) {
try {
const stats = await fs.stat(localPath);
const extractPath = path.join(this.storagePath, docsetId);
await fs.ensureDir(extractPath);
let docsetPath;
if (stats.isDirectory() && localPath.endsWith('.docset')) {
// Direct docset directory - copy it
docsetPath = path.join(extractPath, path.basename(localPath));
await fs.copy(localPath, docsetPath);
} else if (stats.isFile()) {
// Handle different archive formats
if (localPath.endsWith('.tgz') || localPath.endsWith('.tar.gz')) {
// Tar archive - extract it
await tar.extract({
file: localPath,
cwd: extractPath,
strip: 0
});
} else if (localPath.endsWith('.zip')) {
// ZIP archive - extract it
const zip = new AdmZip(localPath);
zip.extractAllTo(extractPath, true);
} else {
throw new Error('Local file must be a .tgz, .tar.gz, or .zip archive');
}
// Find the .docset directory
const files = await fs.readdir(extractPath);
const docsetDir = files.find(f => f.endsWith('.docset'));
if (!docsetDir) {
throw new Error('No .docset directory found in the archive');
}
docsetPath = path.join(extractPath, docsetDir);
} else {
throw new Error('Local path must be a .docset directory or archive file (.tgz, .tar.gz, .zip)');
}
// Read docset metadata
const metadata = await this.readDocsetMetadata(docsetPath);
// Get docset size
const size = await this.getDirectorySize(docsetPath);
// Create docset info
const docsetInfo = {
id: docsetId,
name: metadata.CFBundleName || path.basename(docsetPath).replace('.docset', ''),
source: localPath,
path: docsetPath,
version: metadata.CFBundleVersion,
platform: metadata.DocSetPlatformFamily,
downloadedAt: new Date(),
size
};
// Save to metadata
this.docsets.set(docsetId, docsetInfo);
await this.saveMetadata();
return docsetInfo;
} catch (error) {
// Clean up on failure
try {
const extractPath = path.join(this.storagePath, docsetId);
await fs.remove(extractPath);
} catch (cleanupError) {
// Ignore cleanup errors
}
throw error;
}
}
async addDocsetFromUrl(url, docsetId) {
const progress = {
docsetId,
url,
totalBytes: 0,
downloadedBytes: 0,
percentage: 0,
status: 'downloading'
};
this.downloadProgress.set(docsetId, progress);
try {
// Determine file extension from URL
const urlLower = url.toLowerCase();
let fileExt = '.tgz';
if (urlLower.endsWith('.zip')) {
fileExt = '.zip';
} else if (urlLower.endsWith('.tar.gz')) {
fileExt = '.tar.gz';
}
// Download the docset
const downloadPath = path.join(this.storagePath, `${docsetId}${fileExt}`);
await this.downloadFile(url, downloadPath, docsetId);
// Update progress
progress.status = 'extracting';
// Extract the docset
const extractPath = path.join(this.storagePath, docsetId);
await fs.ensureDir(extractPath);
if (fileExt === '.zip') {
// Extract ZIP file
const zip = new AdmZip(downloadPath);
zip.extractAllTo(extractPath, true);
} else {
// Extract tar archive
await tar.extract({
file: downloadPath,
cwd: extractPath,
strip: 0
});
}
// Clean up downloaded file
await fs.unlink(downloadPath);
// Find the .docset directory
const files = await fs.readdir(extractPath);
const docsetDir = files.find(f => f.endsWith('.docset'));
if (!docsetDir) {
throw new Error('No .docset directory found in the downloaded archive');
}
const docsetPath = path.join(extractPath, docsetDir);
// Read docset metadata
const metadata = await this.readDocsetMetadata(docsetPath);
// Get docset size
const size = await this.getDirectorySize(docsetPath);
// Create docset info
const docsetInfo = {
id: docsetId,
name: metadata.CFBundleName || docsetDir.replace('.docset', ''),
source: url,
path: docsetPath,
version: metadata.CFBundleVersion,
platform: metadata.DocSetPlatformFamily,
downloadedAt: new Date(),
size
};
// Save to metadata
this.docsets.set(docsetId, docsetInfo);
await this.saveMetadata();
// Update progress
progress.status = 'completed';
progress.percentage = 100;
return docsetInfo;
} catch (error) {
progress.status = 'failed';
progress.error = error.message;
// Clean up any partial downloads
try {
const extractPath = path.join(this.storagePath, docsetId);
await fs.remove(extractPath);
} catch (cleanupError) {
// Ignore cleanup errors
}
throw error;
} finally {
// Clean up progress after a delay
setTimeout(() => {
this.downloadProgress.delete(docsetId);
}, 5000);
}
}
async downloadFile(url, destination, docsetId) {
const response = await axios({
method: 'GET',
url,
responseType: 'stream',
onDownloadProgress: (progressEvent) => {
const progress = this.downloadProgress.get(docsetId);
if (progress && progressEvent.total) {
progress.totalBytes = progressEvent.total;
progress.downloadedBytes = progressEvent.loaded;
progress.percentage = Math.round((progressEvent.loaded / progressEvent.total) * 100);
}
}
});
const writer = createWriteStream(destination);
await pipeline(response.data, writer);
}
async readDocsetMetadata(docsetPath) {
const plistPath = path.join(docsetPath, 'Contents', 'Info.plist');
const plistContent = await fs.readFile(plistPath, 'utf-8');
return plist.parse(plistContent);
}
async getDirectorySize(dirPath) {
let size = 0;
const files = await fs.readdir(dirPath, { withFileTypes: true });
for (const file of files) {
const filePath = path.join(dirPath, file.name);
if (file.isDirectory()) {
size += await this.getDirectorySize(filePath);
} else {
const stats = await fs.stat(filePath);
size += stats.size;
}
}
return size;
}
async removeDocset(docsetId) {
const docset = this.docsets.get(docsetId);
if (!docset) {
throw new Error(`Docset ${docsetId} not found`);
}
// Remove the docset directory
const docsetDir = path.dirname(docset.path);
await fs.remove(docsetDir);
// Update metadata
this.docsets.delete(docsetId);
await this.saveMetadata();
}
async listDocsets() {
return Array.from(this.docsets.values());
}
getDocset(docsetId) {
return this.docsets.get(docsetId);
}
getDownloadProgress(docsetId) {
return this.downloadProgress.get(docsetId);
}
getAllDownloadProgress() {
return Array.from(this.downloadProgress.values());
}
}