paper-search-mcp-nodejs
Version:
A Node.js MCP server for searching and downloading academic papers from multiple sources, including arXiv, PubMed, bioRxiv, Web of Science, and more.
147 lines • 6.22 kB
JavaScript
/**
* Wiley TDM (Text and Data Mining) API - PDF Download Only
*
* Documentation: https://onlinelibrary.wiley.com/library-info/resources/text-and-datamining
* GitHub Client: https://github.com/WileyLabs/tdm-client
*
* IMPORTANT: Wiley TDM API does NOT support keyword search.
* It only supports downloading PDFs by DOI.
* For searching Wiley content, use Crossref API with publisher filter.
*
* API Endpoint: https://api.wiley.com/onlinelibrary/tdm/v1/articles/{DOI}
* Header: Wiley-TDM-Client-Token: <token>
*
* Rate limits:
* - Up to 3 articles per second
* - Up to 60 requests per 10 minutes (build in 10 second delay between requests)
*/
import axios from 'axios';
import { PaperSource } from './PaperSource.js';
import { PaperFactory } from '../models/Paper.js';
import { RateLimiter } from '../utils/RateLimiter.js';
import { sanitizeDoi } from '../utils/SecurityUtils.js';
import { TIMEOUTS } from '../config/constants.js';
export class WileySearcher extends PaperSource {
client;
rateLimiter;
constructor(tdmToken) {
super('wiley', 'https://api.wiley.com/onlinelibrary/tdm/v1', tdmToken);
this.client = axios.create({
baseURL: 'https://api.wiley.com/onlinelibrary/tdm/v1',
headers: {
'Accept': 'application/pdf',
...(tdmToken ? { 'Wiley-TDM-Client-Token': tdmToken } : {})
},
maxRedirects: 5,
timeout: TIMEOUTS.EXTENDED
});
// Wiley rate limits: 3 articles/sec, 60 requests/10min
this.rateLimiter = new RateLimiter({
requestsPerSecond: 0.1, // Conservative: ~6 per minute
burstCapacity: 3
});
}
/**
* Search is NOT supported by Wiley TDM API.
* Use Crossref API to search for Wiley articles, then use download() to get PDFs.
*/
async search(query, options = {}) {
throw new Error('Wiley TDM API does not support keyword search. ' +
'Use Crossref API (search_crossref) to find Wiley articles by DOI, ' +
'then use download_paper with platform="wiley" to download PDFs.');
}
/**
* Download PDF by DOI using Wiley TDM API
* @param doi - The DOI of the article (e.g., "10.1111/jtsb.12390")
* @param options - Download options including savePath
*/
async downloadPdf(doi, options = {}) {
if (!this.apiKey) {
throw new Error('Wiley TDM token is required. Set WILEY_TDM_TOKEN environment variable.');
}
// Clean and validate DOI format
const doiResult = sanitizeDoi(doi);
if (!doiResult.valid) {
throw new Error(`Invalid DOI format: ${doi}. ${doiResult.error || ''}`);
}
const cleanDoi = doiResult.sanitized;
const fs = await import('fs');
const path = await import('path');
const savePath = options.savePath || './downloads';
if (!fs.existsSync(savePath)) {
fs.mkdirSync(savePath, { recursive: true });
}
// Encode DOI for URL (replace / with %2F)
const encodedDoi = encodeURIComponent(cleanDoi);
const url = `/articles/${encodedDoi}`;
await this.rateLimiter.waitForPermission();
try {
const response = await this.client.get(url, {
responseType: 'stream',
headers: {
'Wiley-TDM-Client-Token': this.apiKey,
'Accept': 'application/pdf'
}
});
// Generate filename from DOI
const fileName = `${cleanDoi.replace(/[\/\\:*?"<>|]/g, '_')}.pdf`;
const filePath = path.join(savePath, fileName);
const writer = fs.createWriteStream(filePath);
response.data.pipe(writer);
return new Promise((resolve, reject) => {
writer.on('finish', () => resolve(filePath));
writer.on('error', reject);
});
}
catch (error) {
const status = error.response?.status;
const errorMessages = {
400: 'No TDM Client Token was found in the request',
403: 'TDM Client Token is invalid or not registered',
404: 'Access denied - you or your institution does not have access to this content. Check your subscription.',
429: 'Rate limit exceeded. Please reduce request frequency (max 60 requests per 10 minutes).'
};
if (status && errorMessages[status]) {
throw new Error(`Wiley TDM Error (${status}): ${errorMessages[status]}`);
}
throw new Error(`Failed to download PDF: ${error.message}`);
}
}
/**
* Get article metadata and download link (without downloading)
*/
async getArticleInfo(doi) {
// Clean and validate DOI
const doiResult = sanitizeDoi(doi);
const cleanDoi = doiResult.valid ? doiResult.sanitized : doi;
// Since TDM API only provides PDF download, we create basic paper info from DOI
return PaperFactory.create({
paperId: cleanDoi,
title: `Wiley Article: ${cleanDoi}`,
authors: [],
abstract: '',
doi: cleanDoi,
publishedDate: null,
pdfUrl: `https://api.wiley.com/onlinelibrary/tdm/v1/articles/${encodeURIComponent(cleanDoi)}`,
url: `https://doi.org/${cleanDoi}`,
source: 'wiley',
extra: {
note: 'Use Crossref API for full metadata. This endpoint only provides PDF download.'
}
});
}
getCapabilities() {
return {
search: false, // TDM API does not support search
download: true, // PDF download by DOI
fullText: true, // Full PDF available
citations: false,
requiresApiKey: true,
supportedOptions: [] // No search options - only DOI-based download
};
}
async readPaper(paperId, options = {}) {
return 'Wiley TDM API only supports PDF download. Use downloadPdf() method with a DOI to get the full PDF.';
}
}
//# sourceMappingURL=WileySearcher.js.map