SCRAPE-3: Implement fetchPage() with axios, timeout, retry (#8)

Co-authored-by: Stephen Minakian <stephenminakian@gmail.com>
Co-committed-by: Stephen Minakian <stephenminakian@gmail.com>
This commit is contained in:
2026-02-05 15:58:07 -07:00
committed by stephen
parent 2993d019c5
commit c6beaf8333
3 changed files with 651 additions and 2 deletions

114
services/scraperService.js Normal file
View File

@ -0,0 +1,114 @@
/**
* Scraper Service
*
* Core scraper logic including HTTP fetching, HTML parsing,
* data transformation, and database operations.
*/
const axios = require('axios');
const config = require('../config/scraper');
/**
* Sleep utility for retry delays
* @param {number} ms - Milliseconds to sleep
* @returns {Promise<void>}
*/
function sleep(ms) {
return new Promise(resolve => setTimeout(resolve, ms));
}
/**
* Determine if an error is retryable
* @param {Error} error - Axios error
* @returns {boolean} True if should retry
*/
function isRetryableError(error) {
// Network errors (timeout, DNS, connection) don't have a response property
if (!error.response) {
return true;
}
const status = error.response.status;
// 5xx server errors are retryable
if (status >= 500) {
return true;
}
// 429 Too Many Requests - do not retry immediately
if (status === 429) {
return false;
}
// 4xx client errors - do not retry
return false;
}
/**
* Fetch page HTML with retry logic
* @param {string} url - Target URL
* @param {Object} logger - Logger instance
* @returns {Promise<string>} HTML content
* @throws {Error} After all retries exhausted
*/
async function fetchPage(url, logger) {
const { maxRetries, baseDelay, timeout } = config.RETRY_CONFIG;
let lastError;
for (let attempt = 1; attempt <= maxRetries + 1; attempt++) {
try {
logger.info('Fetching page', { url, attempt });
const response = await axios.get(url, {
timeout,
headers: {
'User-Agent': config.USER_AGENT
},
maxRedirects: 5,
validateStatus: (status) => status < 400 // Accept 2xx and 3xx
});
logger.info('Page fetched successfully', {
status: response.status,
contentLength: response.data.length
});
return response.data;
} catch (error) {
lastError = error;
// Determine if error is retryable
const isRetryable = isRetryableError(error);
logger.warn('Fetch attempt failed', {
attempt,
errorType: error.name,
errorMessage: error.message,
statusCode: error.response?.status,
isRetryable
});
// Don't retry non-retryable errors (4xx)
if (!isRetryable) {
throw error;
}
// Don't wait after last attempt
if (attempt <= maxRetries) {
const delay = baseDelay * Math.pow(2, attempt - 1); // Exponential backoff
logger.info('Waiting before retry', { delay });
await sleep(delay);
}
}
}
throw lastError;
}
module.exports = {
fetchPage,
// Export for testing
isRetryableError,
sleep
};