using winston logger

This commit is contained in:
orangecoding
2025-09-13 18:57:56 +02:00
parent e868cdce86
commit 21415dcff3
19 changed files with 230 additions and 49 deletions

View File

@@ -1,6 +1,7 @@
import { setDebug } from './utils.js';
import puppeteerExtractor from './puppeteerExtractor.js';
import { loadParser, parse } from './parser/parser.js';
import logger from '../logger.js';
const DEFAULT_OPTIONS = {
debug: false,
@@ -32,7 +33,7 @@ export default class Extractor {
loadParser(this.responseText);
}
} catch (error) {
console.error('Error trying to load page.', error);
logger.error('Error trying to load page.', error);
}
return this;
};

View File

@@ -1,4 +1,5 @@
import * as cheerio from 'cheerio';
import logger from '../../logger.js';
let $ = null;
@@ -8,19 +9,19 @@ export function loadParser(text) {
export function parse(crawlContainer, crawlFields, text, url) {
if (!text) {
console.warn('No content found for ', url);
logger.warn('No content found for ', url);
return null;
}
if (!crawlContainer || !crawlFields) {
console.warn('Cannot parse, selector was empty for url ', url);
logger.warn('Cannot parse, selector was empty for url ', url);
return null;
}
const result = [];
if ($(crawlContainer).length === 0) {
console.warn('No elements in crawl container found for url ', url);
logger.warn('No elements in crawl container found for url ', url);
return null;
}
@@ -58,7 +59,7 @@ export function parse(crawlContainer, crawlFields, text, url) {
parsedObject[key] = value || null;
} catch (error) {
console.error(`Error parsing field '${key}' with selector '${fieldSelector}':`, error);
logger.error(`Error parsing field '${key}' with selector '${fieldSelector}':`, error);
parsedObject[key] = null;
}
}
@@ -66,9 +67,7 @@ export function parse(crawlContainer, crawlFields, text, url) {
if (parsedObject.id != null) {
result.push(parsedObject);
} else {
/* eslint-disable no-console */
console.debug('ID not found. Not relaying object.');
/* eslint-enable no-console */
logger.debug('ID not found. Not relaying object.');
}
});
@@ -91,7 +90,7 @@ function applyModifiers(value, modifiers) {
value = value.replace(/\n/g, ' ');
break;
default:
console.warn(`Unknown modifier: ${modifier}`);
logger.warn(`Unknown modifier: ${modifier}`);
}
});

View File

@@ -1,6 +1,7 @@
import puppeteer from 'puppeteer-extra';
import StealthPlugin from 'puppeteer-extra-plugin-stealth';
import { debug, DEFAULT_HEADER, botDetected } from './utils.js';
import logger from '../logger.js';
puppeteer.use(StealthPlugin());
@@ -33,13 +34,13 @@ export default async function execute(url, waitForSelector, options) {
const statusCode = response.status();
if (botDetected(pageSource, statusCode)) {
console.warn('We have been detected as a bot :-/ Tried url: => ', url);
logger.warn('We have been detected as a bot :-/ Tried url: => ', url);
return null;
}
return await page.content();
} catch (error) {
console.error('Error executing with puppeteer executor', error);
logger.error('Error executing with puppeteer executor', error);
return null;
} finally {
if (browser != null) {

View File

@@ -1,3 +1,5 @@
import logger from '../logger.js';
let debuggingOn = false;
export const DEFAULT_HEADER = {
@@ -15,9 +17,7 @@ export const setDebug = (options) => {
export const debug = (message) => {
if (debuggingOn) {
/* eslint-disable no-console */
console.debug(message);
/* eslint-enable no-console */
logger.debug(message);
}
};