tykayn/scripts
1
0
Fork 0
Code Issues 1 Pull Requests Projects Releases Wiki Activity

ajotu du scraping de bibliothèque

Browse Source 
Signed-off-by: tykayn <contact@cipherbliss.com>
This commit is contained in:
Tykayn 2023-08-18 10:18:14 +02:00 committed by tykayn
parent 265f50291d
commit 4f828962c1
2 changed files with 113869 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
scraping/main.ts
View File

@ -22,7 +22,7 @@ const books: Book[] = [];
// autres pages: // autres pages:
// http://www.mediatheque-de-briis-sous-forges.net/mediatheque-de-briis-sous-forges.net/opac/recherche/catalogue?node=0&value=0&page=2 // http://www.mediatheque-de-briis-sous-forges.net/mediatheque-de-briis-sous-forges.net/opac/recherche/catalogue?node=0&value=0&page=2
let page_max = 1927 let page_max = 1927
page_max = 4 // page_max = 4
const getTables = (html: string): any => { const getTables = (html: string): any => {
const $ = cheerio.load(html); const $ = cheerio.load(html);
@ -93,10 +93,10 @@ const scrapOnePage = (tables: any) => {
let img_src = $(table).find('td img').attr('src'); let img_src = $(table).find('td img').attr('src');
console.log(img_src); console.log(img_src);
books.push({ books.push({
author: boom[0].trim(), author: boom[0],
title: splitting[0].trim(), title: splitting[0],
description: splitting[1].trim(), description: splitting[1],
format: format[1].trim(), format: format[1],
img: img_src img: img_src
}) })
// console.log(cheerio.load(table).html()) // console.log(cheerio.load(table).html())

113864
scraping/output/books.json
View File

File diff suppressed because it is too large Load Diff