Permalink
Cannot retrieve contributors at this time
Name already in use
A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
AnimeHigherLower/scraper.js
Go to fileThis commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
412 lines (343 sloc)
19.2 KB
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
const request = require("request-promise"); | |
const cheerio = require("cheerio"); | |
const fs = require('fs'); | |
async function main() { | |
let shows = new Array(600); | |
let members = new Array(600); | |
let images = new Array(600); | |
let smallImages = new Array(600); | |
let scores = new Array(600); | |
var showDict = {}; | |
// Scrape Page 1 | |
var result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity"); | |
var $ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index] = $(element).text(); | |
}); | |
// Scrape Page 2 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=50"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 50] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 50] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 50] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 50] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 50] = $(element).text(); | |
}); | |
// Scrape Page 3 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=100"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 100] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 100] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 100] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 100] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 100] = $(element).text(); | |
}); | |
// Scrape Page 4 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=150"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 150] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 150] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 150] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 150] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 150] = $(element).text(); | |
}); | |
// Scrape Page 5 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=200"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 200] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 200] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 200] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 200] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 200] = $(element).text(); | |
}); | |
// Scrape Page 6 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=250"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 250] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 250] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 250] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 250] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 250] = $(element).text(); | |
}); | |
// Scrape Page 7 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=300"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 300] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 300] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 300] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 300] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 300] = $(element).text(); | |
}); | |
// Scrape Page 8 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=350"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 350] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 350] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 350] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 350] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 350] = $(element).text(); | |
}); | |
// Scrape Page 9 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=400"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 400] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 400] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 400] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 400] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 400] = $(element).text(); | |
}); | |
// Scrape Page 10 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=450"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 450] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 450] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 450] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 450] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 450] = $(element).text(); | |
}); | |
// Scrape Page 11 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=500"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 500] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 500] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 500] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 500] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 500] = $(element).text(); | |
}); | |
// Scrape Page 12 | |
result = await request.get("https://myanimelist.net/topanime.php?type=bypopularity&limit=550"); | |
$ = cheerio.load(result); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.di-ib.clearfix > h3").each((index, element) => { | |
shows[index + 550] = $(element).text(); | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > div > div.information.di-ib.mt4").each((index, element) => { | |
var tempStr = $(element).text(); | |
var tempArr = tempStr.split('\n'); | |
var tempArr2 = tempArr[3].trim().split(' '); | |
members[index + 550] = tempArr2[0]; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.title.al.va-t.word-break > a > img").each((index, element) => { | |
var splitArr = $(element).attr("data-src").split('/'); | |
var first = splitArr[7]; | |
var secondArr = splitArr[8].split('.'); | |
var second = secondArr[0]; | |
var img = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + "l.webp"; | |
images[index + 550] = img; | |
var smallImg = "https://cdn.myanimelist.net/images/anime/" + first + "/" + second + ".jpg"; | |
smallImages[index + 550] = smallImg; | |
}); | |
$("#content > div.pb12 > table > tbody > tr > td.score.ac.fs14 > div > span").each((index, element) => { | |
scores[index + 550] = $(element).text(); | |
}); | |
// Write to files | |
var output = shows.join("\n"); | |
fs.writeFile("/home/jason/popularanime.txt", output, function(err) { | |
if (err) { | |
return console.log(err); | |
} | |
}); | |
output = members.join("\n"); | |
fs.writeFile("/home/jason/animemembers.txt", output, function(err) { | |
if (err) { | |
return console.log(err); | |
} | |
}); | |
output = images.join("\n"); | |
fs.writeFile("/home/jason/popularanimepictures.txt", output, function(err) { | |
if (err) { | |
return console.log(err); | |
} | |
}); | |
output = scores.join("\n"); | |
fs.writeFile("/home/jason/animescores.txt", output, function(err) { | |
if (err) { | |
return console.log(err); | |
} | |
}); | |
output = smallImages.join("\n"); | |
fs.writeFile("/home/jason/smallanimepictures.txt", output, function(err) { | |
if (err) { | |
return console.log(err); | |
} | |
}); | |
} | |
main(); |