Files
scraptiss/nodelib/lib.js

81 lines
2.6 KiB
JavaScript

const fs = require("fs");
const YAML = require("yaml");
const child_process = require("child_process");
const puppeteer = require('puppeteer');
async function read_html(filename){
return fs.readFileSync(filename,"utf8");
}
function parse_du_line(text) {
res={}
r=text.split("\t")
if (r.length <2) return [text]; // valid line mus have a column size and folder
foldername=r[1]
res["size"]=r[0]
res["folderpath"]=foldername
res["folder"]=foldername.split("/")[foldername.split("/").length-1]
//console.log([foldername.split("/")[foldername.split("/").length-1], foldername.split("/"),foldername.split("/").length])
r1=foldername.replace(/(\d{3})[\._]?([A\d]{3})/i,"$1.$2")
r1=r1.match(/(\d{3})\.([A\d]{3})/i)
if (!r1) return res;
res["lvanr"] = r1[0]
return res
}
function du_unizeug(){
buf = child_process.execSync("du /mnt/save/daten/Unizeug/ -d 1 -h",{"maxBuffer": 1024*1024*48}).toString()
tab=[];
buf.split("\n").forEach(value => {
text=parse_du_line(value)
tab.push(text)
})
lookup=tab.reduce((acc,item,index)=>{
if (item["lvanr"]) {
if (acc[item["lvanr"]]) acc[item["lvanr"]].push(index);
else acc[item["lvanr"]] =[index];
}
return acc;
}, {});
return {"folders": tab, "lookup": lookup}
}
async function fetch_page(url, selector) {
const browser = await puppeteer.launch({args: ['--no-sandbox']});
const page = await browser.newPage();
await page.goto(url, {
waitUntil: 'networkidle2'
});
//console.log("Waiting for selector: "+selector)
await page.waitForSelector(selector,{ timeout: 7000 }).catch((err)=>{console.log(err); page.screenshot({path: 'exampleerr.png'}); await browser.close();})
await page.screenshot({path: 'example.png'});
let body = await page.evaluate((selector)=>{
let b = $(selector)[0].innerHTML;
return b
},selector).catch((err)=>{console.log(err); await browser.close()});
//console.log(await page.text());
await browser.close();
//console.log(body)
return body
}
function merge_folders_courselist(courselist,folders) {
Object.entries(courselist["tab2"]).forEach(([index0,value0])=> {
value0["courses"].forEach((value)=> {
i=folders["lookup"][value["lvanr"]] // lookup the LVAnr and get an Array with indexes
if (i) value["folders"] = i.reduce((acc,item)=>{
acc.push(folders["folders"][item]) // get one of the folders that are related to LVAnr
return acc
},[])
},value0)
})
return courselist;
}
module.exports = {read_html,du_unizeug, fetch_page,merge_folders_courselist}