multiple added features including tokens

This commit is contained in:
www
2020-08-30 11:25:12 +00:00
parent 5458633cd3
commit 4c8ef46c6a
8 changed files with 283 additions and 61 deletions

81
nodelib/lib.js Normal file
View File

@@ -0,0 +1,81 @@
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}

63
nodelib/tissparse.js Normal file
View File

@@ -0,0 +1,63 @@
const cheerio = require('cheerio');
const slugify=require('slugify');
function courselist(html) {
$ = cheerio.load(html);
tab2={};
tab=[];
studium="";
pruefungsfach="";
lvaname="";
$('table').find('tr').each((index,element)=>{
let element_first=$(element).find('td.nodeTable-title > div.ui-widget');
if (element_first.hasClass("nodeTable-level-0")) {
studium=element_first.text().replace(/^[\s\n]+|[\s\n]+$/g, '')
}
if (element_first.hasClass("nodeTable-level-1")) {
pruefungsfach=element_first.text().replace(/^[\s\n]+|[\s\n]+$/g, '')
}
if (element_first.hasClass("nodeTable-level-2")) {
modul=element_first.text().replace(/^[\s\n]+|[\s\n]+$/g, '')
}
if (element_first.hasClass("nodeTable-level-3")) {
lvaname=element_first.text().replace(/^[\s\n]+|[\s\n]+$/g, '')
tab2[lvaname]={"courses": []};
}
if (element_first.hasClass("nodeTable-level-4")) {
let course_key=element_first.find("div.courseKey").text();
course_key=course_key.replace(/^[\s\n]+|[\s\n]+$/g, '');
let ects=$(($(element).find("td.nodeTable-short")).toArray()[2]).text()
let std=$(($(element).find("td.nodeTable-short")).toArray()[1]).text()
tab2[lvaname]["studium"] = studium
tab2[lvaname]["modul"] = modul
tab2[lvaname]["prufungsfach"] = pruefungsfach
tab2[lvaname]["lvaname"] = lvaname
tab2[lvaname]["ects"]=ects
tab2[lvaname]["std"]=std
tab2[lvaname]["courses"].push({
"href": element_first.find("a").attr("href"),
"lvanr": course_key.split(" ")[0],
"lvatyp": course_key.split(" ")[1],
"lvasem": course_key.split(" ")[2],
"courseKey": course_key,
"courseTitle":element_first.find("div.courseTitle").text().replace(/^[\s\n]+|[\s\n]+$/g, '')
})
tab.push({"href": element_first.find("a").attr("href"),
"studium": studium,
"pruefungsfach": pruefungsfach,
"modul":modul,
"ects":ects,
"std":std,
"courseKey": course_key,
"lvaname": lvaname,
"lvanr": course_key.split(" ")[0],
"lvatyp": course_key.split(" ")[1],
"lvasem": course_key.split(" ")[2],
"courseTitle":element_first.find("div.courseTitle").text().replace(/^[\s\n]+|[\s\n]+$/g, '')
})
}
})
return {tab2};
}
module.exports = {courselist}