Trying to backup my Instagram stuff because their sandboxed API limitations only let you get the most recent 20 items

build-list.js

// Step 1: Scroll down until all items have loaded.
// Step 2: Paste the following into the console.

// TODO: Probably the classNames and document structure change ¯\_(ツ)_/¯
const items = document.querySelectorAll('._mck9w._gvoze._f2mse');
const media = [];

let item;

for (let i = 0, len = items.length; i < len; i++) {
  item = items[i];

  let is_video = false;

  try {
    is_video = 'Video' === item.childNodes[0].childNodes[1].childNodes[0].childNodes[0].innerText;
  } catch (e) {}

  media.push({
    href: item.childNodes[0].href,
    src: item.childNodes[0].childNodes[0].childNodes[0].childNodes[0].src,
    is_video: is_video
  })
}

// Step 3: Right click this line and “Save as…”
console.log(JSON.stringify(media, null, 2))

save-files.js

// Run this with node
const fs = require('fs');
const https = require('https');
const { exec } = require('child_process');

// Whatever you called this file
const items = require('./data.json');
const dir = `${__dirname}/media`;

if (!fs.existsSync(dir)) {
  fs.mkdirSync(dir);
}

function getVideo(item) {
  https.get(item.href, (res) => {
    const { statusCode } = res;

    if (200 !== statusCode) {
      console.log(`:( ${statusCode} for ${item.href}`);
      res.resume();
      return;
    }

    res.setEncoding('utf8');
    let data = '';
    res.on('data', (chunk) => { data += chunk; });
    res.on('end', () => {
      let url = data.match(/\<meta\sproperty\=\"og:video\"\scontent\=\"(.*)\"\s\/\>/i)[1];
      downloadMedia(url);
    });
  })
}

function basename(path) {
  return path.replace(/\\/g,'/').replace(/.*\//, '');
}

function downloadMedia(url) {
  if (fs.existsSync(`${dir}/${basename(url)}`)) {
    console.log(`${dir}/${basename(url)} exists`);
    return;
  }

  // I’m lazy
  console.log(`Fetching ${url}`);
  exec(`cd ${dir}; wget ${url}`);
}

let item;

for (let i = 0, len = items.length; i < len; i++) {
  item = items[i];

  if (item.is_video) {
    getVideo(item);
    continue;
  }

  downloadMedia(item.src)
}