Parse les annotations des videos de l'Assemblée Nationale

parse.js

var links = document.querySelectorAll('.mediaIndex a')
var player = document.querySelector('#html5_player')

var results = []

var i = 0;
function next() {
	if (links[i]) {
		links[i].click()
		setTimeout(() => {
			var result = [player.currentTime, links[i].innerText, links[i].className];
			results.push(result)
			console.log(i, result)
			i += 1
			next()
		}, 1000)
	} else {
		document.body.innerText = JSON.stringify(results)
	}
}
next()

parse.py

import json, time, sys
import datetime

from selenium import webdriver
from selenium.webdriver.chrome.options import Options

driver = webdriver.Chrome()
driver.get(sys.argv[1])
time.sleep(20)
driver.execute_script(open('parse.js').read())
while True:
	content = driver.find_element_by_tag_name('body').get_attribute('innerText')
	try:
		content = json.loads(content)
		break
	except:
		time.sleep(2)

transformed = []

prev_time = None
for el in content:
	new_el = {}
	new_el['time'] = str(datetime.timedelta(seconds=el[0])).split('.')[0]
	if new_el['time'] == prev_time:
		continue
	prev_time = new_el['time']
	new_el['titre'] = el[1]
	new_el['level'] = int(el[2].split('level')[1])
	transformed.append(new_el)


print(json.dumps(transformed, indent=2, ensure_ascii=False))