from __future__ import division

import time
import csv

import lxml.html

URL = 'http://languagelog.ldc.upenn.edu/nll/?p=8199'

start_time = None
times = []

tree = lxml.html.parse(URL)
for p in tree.iter('p'):
    if p.get('class') == 'postmeta':
        time_str = p.text.split('\n')[1].strip()
        time_struct = time.strptime(time_str, '%B %d, %Y @ %I:%M %p')
        start_time = time_struct
    if p.get('class') == 'commentmeta':
        date_time = p.text.strip()
        date_time += ' ' + p[0].text
        time_struct = time.strptime(date_time, '%B %d, %Y @ %I:%M %p')
        times.append(time_struct)


first = time.mktime(start_time)
last = time.mktime(times[-1])


with open('test.csv', 'w', newline='') as csvfile:
    writer = csv.writer(csvfile, delimiter=',',
                        quotechar='"', quoting=csv.QUOTE_MINIMAL)
    writer.writerow(['Number of Comments'])
    comments = 0
    for second in range(int(first), int(last), 60):
        if time.mktime(times[comments]) <= second:
            comments += 1
        writer.writerow([comments])