Script parses Radio T chat log and outputs each subject starting time

radio-t-timestamps.py

# -*- coding: utf-8 -*-

import lxml.html
import math
from datetime import datetime

radiot_nick = 'jc-radio-t'
podcast = 'http://chat.radio-t.com/logs/radio-t-{number}.html'.format(number=422)

# parsing
page = lxml.html.parse(podcast)
root = page.getroot()
chat_messages = root.find_class('message')

# filter messages
radiot_subjects = []
for message in chat_messages:
    user_nick = message.find_class('nickname')[0].text_content()
    user_message = message.find_class('text')[0].text_content()
    user_date = message.find_class('date')[0].text_content()

    if ((user_nick == radiot_nick) and
            (user_message.find('-->') >= 0) and
            (user_message.find(u'Подкаст выходного дня - импровизации на темы высоких технологий') == -1)):
        radiot_subjects.append({'subject': user_message, 'date': user_date})

# print
# for subject in radiot_subjects:
#    print subject['date'], subject['subject']

# calculate and print each subject and its starting time
podcast_length = 0
for ind, subject in enumerate(radiot_subjects):
    if ind == len(radiot_subjects) - 1:
        radiot_subjects[ind]['start'] = podcast_length
        break
    seconds = (datetime.strptime(radiot_subjects[ind+1]['date'], '%H:%M:%S') - datetime.strptime(radiot_subjects[ind]['date'], '%H:%M:%S')).seconds
    radiot_subjects[ind]['start'] = podcast_length
    podcast_length += int(math.ceil(seconds / 60.0))

# print the result
for subject in radiot_subjects:
    print u"{stime}:00 - {subject}".format(stime=subject['start'], subject=subject['subject'])