Generate xmltv from Firstmedia EPG

fetch.py

#!/usr/bin/env python2

import sys
import httplib
import json
import xml.etree.ElementTree as ET
from datetime import datetime

channels = {
  'net': 'NET.Digital',
  'kompas-tv': 'KOMPAS TV',
  'metro-tv': 'METRO TV HD',
  'tv-one': 'TVONE DIGITAL',
  'trans-tv': 'TRANS TV HD',
  'trans-7': 'TRANS7 HD TV',
  'tvri-nasional': 'TVRI HD',
  'indosiar': 'INDOSIAR NETWORK'
}

def main():
  if len(sys.argv) > 2:
    current_date = sys.argv[1]
  else:
    current_date = datetime.now().strftime('%Y-%m-%d')
  epg = fetch_epg(','.join(channels.keys()), current_date)
  xml = build_xml(epg, current_date)
  #print('<?xml version="1.0" encoding="ISO-8859-1"?><!DOCTYPE tv SYSTEM "xmltv.dtd">')
  ET.dump(xml)

def fetch_epg(channel, date):
  client = httplib.HTTPConnection('jkt.firstmedia.com')
  url = '/tv-guides/selected_date:%s/selected_channels:%s/get.json' % (date, channel)
  headers = {"X-Requested-With": "XMLHttpRequest", "Referrer": "http://jkt.firstmedia.com/tv-guides/"}
  
  client.request('GET', url, headers=headers)
  resp = client.getresponse().read()
  return json.loads(resp)


def build_xml(epg, current_date):
  root = ET.Element('tv', attrib={'source-info-url': 'http://jkt.firstmedia.com/tv-guides/'})
  for channel in epg:
    channel_id = channel['Channel']['slug']
    xch = ET.SubElement(root, 'channel', attrib={'id': channel_id})
    if channel_id in channels:
      ET.SubElement(xch, 'display-name').text=channels[channel_id]
    ET.SubElement(xch, 'display-name').text=channel['Channel']['title']
    
    programs = channel['Schedule']
    if isinstance(programs, dict):
      programs = programs.values()
    for program in programs:
      duration = int(program['length']) * 60
      
      # bugshit
      if duration > 0 and program['date'][0:10] == current_date and program['title'] != 'no schedule':
        start_time_unix = int(program['unix_date']) / 1000
        start_time = datetime.fromtimestamp(start_time_unix)
        stop_time = datetime.fromtimestamp(start_time_unix + duration)
        attrib = {
          'channel': channel_id,
          'start': start_time.strftime('%Y%m%d%H%M%S +0700'),
          'stop': stop_time.strftime('%Y%m%d%H%M%S +0700')
        }
        xprog = ET.SubElement(root, 'programme', attrib=attrib)
        ET.SubElement(xprog, 'title', attrib={'lang':'id'}).text = program['title']
        ET.SubElement(xprog, 'desc', attrib={'lang':'id'}).text=program['long_description']
  return root

def log(s):
  sys.stderr.write(s + "\n")

main()

fetch.sh

#!/bin/sh

set -euo pipefail

SCRIPT=$(readlink -f "$0")
SCRIPTPATH=$(dirname "$SCRIPT")

PYSCRIPT=$SCRIPTPATH/fetch.py

CACHEDIR=$SCRIPTPATH/cache

DATE=$(date +%Y-%m-%d)
CACHEFILE=$CACHEDIR/$DATE.xml

mkdir -p $CACHEDIR
[ -f $CACHEFILE ] || $PYSCRIPT $DATE > $CACHEFILE
cat $CACHEFILE