_ Git - youtube-dl/blob - youtube_dl/extractor/videoesri.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import os
   5 import re
   6
   7 from .common import InfoExtractor
   8
   9 from ..utils import (
  10     unified_strdate
  11 )
  12
  13
  14 class VideoEsriIE(InfoExtractor):
  15     _VALID_URL = r'https?://video\.esri\.com/watch/(?P<id>[0-9]+)'
  16     _TEST = {
  17         'url': 'https://video.esri.com/watch/4228',
  18         'md5': '170b4d513c2466ed483c150a48384133',
  19         'info_dict': {
  20             'id': '4228',
  21             'ext': 'mp4',
  22             'title': 'AppStudio for ArcGIS',
  23             'thumbnail': 're:^https?://.*\.jpg$',
  24             'upload_date': '20150310',
  25         }
  26     }
  27
  28     def _real_extract(self, url):
  29         video_id = self._match_id(url)
  30         webpage = self._download_webpage(url, video_id)
  31
  32         title = self._html_search_regex(r'<h1>(.*?)</h1>', webpage, 'title')
  33
  34         upload_date_raw = self._search_regex(
  35             r'http-equiv="last-modified" content="(.*)"',
  36             webpage, 'upload date')
  37         upload_date = unified_strdate(upload_date_raw)
  38
  39         settings_info = self._search_regex(
  40             r'evPlayerSettings = {(.*?);\s*$',
  41             webpage, 'settings info', flags=re.MULTILINE | re.DOTALL)
  42
  43         # thumbnail includes '_x' for large, also has {_m,_t,_s} or
  44         # without size suffix returns full image
  45         thumbnail_path = re.findall(
  46             r'image\': \'(\/thumbs.*)\'',
  47             settings_info)[0]
  48
  49         if thumbnail_path:
  50             thumbnail = '/'.join(['http://video.esri.com', thumbnail_path])
  51
  52         # note that this misses the (exceedly rare) webm files
  53         video_paths = re.findall(r'mp4:(.*)\'', settings_info)
  54
  55         # find possible http servers of the mp4 files (also has rtsp)
  56         base_url = re.findall(
  57             r'netstreambasepath\':\s\'(h.*)\'', settings_info)[0]
  58
  59         # these are the numbers used internally, but really map
  60         # to other resolutions, e.g. 960 is 720p.
  61         heights = [480, 720, 960]
  62         videos_by_res = {}
  63         for video_path in video_paths:
  64             url = "{base_url}{video_path}".format(
  65                 base_url=base_url,
  66                 video_path=video_path)
  67             filename, ext = os.path.splitext(video_path)
  68             height_label = int(filename.split('_')[1])
  69             videos_by_res[height_label] = {
  70                 'url': url,
  71                 'ext': ext[1:],
  72                 'protocol': 'http',  # http-only supported currently
  73             }
  74
  75         formats = []
  76         for height in heights:
  77             if height in videos_by_res:
  78                 formats.append(videos_by_res[height])
  79
  80         result = {
  81             'id': video_id,
  82             'title': title,
  83             'upload_date': upload_date,
  84             'formats': formats,
  85         }
  86
  87         if thumbnail:
  88             result['thumbnail'] = thumbnail
  89
  90         return result