You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

128 lines
4.7 KiB

  1. from __future__ import unicode_literals
  2. import re
  3. from .common import InfoExtractor
  4. from ..utils import (
  5. parse_duration,
  6. parse_iso8601,
  7. )
  8. class MLBIE(InfoExtractor):
  9. _VALID_URL = r'https?://m(?:lb)?\.mlb\.com/(?:(?:.*?/)?video/(?:topic/[\da-z_-]+/)?v|(?:shared/video/embed/embed\.html|[^/]+/video/play\.jsp)\?.*?\bcontent_id=)(?P<id>n?\d+)'
  10. _TESTS = [
  11. {
  12. 'url': 'http://m.mlb.com/sea/video/topic/51231442/v34698933/nymsea-ackley-robs-a-home-run-with-an-amazing-catch/?c_id=sea',
  13. 'md5': 'ff56a598c2cf411a9a38a69709e97079',
  14. 'info_dict': {
  15. 'id': '34698933',
  16. 'ext': 'mp4',
  17. 'title': "Ackley's spectacular catch",
  18. 'description': 'md5:7f5a981eb4f3cbc8daf2aeffa2215bf0',
  19. 'duration': 66,
  20. 'timestamp': 1405980600,
  21. 'upload_date': '20140721',
  22. 'thumbnail': 're:^https?://.*\.jpg$',
  23. },
  24. },
  25. {
  26. 'url': 'http://m.mlb.com/video/topic/81536970/v34496663/mianym-stanton-practices-for-the-home-run-derby',
  27. 'md5': 'd9c022c10d21f849f49c05ae12a8a7e9',
  28. 'info_dict': {
  29. 'id': '34496663',
  30. 'ext': 'mp4',
  31. 'title': 'Stanton prepares for Derby',
  32. 'description': 'md5:d00ce1e5fd9c9069e9c13ab4faedfa57',
  33. 'duration': 46,
  34. 'timestamp': 1405105800,
  35. 'upload_date': '20140711',
  36. 'thumbnail': 're:^https?://.*\.jpg$',
  37. },
  38. },
  39. {
  40. 'url': 'http://m.mlb.com/video/topic/vtp_hrd_sponsor/v34578115/hrd-cespedes-wins-2014-gillette-home-run-derby',
  41. 'md5': '0e6e73d509321e142409b695eadd541f',
  42. 'info_dict': {
  43. 'id': '34578115',
  44. 'ext': 'mp4',
  45. 'title': 'Cespedes repeats as Derby champ',
  46. 'description': 'md5:08df253ce265d4cf6fb09f581fafad07',
  47. 'duration': 488,
  48. 'timestamp': 1405399936,
  49. 'upload_date': '20140715',
  50. 'thumbnail': 're:^https?://.*\.jpg$',
  51. },
  52. },
  53. {
  54. 'url': 'http://m.mlb.com/video/v34577915/bautista-on-derby-captaining-duties-his-performance',
  55. 'md5': 'b8fd237347b844365d74ea61d4245967',
  56. 'info_dict': {
  57. 'id': '34577915',
  58. 'ext': 'mp4',
  59. 'title': 'Bautista on Home Run Derby',
  60. 'description': 'md5:b80b34031143d0986dddc64a8839f0fb',
  61. 'duration': 52,
  62. 'timestamp': 1405390722,
  63. 'upload_date': '20140715',
  64. 'thumbnail': 're:^https?://.*\.jpg$',
  65. },
  66. },
  67. {
  68. 'url': 'http://m.mlb.com/shared/video/embed/embed.html?content_id=35692085&topic_id=6479266&width=400&height=224&property=mlb',
  69. 'only_matching': True,
  70. },
  71. {
  72. 'url': 'http://mlb.mlb.com/shared/video/embed/embed.html?content_id=36599553',
  73. 'only_matching': True,
  74. },
  75. {
  76. 'url': 'http://mlb.mlb.com/es/video/play.jsp?content_id=36599553',
  77. 'only_matching': True,
  78. },
  79. ]
  80. def _real_extract(self, url):
  81. mobj = re.match(self._VALID_URL, url)
  82. video_id = mobj.group('id')
  83. detail = self._download_xml(
  84. 'http://m.mlb.com/gen/multimedia/detail/%s/%s/%s/%s.xml'
  85. % (video_id[-3], video_id[-2], video_id[-1], video_id), video_id)
  86. title = detail.find('./headline').text
  87. description = detail.find('./big-blurb').text
  88. duration = parse_duration(detail.find('./duration').text)
  89. timestamp = parse_iso8601(detail.attrib['date'][:-5])
  90. thumbnails = [{
  91. 'url': thumbnail.text,
  92. } for thumbnail in detail.findall('./thumbnailScenarios/thumbnailScenario')]
  93. formats = []
  94. for media_url in detail.findall('./url'):
  95. playback_scenario = media_url.attrib['playback_scenario']
  96. fmt = {
  97. 'url': media_url.text,
  98. 'format_id': playback_scenario,
  99. }
  100. m = re.search(r'(?P<vbr>\d+)K_(?P<width>\d+)X(?P<height>\d+)', playback_scenario)
  101. if m:
  102. fmt.update({
  103. 'vbr': int(m.group('vbr')) * 1000,
  104. 'width': int(m.group('width')),
  105. 'height': int(m.group('height')),
  106. })
  107. formats.append(fmt)
  108. self._sort_formats(formats)
  109. return {
  110. 'id': video_id,
  111. 'title': title,
  112. 'description': description,
  113. 'duration': duration,
  114. 'timestamp': timestamp,
  115. 'formats': formats,
  116. 'thumbnails': thumbnails,
  117. }