scrippsnetworks.py 6.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183
  1. # coding: utf-8
  2. from __future__ import unicode_literals
  3. import datetime
  4. import json
  5. import hashlib
  6. import hmac
  7. import re
  8. from .common import InfoExtractor
  9. from .anvato import AnvatoIE
  10. from ..utils import (
  11. urlencode_postdata,
  12. xpath_text,
  13. )
  14. class ScrippsNetworksWatchIE(InfoExtractor):
  15. IE_NAME = 'scrippsnetworks:watch'
  16. _VALID_URL = r'''(?x)
  17. https?://
  18. watch\.
  19. (?P<site>hgtv|foodnetwork|travelchannel|diynetwork|cookingchanneltv)\.com/
  20. (?:
  21. player\.[A-Z0-9]+\.html\#|
  22. show/(?:[^/]+/){2}
  23. )
  24. (?P<id>\d+)
  25. '''
  26. _TESTS = [{
  27. 'url': 'http://watch.hgtv.com/show/HGTVE/Best-Ever-Treehouses/2241515/Best-Ever-Treehouses/',
  28. 'md5': '26545fd676d939954c6808274bdb905a',
  29. 'info_dict': {
  30. 'id': '4173834',
  31. 'ext': 'mp4',
  32. 'title': 'Best Ever Treehouses',
  33. 'description': "We're searching for the most over the top treehouses.",
  34. 'uploader': 'ANV',
  35. 'upload_date': '20170922',
  36. 'timestamp': 1506056400,
  37. },
  38. 'params': {
  39. 'skip_download': True,
  40. },
  41. 'add_ie': [AnvatoIE.ie_key()],
  42. }, {
  43. 'url': 'http://watch.diynetwork.com/show/DSAL/Salvage-Dawgs/2656646/Covington-Church/',
  44. 'only_matching': True,
  45. }, {
  46. 'url': 'http://watch.diynetwork.com/player.HNT.html#2656646',
  47. 'only_matching': True,
  48. }]
  49. _SNI_TABLE = {
  50. 'hgtv': 'hgtv',
  51. 'diynetwork': 'diy',
  52. 'foodnetwork': 'food',
  53. 'cookingchanneltv': 'cook',
  54. 'travelchannel': 'trav',
  55. 'geniuskitchen': 'geniuskitchen',
  56. }
  57. _SNI_HOST = 'web.api.video.snidigital.com'
  58. _AWS_REGION = 'us-east-1'
  59. _AWS_IDENTITY_ID_JSON = json.dumps({
  60. 'IdentityId': '%s:7655847c-0ae7-4d9b-80d6-56c062927eb3' % _AWS_REGION
  61. })
  62. _AWS_USER_AGENT = 'aws-sdk-js/2.80.0 callback'
  63. _AWS_API_KEY = 'E7wSQmq0qK6xPrF13WmzKiHo4BQ7tip4pQcSXVl1'
  64. _AWS_SERVICE = 'execute-api'
  65. _AWS_REQUEST = 'aws4_request'
  66. _AWS_SIGNED_HEADERS = ';'.join([
  67. 'host', 'x-amz-date', 'x-amz-security-token', 'x-api-key'])
  68. _AWS_CANONICAL_REQUEST_TEMPLATE = '''GET
  69. %(uri)s
  70. host:%(host)s
  71. x-amz-date:%(date)s
  72. x-amz-security-token:%(token)s
  73. x-api-key:%(key)s
  74. %(signed_headers)s
  75. %(payload_hash)s'''
  76. def _real_extract(self, url):
  77. mobj = re.match(self._VALID_URL, url)
  78. site_id, video_id = mobj.group('site', 'id')
  79. def aws_hash(s):
  80. return hashlib.sha256(s.encode('utf-8')).hexdigest()
  81. token = self._download_json(
  82. 'https://cognito-identity.us-east-1.amazonaws.com/', video_id,
  83. data=self._AWS_IDENTITY_ID_JSON.encode('utf-8'),
  84. headers={
  85. 'Accept': '*/*',
  86. 'Content-Type': 'application/x-amz-json-1.1',
  87. 'Referer': url,
  88. 'X-Amz-Content-Sha256': aws_hash(self._AWS_IDENTITY_ID_JSON),
  89. 'X-Amz-Target': 'AWSCognitoIdentityService.GetOpenIdToken',
  90. 'X-Amz-User-Agent': self._AWS_USER_AGENT,
  91. })['Token']
  92. sts = self._download_xml(
  93. 'https://sts.amazonaws.com/', video_id, data=urlencode_postdata({
  94. 'Action': 'AssumeRoleWithWebIdentity',
  95. 'RoleArn': 'arn:aws:iam::710330595350:role/Cognito_WebAPIUnauth_Role',
  96. 'RoleSessionName': 'web-identity',
  97. 'Version': '2011-06-15',
  98. 'WebIdentityToken': token,
  99. }), headers={
  100. 'Referer': url,
  101. 'X-Amz-User-Agent': self._AWS_USER_AGENT,
  102. 'Content-Type': 'application/x-www-form-urlencoded; charset=utf-8',
  103. })
  104. def get(key):
  105. return xpath_text(
  106. sts, './/{https://sts.amazonaws.com/doc/2011-06-15/}%s' % key,
  107. fatal=True)
  108. access_key_id = get('AccessKeyId')
  109. secret_access_key = get('SecretAccessKey')
  110. session_token = get('SessionToken')
  111. # Task 1: http://docs.aws.amazon.com/general/latest/gr/sigv4-create-canonical-request.html
  112. uri = '/1/web/brands/%s/episodes/scrid/%s' % (self._SNI_TABLE[site_id], video_id)
  113. datetime_now = datetime.datetime.utcnow().strftime('%Y%m%dT%H%M%SZ')
  114. date = datetime_now[:8]
  115. canonical_string = self._AWS_CANONICAL_REQUEST_TEMPLATE % {
  116. 'uri': uri,
  117. 'host': self._SNI_HOST,
  118. 'date': datetime_now,
  119. 'token': session_token,
  120. 'key': self._AWS_API_KEY,
  121. 'signed_headers': self._AWS_SIGNED_HEADERS,
  122. 'payload_hash': aws_hash(''),
  123. }
  124. # Task 2: http://docs.aws.amazon.com/general/latest/gr/sigv4-create-string-to-sign.html
  125. credential_string = '/'.join([date, self._AWS_REGION, self._AWS_SERVICE, self._AWS_REQUEST])
  126. string_to_sign = '\n'.join([
  127. 'AWS4-HMAC-SHA256', datetime_now, credential_string,
  128. aws_hash(canonical_string)])
  129. # Task 3: http://docs.aws.amazon.com/general/latest/gr/sigv4-calculate-signature.html
  130. def aws_hmac(key, msg):
  131. return hmac.new(key, msg.encode('utf-8'), hashlib.sha256)
  132. def aws_hmac_digest(key, msg):
  133. return aws_hmac(key, msg).digest()
  134. def aws_hmac_hexdigest(key, msg):
  135. return aws_hmac(key, msg).hexdigest()
  136. k_secret = 'AWS4' + secret_access_key
  137. k_date = aws_hmac_digest(k_secret.encode('utf-8'), date)
  138. k_region = aws_hmac_digest(k_date, self._AWS_REGION)
  139. k_service = aws_hmac_digest(k_region, self._AWS_SERVICE)
  140. k_signing = aws_hmac_digest(k_service, self._AWS_REQUEST)
  141. signature = aws_hmac_hexdigest(k_signing, string_to_sign)
  142. auth_header = ', '.join([
  143. 'AWS4-HMAC-SHA256 Credential=%s' % '/'.join(
  144. [access_key_id, date, self._AWS_REGION, self._AWS_SERVICE, self._AWS_REQUEST]),
  145. 'SignedHeaders=%s' % self._AWS_SIGNED_HEADERS,
  146. 'Signature=%s' % signature,
  147. ])
  148. mcp_id = self._download_json(
  149. 'https://%s%s' % (self._SNI_HOST, uri), video_id, headers={
  150. 'Accept': '*/*',
  151. 'Referer': url,
  152. 'Authorization': auth_header,
  153. 'X-Amz-Date': datetime_now,
  154. 'X-Amz-Security-Token': session_token,
  155. 'X-Api-Key': self._AWS_API_KEY,
  156. })['results'][0]['mcpId']
  157. return self.url_result(
  158. 'anvato:anvato_scripps_app_web_prod_0837996dbe373629133857ae9eb72e740424d80a:%s' % mcp_id,
  159. AnvatoIE.ie_key(), video_id=mcp_id)