Coverage for src/bob/bio/video/database/youtube.py: 100%

1import logging 

2import os 

3 

4from clapper.rc import UserDefaults 

5 

6from bob.bio.base.database import CSVDatabase 

7 

8logger = logging.getLogger(__name__) 

9rc = UserDefaults("bobrc.toml") 

10 

11 

12class YoutubeDatabase(CSVDatabase): 

13 

14 """ 

15 This package contains the access API and descriptions for the `YouTube Faces` database. 

16 It only contains the Bob accessor methods to use the DB directly from python, with our certified protocols. 

17 The actual raw data for the `YouTube Faces` database should be downloaded from the original URL (though we were not able to contact the corresponding Professor). 

18 

19 .. warning:: 

20 

21 To use this dataset protocol, you need to have the original files of the YOUTUBE datasets. 

22 Once you have it downloaded, please run the following command to set the path for Bob 

23 

24 .. code-block:: sh 

25 

26 bob config set bob.bio.face.youtube.directory [YOUTUBE PATH] 

27 

28 

29 

30 In this interface we implement the 10 original protocols of the `YouTube Faces` database ('fold1', 'fold2', 'fold3', 'fold4', 'fold5', 'fold6', 'fold7', 'fold8', 'fold9', 'fold10') 

31 

32 

33 The code below allows you to fetch the gallery and probes of the "fold0" protocol. 

34 

35 .. code-block:: python 

36 

37 >>> from bob.bio.video.database import YoutubeDatabase 

38 >>> youtube = YoutubeDatabase(protocol="fold0") 

39 >>> 

40 >>> # Fetching the gallery 

41 >>> references = youtube.references() 

42 >>> # Fetching the probes 

43 >>> probes = youtube.probes() 

44 

45 

46 Parameters 

47 ---------- 

48 

49 protocol: str 

50 One of the Youtube above mentioned protocols 

51 

52 annotation_type: str 

53 One of the supported annotation types 

54 

55 original_directory: str 

56 Original directory 

57 

58 extension: str 

59 Default file extension 

60 

61 annotation_extension: str 

62 

63 frame_selector: 

64 Pointer to a function that does frame selection. 

65 

66 """ 

67 

68 name = "youtube" 

69 category = "video" 

70 dataset_protocols_name = "youtube.tar.gz" 

71 dataset_protocols_urls = [ 

72 "https://www.idiap.ch/software/bob/databases/latest/video/youtube-51c1fb2a.tar.gz", 

73 "http://www.idiap.ch/software/bob/databases/latest/video/youtube-51c1fb2a.tar.gz", 

74 ] 

75 dataset_protocols_hash = "51c1fb2a" 

76 

77 def __init__( 

78 self, 

79 protocol, 

80 annotation_type="bounding-box", 

81 fixed_positions=None, 

82 original_directory=rc.get("bob.bio.video.youtube.directory", ""), 

83 extension=".jpg", 

84 annotation_extension=".labeled_faces.txt", 

85 frame_selector=None, 

86 ): 

87 original_directory = original_directory or "" 

88 if not os.path.exists(original_directory): 

89 logger.warning( 

90 f"Invalid or non existent `original_directory`: {original_directory}." 

91 "Please, do `bob config set bob.bio.video.youtube.directory PATH` to set the Youtube data directory." 

92 ) 

93 

94 self.references_dict = {} 

95 self.probes_dict = {} 

96 

97 # Dict that holds a `subject_id` as a key and has 

98 # filenames as values 

99 self.subject_id_files = {} 

100 self.template_id_to_subject_id = None 

101 self.template_id_to_sample = None 

102 self.original_directory = original_directory 

103 self.extension = extension 

104 self.annotation_extension = annotation_extension 

105 self.frame_selector = frame_selector 

106 

107 super().__init__( 

108 name=self.name, 

109 protocol=protocol, 

110 annotation_type=annotation_type, 

111 fixed_positions=fixed_positions, 

112 memory_demanding=True, 

113 )