filename_annotations = 'https://docs.google.com/\
spreadsheets/d/1Rqu1sJiD-ogc4a6R491JTiaYacptOTqh6DKqhwTa8NA/gviz/tq?tqx=out:csv&sheet=Template'
#Retrieve labels
df_annotations = pd.read_csv(filename_annotations,
header=None).drop([0, 1, 2, 3])
video_names = set(df_annotations[1].values)
dict = {}
for (j, video) in enumerate(video_names):
for i in range(1, 5):
text_file = f'{video}_{i}.txt'
label = get_annotations_video(filename_annotations, video, 'max')[2]
gender = get_annotations_video(filename_annotations, video, 'max')[4]
gender_bool = 1.0 if gender == 'H' else 0.0
group = j
dict[text_file] = (label[i - 1], gender_bool, group)
df_labels = pd.DataFrame.from_dict(dict,
columns=['Label', 'Gender', 'Group'],
orient='index')
text_files = df_labels.index
#Retrieve Word Embedding
X_vect = np.load(file_word_embedding)
#videos_excluded = ['WIN_20210329_14_13_45_Pro','WIN_20210402_14_27_50_Pro']
#multi_feat = multi_feat.drop(videos_excluded,axis=0)
#Retrieve labels
df_annotations = pd.read_csv(filename_annotations,
header=None).drop([0, 1, 2, 3])
diapos = [1, 8, 9, 10, 11, 12, 17, 18]
video_names = set(df_annotations[1].values)
dict = {}
for video_name in video_names:
labels = get_annotations_video(filename_annotations, video_name, 'max')[2]
dict_diapo = {}
for (i, diapo) in enumerate(diapos):
dict_diapo[diapo] = labels[i]
dict[video_name] = dict_diapo
df_labels = pd.DataFrame.from_dict(dict, orient='index')
df_labels = df_labels.stack()
#df_labels = df_labels.drop(videos_excluded,axis=0)
#Merge multi_feeatures and labels
data = pd.concat([multi_feat, df_labels], axis=1)