def run(argv):
# Now let's parse the arguements
try:
opts, args = getopt.getopt(argv,'hi:o:w:')
except getopt.GetoptError:
print "You did something wrong"
sys.exit(0)
video_file = None
output_folder = None
wavfile_dir = None
for opt, arg in opts:
if opt in ('-h'):
print "HELP!"
sys.exit(0)
elif opt in ('-i'):
video_file = arg
elif opt in ('-o'):
output_folder = arg
elif opt in ('-w'):
wavfile_dir = arg
if not (video_file and output_folder and wavfile_dir):
print "You need more arguments to run this code"
sys.exit(0)
### THIS SECTION OF CODE USES THE SHoUT Toolbox ########
# Now do the diarization to start us out
file_id_with_ext = reader.GetFileOnly(video_file)
file_id = reader.ReplaceExt(file_id_with_ext,"")
raw_file = file_id + ".raw"
raw_file = os.path.join(wavfile_dir,raw_file)
# Perform turning the video into ."raw" audiofile
return_code = shout_vid2wav(video_file,raw_file)
if return_code:
print "We had an issue in: VIDEO TRANSCODING"
sys.exit(0)
# Perform segmentation
seg_file = os.path.join(output_folder,file_id + ".seg")
return_code = shout_segment(raw_file,seg_file)
if return_code:
print "We had an issue in: SEGMENTATION"
sys.exit(0)
# Perform Diarization
dia_file = reader.ReplaceExt(seg_file,".dia")
return_code = shout_cluster(raw_file,seg_file,dia_file)
if return_code:
print "We had an issue in: CLUSTERING"
sys.exit(0)
###################################################
# Now let's parse through the clustering file, and find the speaker times for each video
person_segs = reader.read_diafile(dia_file)
# Now we need to combine the sections of speech that are concurrent between people
segments = reader.ConnectSpkrSegs(person_segs)
# This cuts up the video and outputs it to the desired directory
# Create a directory to hold all of the videos for this particular youtube program
output_vid_segs_dir = os.path.join(output_folder,file_id)
if not os.path.exists(output_vid_segs_dir):
os.makedirs(output_vid_segs_dir)
####### Video Cutting ########
# This is for if we just want to use the smaller non-connected speech segments
#cut_video(person_segs,video_file,output_vid_segs_dir,file_id)
# This is for if we want to use the diarization results
cut_video(segments,video_file,output_vid_segs_dir,file_id)
# Now we need to output a file that has the time segments available
output_seg_time_file = os.path.join(output_vid_segs_dir,file_id + ".times")
with open(output_seg_time_file, "w") as f:
for i,segment in enumerate(segments):
output = "%03d,%.2f,%.2f\n" % (i,segment[0],segment[0]+segment[1])
f.write(output)
# TODO: NEED TO REMOVE THE AUDIO FILES THAT ARE NOT NEEDED BUT CREATED
os.remove(raw_file)
print "Finished Processing Video: %s" % file_id
