def start_full_process():
global THRET_VAL
st_time = timeit.default_timer()
vodid = _G.StreamFileIndex
data = get_video_info(vodid)
print(f"Retrieved vod data: {data}")
data = data.json()['data'][0]
print(f"VodId: {vodid}\n{data}")
_G.StreamFilePrefix = data['user_name'].upper()
_G.FLAG_POSITIVE_PROC = False
_G.init()
duration = hms2sec(data['duration'])
print(f"Duration: {data['duration']} ({duration})")
download_full_vod(vodid, duration)
errno, stream_filename = None, ''
while True:
try:
errno, stream_filename = THRET_VAL[0]
break
except Exception:
_G.wait(0.3)
inp = ''
if _G.FLAG_ALWAYS_YES:
inp = 'Y'
elif _G.FLAG_ALWAYS_NO:
inp = 'N'
if errno == ERRNO_EXIST:
while inp != 'Y' and inp != 'N':
inp = input("Stream video already exists, process anyway? (Y/N): "
).upper()
if errno == ERRNO_OK or inp == 'Y':
_G.wait(1)
clip.spawn_extracting_proc(_G.StreamFileIndex, None,
_G.StreamFilePrefix, _G.PROC_FULL)
_G.wait(1)
analyzer.spawn_analyze_proc(_G.StreamFileIndex, None,
_G.StreamFilePrefix, _G.PROC_FULL)
print("Complete, time taken: ", timeit.default_timer() - st_time)
from sklearn.neighbors import KNeighborsClassifier
from sklearn.ensemble import RandomForestRegressor
from collections import defaultdict
from pprint import pprint
VERBOSE = 1
N_JOBS = 1
GRID_CV = 5
Category = 'zcr'
TRAIN_SVM = True
TRAIN_KNN = True
TRAIN_RFR = True
if __name__ == "__main__":
argv_parse.init()
_G.init()
N_JOBS = -1
# parts: splited path of the origin data
# used to locate postive label file path
def load_postive_label(parts):
broardcaster = parts[1]
episode = parts[2]
filename = f"{_G.PositiveSampleFolder}/{broardcaster}/{episode}/{_G.PostiveLabelFilename}"
ret = {}
with open(filename, 'r') as fp:
sam_datas = json.load(fp)
for sdata in sam_datas:
slug = sdata['slug']
st = sdata['start_t']
def start_sample_process():
global THRET_VAL
st_time = timeit.default_timer()
slug = _G.ClipName
data = None
_cnt = 0
while not data and _cnt < 3:
_cnt += 1
data = getkarkan_clip_info(slug).json()
if not data:
print(f"Failed to get clip data ({data})...retry({_cnt})")
time.sleep(1)
if not data['vod']:
print("Source unavailable, throwing")
raise RuntimeError("Source VOD unavailable")
id = get_id_from_data(data)
start_t = get_ori_timestamp(data)
_G.StreamFileIndex = int(id)
_G.StreamFilePrefix = data['broadcaster']['name'].upper()
_G.FLAG_POSITIVE_PROC = True
_G.init()
errno, clip_fname = download_clip(id, slug)
inp = ''
if _G.FLAG_ALWAYS_YES:
inp = 'Y'
elif _G.FLAG_ALWAYS_NO:
inp = 'N'
if errno == ERRNO_EXIST:
print(f"{clip_fname} already downloaded, skip")
while inp != 'Y' and inp != 'N':
inp = input("Process clip anyway? (Y/N): ").upper()
if errno == ERRNO_OK or inp == 'Y':
print("Generating labels...")
generate_label(clip_fname, slug, start_t)
print("Extracting clips audio")
_G.wait(1)
clip.spawn_extracting_proc(_G.StreamFileIndex, slug,
_G.StreamFilePrefix, _G.PROC_SAMPLE_POS)
print("Analyzing clip...")
_G.wait(1)
analyzer.spawn_analyze_proc(_G.StreamFileIndex, slug,
_G.StreamFilePrefix, _G.PROC_SAMPLE_POS)
ori_st, duration = _G.get_download_timeinfo(start_t)
print(f"Ori. start time/duraiton: ", ori_st, duration)
THRET_VAL = {}
download_m3u8(id, slug, ori_st, duration)
errno, stream_filename = None, ''
while True:
try:
errno, stream_filename = THRET_VAL[0]
break
except Exception:
_G.wait(0.3)
inp = ''
if _G.FLAG_ALWAYS_YES:
inp = 'Y'
elif _G.FLAG_ALWAYS_NO:
inp = 'N'
if errno == ERRNO_EXIST:
while inp != 'Y' and inp != 'N':
inp = input("Stream video already exists, process anyway? (Y/N): "
).upper()
if errno == ERRNO_OK or inp == 'Y':
_G.wait(1)
clip.spawn_extracting_proc(_G.StreamFileIndex, slug,
_G.StreamFilePrefix, False)
_G.wait(1)
analyzer.spawn_analyze_proc(_G.StreamFileIndex, slug,
_G.StreamFilePrefix, False)
print("Complete, time taken: ", timeit.default_timer() - st_time)