def __init__(self, id):
SimpleModule.__init__(self, id)
config_parser = ConfigurationParser.get_instance()
self.buffering_until = int(config_parser.get_parameter('buffering_until'))
self.max_buffer_size = int(config_parser.get_parameter('max_buffer_size'))
self.playback_step = int(config_parser.get_parameter('playbak_step'))
self.url_mpd = config_parser.get_parameter('url_mpd')
# last pause started at time
self.pause_started_at = None
self.pauses_number = 0
# tag to verify if buffer has an minimal amount of data
self.buffer_initialization = True
# Does the player already started to download a segment?
self.already_downloading = False
# buffer itself
self.buffer = []
# the buffer played position
self.buffer_played = 0
# history of what was played in buffer
self.playback_history = []
#initialize with the first segment sequence number to download
self.segment_id = 1
self.parsed_mpd = ''
self.qi = []
self.timer = Timer.get_instance()
#threading playback
self.playback_thread = threading.Thread(target=self.handle_video_playback)
self.player_thread_events = threading.Event()
self.lock = threading.Lock()
self.kill_playback_thread = False
self.request_time = 0
self.playback_qi = OutVector()
self.playback_quality_qi = OutVector()
self.playback_pauses = OutVector()
self.playback = OutVector()
self.playback_buffer_size = OutVector()
self.throughput = OutVector()
self.whiteboard = Whiteboard.get_instance()
self.whiteboard.add_playback_history(self.playback.get_items())
self.whiteboard.add_playback_qi(self.playback_qi.get_items())
self.whiteboard.add_playback_pauses(self.playback_pauses.get_items())
self.whiteboard.add_playback_buffer_size(self.playback_buffer_size.get_items())
self.whiteboard.add_buffer(self.buffer)
self.whiteboard.add_max_buffer_size(self.max_buffer_size)
def __init__(self, id):
SimpleModule.__init__(self, id)
self.initial_time = 0
self.qi = []
# for traffic shaping
config_parser = ConfigurationParser.get_instance()
self.traffic_shaping_interval = int(
config_parser.get_parameter('traffic_shaping_profile_interval'))
self.traffic_shaping_seed = int(
config_parser.get_parameter('traffic_shaping_seed'))
self.traffic_shaping_values = []
# mark the current traffic shapping interval
self.current_traffic_shaping_interval = 0
self.traffic_shaping_sequence = []
# traffic shaping sequence position
self.tss_position = 0
# traffic shaping values position
self.tsv_position = 0
token = config_parser.get_parameter('traffic_shaping_profile_sequence')
for i in range(len(token)):
if token[i] == 'L':
self.traffic_shaping_sequence.append(0)
elif token[i] == 'M':
self.traffic_shaping_sequence.append(1)
elif token[i] == 'H':
self.traffic_shaping_sequence.append(2)
self.timer = Timer.get_instance()
def __init__(self, id):
IR2A.__init__(self, id)
self.qi = []
self.throughput = 0
self.request_time = 0
self.vM = 0
self.timer = Timer.get_instance()
self.pause_started_at = None
def __init__(self, id):
IR2A.__init__(self, id)
self.parsed_mpd = ''
self.qi = []
# Usamos o tempo para cálculo de vazão, usando o método descrito no documento de especificação.
self.timer = Timer.get_instance()
self.request_time = self.timer.get_started_time()
# Mantém-se uma lista das qualidades escolhidas durante a execução do programa.
# Ela é utilizada no algoritmo de decisão.
self.chosen_qi = [0]
# Também é mantida uma lista de valores de vazão medidos em cada ponto de escolha, para uso no algoritmo.
self.throughput = [0]
def image_upload(track_id,
tpid,
tpe,
seq,
image_type,
content,
task_seq='!!!',
batch='!!!'):
if not task_seq:
task_seq = '!!!'
if not batch:
batch = '!!!'
_timer = Timer()
_timer.stage_begin('image write')
if not image_write(track_id, tpid, tpe, seq, image_type, content, task_seq,
batch):
return False
_timer.stage_begin('set batch')
if not batch_manager.set_batch(track_id, tpid, tpe, seq, image_type,
batch):
return False
_timer.finish()
logger().debug(_timer.dump())
return True
def __init__(self, id):
IR2A.__init__(self, id)
self.parsed_mpd = ''
self.qi = []
self.current_quality = 0
self.segmentos_baixados = 0
self.timer = Timer.get_instance()
self.momento_requisicao = 0
self.taxa_bits = 0
self.menor_taxa = math.inf
self.maior_taxa = 0
self.historico_t = []
self.current_buffer = 0
self.quedas_consecutivas = 0
self.logger = None
def test_step(epoch, data, cfg, model, device):
# get raw output; this can be for several layers
outputs = pytorch_model_outputs(model, data, device)
# cls metrics
confmat = get_confmat_from_raw(outputs, data)
print(confmat)
acc = get_acc_from_confmat(confmat)
recall = get_recall_from_confmat(confmat)
precision = get_precision_from_confmat(confmat)
if cfg.smoothing:
sm_confmat = get_smoothed_confmat_from_raw(cfg, outputs, data)
print(sm_confmat)
sm_acc = get_acc_from_confmat(sm_confmat)
sm_recall = get_recall_from_confmat(sm_confmat)
sm_precision = get_precision_from_confmat(sm_confmat)
# ap = get_ap_from_raw(outputs,data)
# mAP = get_map_from_ap(ap,data)
# similarity metric
features = get_features_from_raw(outputs)
print("simmat")
t = Timer()
t.tic()
sim_mat = get_simmat_from_features(features)
print(sim_mat.shape)
t.toc()
print(t)
print("end [simmat]")
# store results
results = edict()
results.acc = acc
results.recall = recall
results.precision = precision
if cfg.smoothing:
results.sm_acc = sm_acc
results.sm_recall = sm_recall
results.sm_precision = sm_precision
# results.ap = ap
# results.mAP = mAP
print(results)
results.features = features
results.sim_mat = sim_mat
return results