|
|
@ -28,7 +28,7 @@ def unpadded_b64_decode(s):
|
|
|
|
|
|
|
|
|
|
|
|
class SegmentInfo(
|
|
|
|
class SegmentInfo(
|
|
|
|
namedtuple('SegmentInfoBase', [
|
|
|
|
namedtuple('SegmentInfoBase', [
|
|
|
|
'path', 'stream', 'variant', 'start', 'duration', 'is_partial', 'hash'
|
|
|
|
'path', 'stream', 'variant', 'start', 'duration', 'type', 'hash'
|
|
|
|
])
|
|
|
|
])
|
|
|
|
):
|
|
|
|
):
|
|
|
|
"""Info parsed from a segment path, including original path.
|
|
|
|
"""Info parsed from a segment path, including original path.
|
|
|
@ -36,6 +36,9 @@ class SegmentInfo(
|
|
|
|
@property
|
|
|
|
@property
|
|
|
|
def end(self):
|
|
|
|
def end(self):
|
|
|
|
return self.start + self.duration
|
|
|
|
return self.start + self.duration
|
|
|
|
|
|
|
|
@property
|
|
|
|
|
|
|
|
def is_partial(self):
|
|
|
|
|
|
|
|
return self.type != "full"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def parse_segment_path(path):
|
|
|
|
def parse_segment_path(path):
|
|
|
@ -64,7 +67,7 @@ def parse_segment_path(path):
|
|
|
|
variant = variant,
|
|
|
|
variant = variant,
|
|
|
|
start = datetime.datetime.strptime("{}:{}".format(hour, time), "%Y-%m-%dT%H:%M:%S.%f"),
|
|
|
|
start = datetime.datetime.strptime("{}:{}".format(hour, time), "%Y-%m-%dT%H:%M:%S.%f"),
|
|
|
|
duration = datetime.timedelta(seconds=float(duration)),
|
|
|
|
duration = datetime.timedelta(seconds=float(duration)),
|
|
|
|
is_partial = type != "full",
|
|
|
|
type = type,
|
|
|
|
hash = hash,
|
|
|
|
hash = hash,
|
|
|
|
)
|
|
|
|
)
|
|
|
|
except ValueError as e:
|
|
|
|
except ValueError as e:
|
|
|
@ -219,7 +222,9 @@ def best_segments_by_start(hour):
|
|
|
|
# but is easy enough to do, so we might as well.
|
|
|
|
# but is easy enough to do, so we might as well.
|
|
|
|
parsed = (parse_segment_path(os.path.join(hour, name)) for name in segment_paths)
|
|
|
|
parsed = (parse_segment_path(os.path.join(hour, name)) for name in segment_paths)
|
|
|
|
for start_time, segments in itertools.groupby(parsed, key=lambda segment: segment.start):
|
|
|
|
for start_time, segments in itertools.groupby(parsed, key=lambda segment: segment.start):
|
|
|
|
segments = list(segments)
|
|
|
|
# ignore temp segments as they might go away by the time we want to use them
|
|
|
|
|
|
|
|
segments = [segment for segment in segments if segment.type != "temp"]
|
|
|
|
|
|
|
|
|
|
|
|
full_segments = [segment for segment in segments if not segment.is_partial]
|
|
|
|
full_segments = [segment for segment in segments if not segment.is_partial]
|
|
|
|
if full_segments:
|
|
|
|
if full_segments:
|
|
|
|
if len(full_segments) != 1:
|
|
|
|
if len(full_segments) != 1:
|
|
|
|