cutcutcodec.core.analysis.ffprobe

Extract the properties of different streams of a multimedia file.

Functions

`get_metadata`(filename[, ignore_errors])	Call `ffprobe` and parse the result as a dictionary.
`get_slices_metadata`(filename[, slice_type])	Get the packets informations for all streams.
`get_streams_type`(filename[, ignore_errors])	Retrieve in order the stream types present in the file.
`parse_duration`(duration)	Try to convert a duration information into a fraction in second.

Details

cutcutcodec.core.analysis.ffprobe.get_metadata(filename: str | bytes | Path, ignore_errors=False) → dict[str][source]

Call ffprobe and parse the result as a dictionary.

Parameters

filenamepathlike: The pathlike of the file containing streams.
ignore_errorsboolean, default=False: If True, returns an empty dict rather than throwing an exception if invalid data are detected.

Returns

metadatadict: All the metadata containing in the container and each streams.

Examples

>>> from pprint import pprint
>>> from cutcutcodec.core.analysis.ffprobe import get_metadata
>>> pprint(get_metadata("cutcutcodec/examples/audio_5.1_narration.oga"))  
{'format': {'bit_rate': '34982',
            'duration': '8.000000',
            'filename': 'cutcutcodec/examples/audio_5.1_narration.oga',
            'format_long_name': 'Ogg',
            'format_name': 'ogg',
            ...
            'nb_streams': 1,
            'probe_score': 100,
            'size': '34982',
            'start_time': '0.000000'},
 'streams': [{'avg_frame_rate': '0/0',
              'bit_rate': '96000',
              'bits_per_sample': 0,
              'channel_layout': '5.1',
              'channels': 6,
              'codec_long_name': 'Vorbis',
              'codec_name': 'vorbis',
              'codec_tag': '0x0000',
              'codec_tag_string': '[0][0][0][0]',
              'codec_type': 'audio',
              'disposition': {'attached_pic': 0,
                              ...
                              'visual_impaired': 0},
              'duration': '8.000000',
              ...
              'sample_rate': '16000',
              'start_pts': 0,
              'start_time': '0.000000',
              'tags': {'encoder': 'Lavc59.37.100 libvorbis'},
              'time_base': '1/16000'}]}
>>>

cutcutcodec.core.analysis.ffprobe.get_slices_metadata(filename: str | bytes | Path, slice_type: str = 'frame') → tuple[list[list[str]], list[ndarray]][source]

Get the packets informations for all streams.

Parameters

filenamepathlike: The pathlike of the file containing streams.
slice_typestr: The type of slices to decode, “frame” or “packet”. “frame” is slower but more accurate and informative. “packet” is faster but less acurate.

Returns

headerslist[list[str]]: For each stream, the name of the columns.
infos: list[np.ndarray]: For each stream, the 2d str array. Each row correspond to one packet.

Examples

>>> from pprint import pprint
>>> from cutcutcodec.core.analysis.ffprobe import get_slices_metadata
>>> headers, data = (
...     get_slices_metadata("cutcutcodec/examples/audio_5.1_narration.oga", slice_type="packet")
... )
>>> pprint(headers)  
[['codec_type',
  ...
  'stream_index']]
>>> pprint(data)  
[array([['audio', ..., '0'],
       ['audio', ..., '0'],
       ... dtype='<U15')]
>>> headers, data = get_slices_metadata("cutcutcodec/examples/video.mp4", slice_type="packet")
>>> pprint(headers)
[['codec_type',
  'dts',
  'dts_time',
  'duration',
  'duration_time',
  'flags',
  'pos',
  'pts',
  'pts_time',
  'size',
  'stream_index']]
>>> pprint(data)
[array([['video', '-1024', '-0.080000', ..., '0.000000', '5156', '0'],
       ['video', '-512', '-0.040000', ..., '0.240000', '845', '0'],
       ['video', '0', '0.000000', ..., '0.120000', '372', '0'],
       ...,
       ['video', '202240', '15.800000', ..., '15.960000', '354', '0'],
       ['video', '202752', '15.840000', ..., '15.920000', '247', '0'],
       ['video', '203264', '15.880000', ..., '15.880000', '192', '0']],
      shape=(400, 11), dtype='<U9')]
>>>

cutcutcodec.core.analysis.ffprobe.get_streams_type(filename: str | bytes | Path, ignore_errors=False) → list[str][source]

Retrieve in order the stream types present in the file.

Parameters

filenamepathlike: The pathlike of the file containing streams.
ignore_errorsboolean, default=False: If True, returns an empty list rather than throwing an exception if no valid stream is detected.

Returns

streams_typelist[str]: Each item can be « audio », « subtitle » or « video ».

Raises

MissingStreamError: If ignore_errors is False and if one of the indexes is missing or redondant.

Examples

>>> from cutcutcodec.core.analysis.ffprobe import get_streams_type
>>> get_streams_type("cutcutcodec/examples/intro.webm")
['video', 'video', 'audio', 'audio']
>>> get_streams_type("cutcutcodec/__main__.py", ignore_errors=True)
[]
>>>

cutcutcodec.core.analysis.ffprobe.parse_duration(duration: Real | str) → None | Fraction[source]

Try to convert a duration information into a fraction in second.

Parameters

durationnumber or str: The duration to cast in integer

Returns

sec_durationFraction: The decoded duration in second.

Examples

>>> from cutcutcodec.core.analysis.ffprobe import parse_duration
>>> parse_duration(1.5)  # from float
Fraction(3, 2)
>>> parse_duration(2)  # from integer
Fraction(2, 1)
>>> parse_duration(".5")  # from float rep
Fraction(1, 2)
>>> parse_duration("1.")  # from float rep
Fraction(1, 1)
>>> parse_duration("1.5")  # from complete float rep
Fraction(3, 2)
>>> parse_duration("1:01:01")  # from h:m:s
Fraction(3661, 1)
>>>