pymediainfo package

Module contents

This module is a wrapper around the MediaInfo library.

class pymediainfo.MediaInfo(xml: str, encoding_errors: str = 'strict')

An object containing information about a media file.

MediaInfo objects can be created by directly calling code from libmediainfo (in this case, the library must be present on the system):

>>> pymediainfo.MediaInfo.parse("/path/to/file.mp4")

Alternatively, objects may be created from MediaInfo’s XML output. Such output can be obtained using the XML output format on versions older than v17.10 and the OLDXML format on newer versions.

Using such an XML file, we can create a MediaInfo object:

>>> with open("output.xml") as f:
...     mi = pymediainfo.MediaInfo(
  • xml (str) – XML output obtained from MediaInfo.
  • encoding_errors (str) – option to pass to str.encode()’s errors parameter before parsing xml.

xml.etree.ElementTree.ParseError – if passed invalid XML.



A list of Track objects which the media file contains. For instance:

>>> mi = pymediainfo.MediaInfo.parse("/path/to/file.mp4")
>>> for t in mi.tracks:
...     print(t)
<Track track_id='None', track_type='General'>
<Track track_id='1', track_type='Text'>

Returns:All Tracks of type Audio.
Return type:list of Tracks
classmethod can_parse(library_file: Optional[str] = None) → bool

Checks whether media files can be analyzed using libmediainfo.

Parameters:library_file (str) – path to the libmediainfo library, this should only be used if the library cannot be auto-detected.
Return type:bool
Returns:All Tracks of type General.
Return type:list of Tracks
Returns:All Tracks of type Image.
Return type:list of Tracks
Returns:All Tracks of type Menu.
Return type:list of Tracks
Returns:All Tracks of type Other.
Return type:list of Tracks
classmethod parse(filename: Any, library_file: Optional[str] = None, cover_data: bool = False, encoding_errors: str = 'strict', parse_speed: float = 0.5, full: bool = True, legacy_stream_display: bool = False, mediainfo_options: Optional[Dict[str, str]] = None, output: Optional[str] = None) → Union[str, pymediainfo.MediaInfo]

Analyze a media file using libmediainfo.


Because of the way the underlying library works, this method should not be called simultaneously from multiple threads with different arguments. Doing so will cause inconsistencies or failures by changing library options that are shared across threads.

  • filename (str or pathlib.Path or os.PathLike or file-like object.) – path to the media file or file-like object which will be analyzed. A URL can also be used if libmediainfo was compiled with CURL support.
  • library_file (str) – path to the libmediainfo library, this should only be used if the library cannot be auto-detected.
  • cover_data (bool) – whether to retrieve cover data as base64.
  • encoding_errors (str) – option to pass to str.encode()’s errors parameter before parsing MediaInfo’s XML output.
  • parse_speed (float) – passed to the library as ParseSpeed, this option takes values between 0 and 1. A higher value will yield more precise results in some cases but will also increase parsing time.
  • full (bool) – display additional tags, including computer-readable values for sizes and durations, corresponds to the CLI’s --Full/-f parameter.
  • legacy_stream_display (bool) – display additional information about streams.
  • mediainfo_options (dict) – additional options that will be passed to the MediaInfo_Option function, for example: {"Language": "raw"}. Do not use this parameter when running the method simultaneously from multiple threads, it will trigger a reset of all options which will cause inconsistencies or failures.
  • output (str) –

    custom output format for MediaInfo, corresponds to the CLI’s --Output parameter. Setting this causes the method to return a str instead of a MediaInfo object.

    Useful values include:
    • the empty str "" (corresponds to the default text output, obtained when running mediainfo with no additional parameters)
    • "XML"
    • "JSON"
    • %-delimited templates (see mediainfo --Info-Parameters)
Return type:

str if output is set.

Return type:

MediaInfo otherwise.

  • FileNotFoundError – if passed a non-existent file.
  • ValueError – if passed a file-like object opened in text mode.
  • OSError – if the library file could not be loaded.
  • RuntimeError – if parsing fails, this should not happen unless libmediainfo itself fails.
>>> pymediainfo.MediaInfo.parse("tests/data/sample.mkv")
    <pymediainfo.MediaInfo object at 0x7fa83a3db240>
>>> import json
>>> mi = pymediainfo.MediaInfo.parse("tests/data/sample.mkv",
...     output="JSON")
>>> json.loads(mi)["media"]["track"][0]
    {'@type': 'General', 'TextCount': '1', 'FileExtension': 'mkv',
        'FileSize': '5904',  … }
Returns:All Tracks of type Text.
Return type:list of Tracks
to_data() → Dict[str, Any]

Returns a dict representation of the object’s Tracks.

Return type:dict
to_json() → str

Returns a JSON representation of the object’s Tracks.

Return type:str
Returns:All Tracks of type Video.
Return type:list of Tracks
class pymediainfo.Track(xml_dom_fragment: xml.etree.ElementTree.Element)

An object associated with a media file track.

Each Track attribute corresponds to attributes parsed from MediaInfo’s output. All attributes are lower case. Attributes that are present several times such as Duration yield a second attribute starting with other_ which is a list of all alternative attribute values.

When a non-existing attribute is accessed, None is returned.


>>> t = mi.tracks[0]
>>> t
<Track track_id='None', track_type='General'>
>>> t.duration
>>> t.other_duration
['3 s 0 ms', '3 s 0 ms', '3 s 0 ms',
    '00:00:03.000', '00:00:03.000']
>>> type(t.non_existing)

All available attributes can be obtained by calling to_data().

to_data() → Dict[str, Any]

Returns a dict representation of the track attributes.


>>> sorted(track.to_data().keys())[:3]
['codec', 'codec_extensions_usually_used', 'codec_url']
>>> t.to_data()["file_size"]
Return type:dict