pydub-stubs
Pydub version: 0.25.1
pydub-stubs
provides type information for Pydub.
Only the public interface is guaranteed to be typed, however there
are type definitions for some private and undocumented functions.
pip install pydub-stubs
Anticipated Questions
Q: What is _Metadata
and _PartialMetadata
?
These are legacy types that have been replaced by the channels
,
frame_rate
, and sample_rate
keyword arguments.
Changelog
Version 0.25.1.5
- Fixed incorrect
WhiteNoise
generator name
Previously, the generator was incorrectly named Whitenoise
.
(#7 - thanks, @Kakadus!)
Version 0.25.1.4
- Added "effect" functions as methods on
AudioSegment
The methods have always been accessible on AudioSegment, but since they are
not defined on the class, they were not previously defined in the type stubs.
Note that if pydub.scipy_effects
is imported, the effect methods may have
slightly different definitions.
Version 0.25.1.3
- Added
typing.BinaryIO
to _PathLike
Pydub supports using an IO object in place of a file for both creating and
exporting AudioSegments. (Fixes #5)
Version 0.25.1.2
- Added
pathlib.Path
to _PathLike
Allows you to use Path
objects in place of strings. (thanks,
@qthequartermasterman!)
Version 0.25.1.1
-
Removed literal type fallback overloads
The fallbacks removed all the advantages of using literals at all.
-
Modernized the type stubs
Uses new union syntax, and more. (#2
and #3 - thanks, @Viicos!)
-
Add WavSubChunk
, WavData
, and undocumented functions
These previously existed and were available but were untyped. (thanks again, @Viicos!)
Previous versions
Version 0.25.1.0
-
Added v0.25.0 features
This includes pydub.scipy_effects.eq
and new classmethod parameters.
-
Signatures now use literals where possible
Overloaded implementations exist as a fallback.
-
Added missing modules
pydub.silence
and pydub.utils
Version 0.24.1.9
- Add undocumented parameter of
AudioSegment.from_file
read_ahead_limit
is absent from the documentation but is a supported
keyword argument.
Version 0.24.1.8
- Export other modules
Adds exports for effects, exceptions, generators, playback, and scipy_effects
Version 0.24.1.7
-
Added AudioSegment._spawn
(again)
This was accidentally removed in an earlier version.
-
Improved pydub.effects.invert_phase
This is technically less accurate as (0, 0)
is equivalent to (0, 1)
.
Version 0.24.1.6
- Removed testing symbols from
pydub.audio_segment
Version 0.24.1.5
- Fixed
AudioSegment.export
First param is named out_f
and isn't required.
Version 0.24.1.4
Version 0.24.1.3
- Fixed overloads of
AudioSegment.fade
Exactly two of start
, end
, and duration
must be given.
Version 0.24.1.2
-
Improved AudioSegment.fade
Changed to use overloads to prevent invalid method calls.
-
Improved AudioSegment.from_mono_audiosegments
Use a positional-only parameter to ensure there's at least 1 argument.
Version 0.24.1.1
-
Fixed AudioSegment.__init__
Use overloads to model correct parameters.
-
Fixed AudioSegment._spawn
Parameter overrides
accepts a partial dictionary.
-
Fixed pydub.scipy_effects.high_pass_filter
Parameter order
should be int
, not float
.
Version 0.24.1.0
Released