torchaudio 0.2.1 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9ed4c14921f1eee18f5e08ddabfae51e09a9b5a7ef408f1dd67fdf7bfe9622fe
-  data.tar.gz: 1e37d5b9abed9cab7bf56a8c30a769bc8ff8f8a3e15e78bbb772847c444571b2
+  metadata.gz: 40ba5316b924f832befcf352c040528b48da458f53e0a5087530c06604bf7f48
+  data.tar.gz: 6235bf05e51dcb4ec1f63596085243e97ed02cde73a5504c5ced1b03a4679434
 SHA512:
-  metadata.gz: 9ca5436d7e4309dd9659fdce7ee893b122e9da96e9f7b15bf00de5dea32c635e2828a99939f04ef5bf0d9494ab89957829a65002dc3e855fa8a66f54abbbd181
-  data.tar.gz: d62b2a137c19d3b24facb11eda5c1b81be5841120b505877b8617bee2b9f183dbe4b4d42a95af27447346a3d48476d7faec48b57cd89b88c0ddc9709f1b5d51b
+  metadata.gz: c2ec8ae3f0a4fbfb242cf718624828bec8d8b1b741ff119836e7087ea98de43498377c9838eddcb73337db7a56c0d612232aaf7f95d65b7ea05173e93dfae7b8
+  data.tar.gz: 02c22f6eeb6bb037ff1d69d781861a60c29bd6f35e4ef76d1793cfb5b9f1696f6309208ec748f6736dd61239f820c65b1a02859c950cff3b2fd97d3fa96bd2f3

data/CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.3.1 (2023-01-29)
+
+- Added `format` option
+
+## 0.3.0 (2022-07-06)
+
+- Added `center`, `pad_mode`, and `onesided` options to `Spectogram` transform
+- Dropped support for Ruby < 2.7
+
 ## 0.2.1 (2021-07-16)
 
 - Added `create_dct` method

data/LICENSE.txt

@@ -1,7 +1,7 @@
 BSD 2-Clause License
 
 Copyright (c) 2017 Facebook Inc. (Soumith Chintala),
-Copyright (c) 2020-2021 Andrew Kane,
+Copyright (c) 2020-2022 Andrew Kane,
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

data/README.md

@@ -1,8 +1,8 @@
-# TorchAudio
+# TorchAudio Ruby
 
 :fire: An audio library for Torch.rb
 
-[![Build Status](https://github.com/ankane/torchaudio/workflows/build/badge.svg?branch=master)](https://github.com/ankane/torchaudio/actions)
+[![Build Status](https://github.com/ankane/torchaudio-ruby/workflows/build/badge.svg?branch=master)](https://github.com/ankane/torchaudio-ruby/actions)
 
 ## Installation
 
@@ -15,7 +15,7 @@ brew install sox
 Add this line to your application’s Gemfile:
 
 ```ruby
-gem 'torchaudio'
+gem "torchaudio"
 ```
 
 ## Getting Started
@@ -134,22 +134,22 @@ addons:
 
 ## History
 
-View the [changelog](https://github.com/ankane/torchaudio/blob/master/CHANGELOG.md)
+View the [changelog](https://github.com/ankane/torchaudio-ruby/blob/master/CHANGELOG.md)
 
 ## Contributing
 
 Everyone is encouraged to help improve this project. Here are a few ways you can help:
 
-- [Report bugs](https://github.com/ankane/torchaudio/issues)
-- Fix bugs and [submit pull requests](https://github.com/ankane/torchaudio/pulls)
+- [Report bugs](https://github.com/ankane/torchaudio-ruby/issues)
+- Fix bugs and [submit pull requests](https://github.com/ankane/torchaudio-ruby/pulls)
 - Write, clarify, or fix documentation
 - Suggest or add new features
 
 To get started with development:
 
 ```sh
-git clone https://github.com/ankane/torchaudio.git
-cd torchaudio
+git clone https://github.com/ankane/torchaudio-ruby.git
+cd torchaudio-ruby
 bundle install
 bundle exec rake compile
 bundle exec rake test

data/ext/torchaudio/extconf.rb

@@ -20,18 +20,9 @@ $CXXFLAGS += " -D_GLIBCXX_USE_CXX11_ABI=1"
 
 apple_clang = RbConfig::CONFIG["CC_VERSION_MESSAGE"] =~ /apple clang/i
 
-# check omp first
-if have_library("omp") || have_library("gomp")
-  $CXXFLAGS += " -Xclang" if apple_clang
-  $CXXFLAGS += " -fopenmp"
-end
-
 if apple_clang
-  # silence ruby/intern.h warning
-  $CXXFLAGS += " -Wno-deprecated-register"
-
   # silence torch warnings
-  $CXXFLAGS += " -Wno-shorten-64-to-32 -Wno-missing-noreturn"
+  $CXXFLAGS += " -Wno-deprecated-declarations"
 else
   # silence rice warnings
   $CXXFLAGS += " -Wno-noexcept-type"

data/lib/torchaudio/datasets/utils.rb

@@ -15,7 +15,9 @@ module TorchAudio
         end
 
         # follows redirects
-        def download_url_to_file(url, dst, hash_value, hash_type)
+        def download_url_to_file(url, dst, hash_value, hash_type, redirects = 0)
+          raise "Too many redirects" if redirects > 10
+
           uri = URI(url)
           tmp = nil
           location = nil
@@ -41,7 +43,7 @@ module TorchAudio
           end
 
           if location
-            download_url_to_file(location, dst)
+            download_url_to_file(location, dst, hash_value, hash_type, redirects + 1)
           else
             # check hash
             # TODO use hash_type

data/lib/torchaudio/functional.rb

@@ -1,7 +1,7 @@
 module TorchAudio
   module Functional
     class << self
-      def spectrogram(waveform, pad, window, n_fft, hop_length, win_length, power, normalized)
+      def spectrogram(waveform, pad, window, n_fft, hop_length, win_length, power, normalized, center: true, pad_mode: "reflect", onesided: true)
         if pad > 0
           # TODO add "with torch.no_grad():" back when JIT supports it
           waveform = Torch::NN::Functional.pad(waveform, [pad, pad], "constant")
@@ -19,22 +19,25 @@ module TorchAudio
             hop_length: hop_length,
             win_length: win_length,
             window: window,
-            center: true,
-            pad_mode: "reflect",
+            center: center,
+            pad_mode: pad_mode,
             normalized: false,
-            onesided: true
+            onesided: onesided,
+            return_complex: true
           )
 
         # unpack batch
-        spec_f = spec_f.reshape(shape[0..-2] + spec_f.shape[-3..-1])
+        spec_f = spec_f.reshape(shape[0..-2] + spec_f.shape[-2..-1])
 
         if normalized
-          spec_f.div!(window.pow(2.0).sum.sqrt)
+          spec_f /= window.pow(2.0).sum.sqrt
         end
-        if power
-          spec_f = complex_norm(spec_f, power: power)
+        if !power.nil?
+          if power == 1
+            return spec_f.abs
+          end
+          return spec_f.abs.pow(power)
         end
-
         spec_f
       end
 

data/lib/torchaudio/transforms/spectrogram.rb

@@ -3,7 +3,8 @@ module TorchAudio
     class Spectrogram < Torch::NN::Module
       def initialize(
         n_fft: 400, win_length: nil, hop_length: nil, pad: 0,
-        window_fn: Torch.method(:hann_window), power: 2.0, normalized: false, wkwargs: nil
+        window_fn: Torch.method(:hann_window), power: 2.0, normalized: false, wkwargs: nil,
+        center: true, pad_mode: "reflect", onesided: true
       )
 
         super()
@@ -17,10 +18,16 @@ module TorchAudio
         @pad = pad
         @power = power
         @normalized = normalized
+        @center = center
+        @pad_mode = pad_mode
+        @onesided = onesided
       end
 
       def forward(waveform)
-        F.spectrogram(waveform, @pad, @window, @n_fft, @hop_length, @win_length, @power, @normalized)
+        F.spectrogram(
+          waveform, @pad, @window, @n_fft, @hop_length, @win_length, @power, @normalized,
+          center: @center, pad_mode: @pad_mode, onesided: @onesided
+        )
       end
     end
   end

data/lib/torchaudio/version.rb

@@ -1,3 +1,3 @@
 module TorchAudio
-  VERSION = "0.2.1"
+  VERSION = "0.3.1"
 end

data/lib/torchaudio.rb

@@ -31,9 +31,10 @@ module TorchAudio
   class Error < StandardError; end
 
   class << self
+    # TODO remove filetype in 0.4.0
     def load(
       filepath, out: nil, normalization: true, channels_first: true, num_frames: 0,
-      offset: 0, signalinfo: nil, encodinginfo: nil, filetype: nil
+      offset: 0, signalinfo: nil, encodinginfo: nil, filetype: nil, format: nil
     )
 
       filepath = filepath.to_s
@@ -59,7 +60,7 @@ module TorchAudio
 
       # same logic as C++
       # could also make read_audio_file work with nil
-      filetype ||= File.extname(filepath)[1..-1]
+      format ||= filetype || File.extname(filepath)[1..-1]
 
       sample_rate =
         Ext.read_audio_file(
@@ -70,7 +71,7 @@ module TorchAudio
           offset,
           signalinfo,
           encodinginfo,
-          filetype
+          format
         )
 
       # normalize if needed
@@ -95,7 +96,7 @@ module TorchAudio
     end
 
     def save_encinfo(filepath, src, channels_first: true, signalinfo: nil, encodinginfo: nil, filetype: nil)
-      ch_idx, len_idx = channels_first ? [0, 1] : [1, 0]
+      ch_idx, _len_idx = channels_first ? [0, 1] : [1, 0]
 
       # check if save directory exists
       abs_dirpath = File.dirname(File.expand_path(filepath))

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: torchaudio
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.3.1
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-07-16 00:00:00.000000000 Z
+date: 2023-01-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: torch-rb
@@ -16,14 +16,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.7.0
+        version: 0.11.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.7.0
+        version: 0.11.1
 - !ruby/object:Gem::Dependency
   name: rice
   requirement: !ruby/object:Gem::Requirement
@@ -74,7 +74,7 @@ files:
 - lib/torchaudio/transforms/spectrogram.rb
 - lib/torchaudio/transforms/vol.rb
 - lib/torchaudio/version.rb
-homepage: https://github.com/ankane/torchaudio
+homepage: https://github.com/ankane/torchaudio-ruby
 licenses:
 - BSD-2-Clause
 metadata: {}
@@ -86,14 +86,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.6'
+      version: '2.7'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.22
+rubygems_version: 3.4.1
 signing_key:
 specification_version: 4
 summary: Data manipulation and transformation for audio signal processing