torchdata 0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/CHANGELOG.md +3 -0
- data/LICENSE.txt +30 -0
- data/README.md +55 -0
- data/lib/torchdata/data_pipes/iter/util/csv_parser.rb +45 -0
- data/lib/torchdata/data_pipes/iter/util/random_splitter.rb +114 -0
- data/lib/torchdata/version.rb +3 -0
- data/lib/torchdata.rb +23 -0
- metadata +63 -0
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA256:
|
3
|
+
metadata.gz: 3417d94393b7c9636a0c3e0935ecc62ddb2306ff51dcaf3a91f7118859ce8154
|
4
|
+
data.tar.gz: 3ad87bb333a0103db4ec5ff861f088d33471029dc5ae8739e6af27f325efe907
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: e3be548b4ee8575da3b9ed68b748bbf9b6f472472c903529ffaf65703582a9ee0e2c4f827e6383b318e12dfe4200b2533a118884beff406d6b98cdfaff886f19
|
7
|
+
data.tar.gz: '0797aabe0fa0975545f2650960012badccf7458208c2e0cbbcad88869000cafcce4465e389536ec3554f9d22f0322dea0ae674453e8887787fc3758ba14fa142'
|
data/CHANGELOG.md
ADDED
data/LICENSE.txt
ADDED
@@ -0,0 +1,30 @@
|
|
1
|
+
BSD 3-Clause License
|
2
|
+
|
3
|
+
Copyright (c) 2021-present, Facebook, Inc.
|
4
|
+
Copyright (c) 2023, Andrew Kane.
|
5
|
+
All rights reserved.
|
6
|
+
|
7
|
+
Redistribution and use in source and binary forms, with or without
|
8
|
+
modification, are permitted provided that the following conditions are met:
|
9
|
+
|
10
|
+
* Redistributions of source code must retain the above copyright notice, this
|
11
|
+
list of conditions and the following disclaimer.
|
12
|
+
|
13
|
+
* Redistributions in binary form must reproduce the above copyright notice,
|
14
|
+
this list of conditions and the following disclaimer in the documentation
|
15
|
+
and/or other materials provided with the distribution.
|
16
|
+
|
17
|
+
* Neither the name of the copyright holder nor the names of its
|
18
|
+
contributors may be used to endorse or promote products derived from
|
19
|
+
this software without specific prior written permission.
|
20
|
+
|
21
|
+
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
22
|
+
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
23
|
+
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
|
24
|
+
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
|
25
|
+
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
26
|
+
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
|
27
|
+
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
|
28
|
+
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
|
29
|
+
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
30
|
+
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
data/README.md
ADDED
@@ -0,0 +1,55 @@
|
|
1
|
+
# TorchData Ruby
|
2
|
+
|
3
|
+
Composable data loading for Ruby
|
4
|
+
|
5
|
+
[![Build Status](https://github.com/ankane/torchdata-ruby/workflows/build/badge.svg?branch=master)](https://github.com/ankane/torchdata-ruby/actions)
|
6
|
+
|
7
|
+
## Installation
|
8
|
+
|
9
|
+
Add this line to your application’s Gemfile:
|
10
|
+
|
11
|
+
```ruby
|
12
|
+
gem "torchdata"
|
13
|
+
```
|
14
|
+
|
15
|
+
## Getting Started
|
16
|
+
|
17
|
+
This library follows the [Python API](https://pytorch.org/data/). Many methods and options are missing at the moment. PRs welcome!
|
18
|
+
|
19
|
+
```ruby
|
20
|
+
folder = "path/to/csv/folder"
|
21
|
+
datapipe = TorchData::DataPipes::Iter::FileLister.new([folder]).filter { |filename| filename.end_with?(".csv") }
|
22
|
+
datapipe = TorchData::DataPipes::Iter::FileOpener.new(datapipe, mode: "rt")
|
23
|
+
datapipe = datapipe.parse_csv(delimiter: ",")
|
24
|
+
train, valid = datapipe.random_split(total_length: 10000, weights: {train: 0.5, valid: 0.5}, seed: 0)
|
25
|
+
|
26
|
+
train.each do |x|
|
27
|
+
# code
|
28
|
+
end
|
29
|
+
|
30
|
+
valid.each do |y|
|
31
|
+
# code
|
32
|
+
end
|
33
|
+
```
|
34
|
+
|
35
|
+
## History
|
36
|
+
|
37
|
+
View the [changelog](CHANGELOG.md)
|
38
|
+
|
39
|
+
## Contributing
|
40
|
+
|
41
|
+
Everyone is encouraged to help improve this project. Here are a few ways you can help:
|
42
|
+
|
43
|
+
- [Report bugs](https://github.com/ankane/torchdata-ruby/issues)
|
44
|
+
- Fix bugs and [submit pull requests](https://github.com/ankane/torchdata-ruby/pulls)
|
45
|
+
- Write, clarify, or fix documentation
|
46
|
+
- Suggest or add new features
|
47
|
+
|
48
|
+
To get started with development:
|
49
|
+
|
50
|
+
```sh
|
51
|
+
git clone https://github.com/ankane/torchdata-ruby.git
|
52
|
+
cd torchdata-ruby
|
53
|
+
bundle install
|
54
|
+
bundle exec rake test
|
55
|
+
```
|
@@ -0,0 +1,45 @@
|
|
1
|
+
module TorchData
|
2
|
+
module DataPipes
|
3
|
+
module Iter
|
4
|
+
module Util
|
5
|
+
class CSVParser < IterDataPipe
|
6
|
+
functional_datapipe :parse_csv
|
7
|
+
|
8
|
+
def initialize(source_datapipe, delimiter: ",")
|
9
|
+
@source_datapipe = source_datapipe
|
10
|
+
@helper = PlainTextReaderHelper.new
|
11
|
+
@fmtparams = {col_sep: delimiter}
|
12
|
+
end
|
13
|
+
|
14
|
+
def each(&block)
|
15
|
+
@source_datapipe.each do |path, file|
|
16
|
+
stream = @helper.skip_lines(file)
|
17
|
+
stream = @helper.decode(stream)
|
18
|
+
stream = CSV.parse(stream, **@fmtparams)
|
19
|
+
stream = @helper.as_tuple(stream)
|
20
|
+
@helper.return_path(stream, path: path).each(&block)
|
21
|
+
end
|
22
|
+
end
|
23
|
+
end
|
24
|
+
|
25
|
+
class PlainTextReaderHelper
|
26
|
+
def skip_lines(file)
|
27
|
+
file
|
28
|
+
end
|
29
|
+
|
30
|
+
def decode(stream)
|
31
|
+
stream
|
32
|
+
end
|
33
|
+
|
34
|
+
def return_path(stream, path: nil)
|
35
|
+
stream
|
36
|
+
end
|
37
|
+
|
38
|
+
def as_tuple(stream)
|
39
|
+
stream
|
40
|
+
end
|
41
|
+
end
|
42
|
+
end
|
43
|
+
end
|
44
|
+
end
|
45
|
+
end
|
@@ -0,0 +1,114 @@
|
|
1
|
+
module TorchData
|
2
|
+
module DataPipes
|
3
|
+
module Iter
|
4
|
+
module Util
|
5
|
+
class RandomSplitter < IterDataPipe
|
6
|
+
functional_datapipe :random_split
|
7
|
+
|
8
|
+
def self.new(source_datapipe, weights:, seed:, total_length: nil, target: nil)
|
9
|
+
if total_length.nil?
|
10
|
+
begin
|
11
|
+
total_length = source_datapipe.length
|
12
|
+
rescue NoMethodError
|
13
|
+
raise TypeError, "RandomSplitter needs `total_length`, but it is unable to infer it from the `source_datapipe`: #{source_datapipe}."
|
14
|
+
end
|
15
|
+
end
|
16
|
+
|
17
|
+
container = InternalRandomSplitterIterDataPipe.new(source_datapipe, total_length, weights, seed)
|
18
|
+
|
19
|
+
if target.nil?
|
20
|
+
weights.map { |k, _| SplitterIterator.new(container, k) }
|
21
|
+
else
|
22
|
+
raise "todo"
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
|
27
|
+
class InternalRandomSplitterIterDataPipe < IterDataPipe
|
28
|
+
attr_reader :source_datapipe
|
29
|
+
|
30
|
+
def initialize(source_datapipe, total_length, weights, seed)
|
31
|
+
@source_datapipe = source_datapipe
|
32
|
+
@total_length = total_length
|
33
|
+
@remaining_length = @total_length
|
34
|
+
@seed = seed
|
35
|
+
@keys = weights.keys
|
36
|
+
@key_to_index = @keys.map.with_index.to_h
|
37
|
+
@norm_weights = self.class.normalize_weights(@keys.map { |k| weights[k] }, total_length)
|
38
|
+
@weights = @norm_weights.dup
|
39
|
+
@rng = Random.new(@seed)
|
40
|
+
@lengths = []
|
41
|
+
end
|
42
|
+
|
43
|
+
def draw
|
44
|
+
selected_key = choices(@rng, @keys, @weights)
|
45
|
+
index = @key_to_index[selected_key]
|
46
|
+
@weights[index] -= 1
|
47
|
+
@remaining_length -= 1
|
48
|
+
if @weights[index] < 0
|
49
|
+
@weights[index] = 0
|
50
|
+
@weights = self.class.normalize_weights(@weights, @remaining_length)
|
51
|
+
end
|
52
|
+
selected_key
|
53
|
+
end
|
54
|
+
|
55
|
+
def self.normalize_weights(weights, total_length)
|
56
|
+
total_weight = weights.sum
|
57
|
+
weights.map { |w| w.to_f * total_length / total_weight }
|
58
|
+
end
|
59
|
+
|
60
|
+
def reset
|
61
|
+
@rng = Random.new(@seed)
|
62
|
+
@weights = @norm_weights.dup
|
63
|
+
@remaining_length = @total_length
|
64
|
+
end
|
65
|
+
|
66
|
+
def override_seed(seed)
|
67
|
+
@seed = seed
|
68
|
+
self
|
69
|
+
end
|
70
|
+
|
71
|
+
def get_length(target)
|
72
|
+
raise "todo"
|
73
|
+
end
|
74
|
+
|
75
|
+
private
|
76
|
+
|
77
|
+
def choices(rng, keys, weights)
|
78
|
+
total = weights.sum
|
79
|
+
x = rng.rand * total
|
80
|
+
weights.each_with_index do |w, i|
|
81
|
+
return keys[i] if x < w
|
82
|
+
x -= w
|
83
|
+
end
|
84
|
+
keys[-1]
|
85
|
+
end
|
86
|
+
end
|
87
|
+
|
88
|
+
class SplitterIterator < IterDataPipe
|
89
|
+
def initialize(main_datapipe, target)
|
90
|
+
@main_datapipe = main_datapipe
|
91
|
+
@target = target
|
92
|
+
end
|
93
|
+
|
94
|
+
def each
|
95
|
+
@main_datapipe.reset
|
96
|
+
@main_datapipe.source_datapipe.each do |sample|
|
97
|
+
if @main_datapipe.draw == @target
|
98
|
+
yield sample
|
99
|
+
end
|
100
|
+
end
|
101
|
+
end
|
102
|
+
|
103
|
+
def override_seed(seed)
|
104
|
+
@main_datapipe.override_seed(seed)
|
105
|
+
end
|
106
|
+
|
107
|
+
def length
|
108
|
+
@main_datapipe.get_length(@target)
|
109
|
+
end
|
110
|
+
end
|
111
|
+
end
|
112
|
+
end
|
113
|
+
end
|
114
|
+
end
|
data/lib/torchdata.rb
ADDED
@@ -0,0 +1,23 @@
|
|
1
|
+
# dependencies
|
2
|
+
require "torch"
|
3
|
+
|
4
|
+
# stdlib
|
5
|
+
require "csv"
|
6
|
+
|
7
|
+
# modules
|
8
|
+
require_relative "torchdata/version"
|
9
|
+
|
10
|
+
module TorchData
|
11
|
+
class Error < StandardError; end
|
12
|
+
|
13
|
+
module DataPipes
|
14
|
+
module Iter
|
15
|
+
IterDataPipe = Torch::Utils::Data::DataPipes::IterDataPipe
|
16
|
+
FileLister = Torch::Utils::Data::DataPipes::Iter::FileLister
|
17
|
+
FileOpener = Torch::Utils::Data::DataPipes::Iter::FileOpener
|
18
|
+
end
|
19
|
+
end
|
20
|
+
end
|
21
|
+
|
22
|
+
require_relative "torchdata/data_pipes/iter/util/csv_parser"
|
23
|
+
require_relative "torchdata/data_pipes/iter/util/random_splitter"
|
metadata
ADDED
@@ -0,0 +1,63 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: torchdata
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.0.1
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- Andrew Kane
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
date: 2023-01-31 00:00:00.000000000 Z
|
12
|
+
dependencies:
|
13
|
+
- !ruby/object:Gem::Dependency
|
14
|
+
name: torch-rb
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
16
|
+
requirements:
|
17
|
+
- - ">="
|
18
|
+
- !ruby/object:Gem::Version
|
19
|
+
version: 0.12.2
|
20
|
+
type: :runtime
|
21
|
+
prerelease: false
|
22
|
+
version_requirements: !ruby/object:Gem::Requirement
|
23
|
+
requirements:
|
24
|
+
- - ">="
|
25
|
+
- !ruby/object:Gem::Version
|
26
|
+
version: 0.12.2
|
27
|
+
description:
|
28
|
+
email: andrew@ankane.org
|
29
|
+
executables: []
|
30
|
+
extensions: []
|
31
|
+
extra_rdoc_files: []
|
32
|
+
files:
|
33
|
+
- CHANGELOG.md
|
34
|
+
- LICENSE.txt
|
35
|
+
- README.md
|
36
|
+
- lib/torchdata.rb
|
37
|
+
- lib/torchdata/data_pipes/iter/util/csv_parser.rb
|
38
|
+
- lib/torchdata/data_pipes/iter/util/random_splitter.rb
|
39
|
+
- lib/torchdata/version.rb
|
40
|
+
homepage: https://github.com/ankane/torchdata-ruby
|
41
|
+
licenses:
|
42
|
+
- BSD-3-Clause
|
43
|
+
metadata: {}
|
44
|
+
post_install_message:
|
45
|
+
rdoc_options: []
|
46
|
+
require_paths:
|
47
|
+
- lib
|
48
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
49
|
+
requirements:
|
50
|
+
- - ">="
|
51
|
+
- !ruby/object:Gem::Version
|
52
|
+
version: '2.7'
|
53
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
54
|
+
requirements:
|
55
|
+
- - ">="
|
56
|
+
- !ruby/object:Gem::Version
|
57
|
+
version: '0'
|
58
|
+
requirements: []
|
59
|
+
rubygems_version: 3.4.1
|
60
|
+
signing_key:
|
61
|
+
specification_version: 4
|
62
|
+
summary: Composable data loading for Ruby
|
63
|
+
test_files: []
|