Name | torchchronos JSON |
Version |
0.0.3.post2
JSON |
| download |
home_page | https://github.com/mauricekraus/torchchronos |
Summary | PyTorch and Lightning compatible library that provides easy and flexible access to various time-series datasets for classification and regression tasks |
upload_time | 2023-08-25 15:26:33 |
maintainer | |
docs_url | None |
author | Maurice Kraus |
requires_python | >=3.10,<3.12 |
license | MIT |
keywords |
|
VCS |
|
bugtrack_url |
|
requirements |
No requirements were recorded.
|
Travis-CI |
No Travis.
|
coveralls test coverage |
No coveralls.
|
# torchchronos
[![PyPI version](https://img.shields.io/pypi/v/torchchronos.svg?color=blue)](https://pypi.org/project/torchchronos)
[![license](https://img.shields.io/pypi/l/torchchronos.svg?color=blue)](https://github.com/mauricekraus/torchchronos/blob/main/LICENSE)
[![python version](https://img.shields.io/badge/python-3.10+-blue)](https://devguide.python.org/versions/)
[![test](https://github.com/mauricekraus/torchchronos/actions/workflows/main.yml/badge.svg)](https://github.com/mauricekraus/torchchronos/actions/workflows/main.yml)
[![code style](https://img.shields.io/badge/code%20style-black-black)](https://github.com/psf/black)
*torchchronos* is an experimental [PyTorch](https://pytorch.org/) and [Lightning](https://lightning.ai/pytorch-lightning/) compatible library that provides easy and flexible access to various time-series datasets for classification and regression tasks. It also provides a simple and extensible transform API to preprocess data.
It is inspired by the much more complicated [torchtime](https://github.com/philipdarke/torchtime).
## Installation
You can install torchchronos via pip:
`pip install torchchronos`
## Usage
### Datasets
torchchronos currently provides access to several popular time-series datasets, including:
- [UCR/UEA Time Series Classification Repository](https://www.timeseriesclassification.com/): `torchchronos.datasets.UCRUEADataset`
- Time series as preprocessed in the [TFC paper](https://github.com/mims-harvard/TFC-pretraining): `torchchronos.datasets.TFCPretrainDataset` (datasets `Gesture` and `EMG`)
To use a dataset, you can simply import the corresponding dataset class and create an instance:
```python
from torchchronos.datasets import UCRUEADataset
from torchchronos.transforms import PadFront
from torchchronos.download import download_uea_ucr
download_uea_ucr("ECG5000",Path(".cache/data"))
dataset = UCRUEADataset('ECG5000', path=Path(".cache") / "data", transforms=PadFront(10))
```
### Data Modules
torchchronos also provides [Lightning compatible `DataModules`](https://lightning.ai/docs/pytorch/stable/data/datamodule.html) to make it easy to load and preprocess data. They support common use cases like (multi-)GPU training and train/test/val-splitting out of the box. For example:
```python
from torchchronos.lightning import UCRUEADataModule
from torchchronos.transforms import PadFront, PadBack
module = UCRUEAModule('ECG5000', split_ratio= (0.75, 0.15), batch_size= 32,
transforms=Compose([PadFront(10), PadBack(10)]))
```
Analogous the the datasets above, these dataloaders are supported as of now, wrapping the respective datasets:
- `torchchronos.lightning.UCRUEADataModule`
- `torchchronos.lightning.TFCPretrainDataModule`
### Transforms
torchchronos provides a flexible transform API to preprocess time-series data. For example, to normalize a dataset, you can define a custom `Transform` like this:
```python
from torchchronos.transforms import Transform
class Normalize(Transform):
def __init__(self, mean=None, std=None):
self.mean = mean
self.std = std
def fit(self, data) -> Self:
self.mean = data.mean()
self.std = data.std()
return self
def __call__(self, data):
return (data - self.mean) / self.std
```
## Known issues
- The dataset [SpokenArabicDigits](https://www.timeseriesclassification.com/description.php?Dataset=SpokenArabicDigits) does not seem to work due to a missmatch of TRAIN and TEST size
- The dataset [UrbanSound](https://www.timeseriesclassification.com/description.php?Dataset=UrbanSound) does not seem to work due to missing ts files
## Roadmap
The following features are planned for future releases of torchchronos:
- Support for additional time-series datasets, including:
- Energy consumption dataset
- Traffic dataset
- PhysioNet Challenge 2012 (in-hospital mortality)
- PhysioNet Challenge 2019 (sepsis prediction) datasets
- Additional transform classes, including:
- Resampling
- Missing value imputation
If you have any feature requests or suggestions, please open an issue on our GitHub page.
Raw data
{
"_id": null,
"home_page": "https://github.com/mauricekraus/torchchronos",
"name": "torchchronos",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.10,<3.12",
"maintainer_email": "",
"keywords": "",
"author": "Maurice Kraus",
"author_email": "dev@mkraus.io",
"download_url": "https://files.pythonhosted.org/packages/35/45/88ade79ab2acca40e01cb127a949978e37c829e8a660e1beb5ee87254186/torchchronos-0.0.3.post2.tar.gz",
"platform": null,
"description": "# torchchronos\n\n[![PyPI version](https://img.shields.io/pypi/v/torchchronos.svg?color=blue)](https://pypi.org/project/torchchronos)\n[![license](https://img.shields.io/pypi/l/torchchronos.svg?color=blue)](https://github.com/mauricekraus/torchchronos/blob/main/LICENSE)\n[![python version](https://img.shields.io/badge/python-3.10+-blue)](https://devguide.python.org/versions/)\n\n[![test](https://github.com/mauricekraus/torchchronos/actions/workflows/main.yml/badge.svg)](https://github.com/mauricekraus/torchchronos/actions/workflows/main.yml)\n[![code style](https://img.shields.io/badge/code%20style-black-black)](https://github.com/psf/black)\n\n*torchchronos* is an experimental [PyTorch](https://pytorch.org/) and [Lightning](https://lightning.ai/pytorch-lightning/) compatible library that provides easy and flexible access to various time-series datasets for classification and regression tasks. It also provides a simple and extensible transform API to preprocess data.\nIt is inspired by the much more complicated [torchtime](https://github.com/philipdarke/torchtime).\n\n## Installation\nYou can install torchchronos via pip:\n\n`pip install torchchronos`\n\n## Usage\n### Datasets\ntorchchronos currently provides access to several popular time-series datasets, including:\n\n- [UCR/UEA Time Series Classification Repository](https://www.timeseriesclassification.com/): `torchchronos.datasets.UCRUEADataset`\n- Time series as preprocessed in the [TFC paper](https://github.com/mims-harvard/TFC-pretraining): `torchchronos.datasets.TFCPretrainDataset` (datasets `Gesture` and `EMG`)\n\nTo use a dataset, you can simply import the corresponding dataset class and create an instance:\n\n```python\nfrom torchchronos.datasets import UCRUEADataset\nfrom torchchronos.transforms import PadFront\nfrom torchchronos.download import download_uea_ucr\n\ndownload_uea_ucr(\"ECG5000\",Path(\".cache/data\"))\ndataset = UCRUEADataset('ECG5000', path=Path(\".cache\") / \"data\", transforms=PadFront(10))\n```\n\n### Data Modules\ntorchchronos also provides [Lightning compatible `DataModules`](https://lightning.ai/docs/pytorch/stable/data/datamodule.html) to make it easy to load and preprocess data. They support common use cases like (multi-)GPU training and train/test/val-splitting out of the box. For example:\n\n```python\nfrom torchchronos.lightning import UCRUEADataModule\nfrom torchchronos.transforms import PadFront, PadBack\n\nmodule = UCRUEAModule('ECG5000', split_ratio= (0.75, 0.15), batch_size= 32,\n transforms=Compose([PadFront(10), PadBack(10)]))\n```\n\nAnalogous the the datasets above, these dataloaders are supported as of now, wrapping the respective datasets:\n- `torchchronos.lightning.UCRUEADataModule`\n- `torchchronos.lightning.TFCPretrainDataModule`\n\n### Transforms\ntorchchronos provides a flexible transform API to preprocess time-series data. For example, to normalize a dataset, you can define a custom `Transform` like this:\n\n```python\nfrom torchchronos.transforms import Transform\n\nclass Normalize(Transform):\n def __init__(self, mean=None, std=None):\n self.mean = mean\n self.std = std\n\n def fit(self, data) -> Self:\n self.mean = data.mean()\n self.std = data.std()\n return self\n\n def __call__(self, data):\n return (data - self.mean) / self.std\n```\n\n## Known issues\n- The dataset [SpokenArabicDigits](https://www.timeseriesclassification.com/description.php?Dataset=SpokenArabicDigits) does not seem to work due to a missmatch of TRAIN and TEST size\n- The dataset [UrbanSound](https://www.timeseriesclassification.com/description.php?Dataset=UrbanSound) does not seem to work due to missing ts files\n\n\n## Roadmap\nThe following features are planned for future releases of torchchronos:\n\n- Support for additional time-series datasets, including:\n - Energy consumption dataset\n - Traffic dataset\n - PhysioNet Challenge 2012 (in-hospital mortality)\n - PhysioNet Challenge 2019 (sepsis prediction) datasets\n- Additional transform classes, including:\n - Resampling\n - Missing value imputation\n\nIf you have any feature requests or suggestions, please open an issue on our GitHub page.\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "PyTorch and Lightning compatible library that provides easy and flexible access to various time-series datasets for classification and regression tasks",
"version": "0.0.3.post2",
"project_urls": {
"Homepage": "https://github.com/mauricekraus/torchchronos"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "81f4212be10e892dee365316bdd45d1c322031296cf3f63a9027bb61ee1b12a7",
"md5": "b8266f3e9f7217b7be6443748e87065b",
"sha256": "88e4718ff6e1d29f345c072ff98e2a792f686af43a859bccc596675e9c1d56b9"
},
"downloads": -1,
"filename": "torchchronos-0.0.3.post2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "b8266f3e9f7217b7be6443748e87065b",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10,<3.12",
"size": 13489,
"upload_time": "2023-08-25T15:26:32",
"upload_time_iso_8601": "2023-08-25T15:26:32.375516Z",
"url": "https://files.pythonhosted.org/packages/81/f4/212be10e892dee365316bdd45d1c322031296cf3f63a9027bb61ee1b12a7/torchchronos-0.0.3.post2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "354588ade79ab2acca40e01cb127a949978e37c829e8a660e1beb5ee87254186",
"md5": "cb8bfd3a7eb45965794688e8eb965d9b",
"sha256": "471f287fb37eee8c8a6d0b0fe91140d4e9255c3c1ed32a47103ffb2899094149"
},
"downloads": -1,
"filename": "torchchronos-0.0.3.post2.tar.gz",
"has_sig": false,
"md5_digest": "cb8bfd3a7eb45965794688e8eb965d9b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10,<3.12",
"size": 10781,
"upload_time": "2023-08-25T15:26:33",
"upload_time_iso_8601": "2023-08-25T15:26:33.422192Z",
"url": "https://files.pythonhosted.org/packages/35/45/88ade79ab2acca40e01cb127a949978e37c829e8a660e1beb5ee87254186/torchchronos-0.0.3.post2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-08-25 15:26:33",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "mauricekraus",
"github_project": "torchchronos",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "torchchronos"
}