# ezDSP: Easy Digital Signal Processing
ezDSP is a digital signal processing toolbox written in PyTorch (`./src/ezdsp/nn/`). However, ezDSP not only processes torch.tensors (CPU & GPU) but also handles numpy.ndarray and pd.DataFrame, enabling a consistent and intensive workflow.
## Installation
```bash
$ pip install ezdsp
$ python ./example.py # ./example_outputs/ will be generated.
```
## Samples
#### Normalization, Resampling, Noise Addition, Filtering, Hilbert Transformation
<div align="center">
<img src="./example_outputs/3_chirp/1_signals.png" height="400">
</div>
#### Wavelet Transformation
<div align="center">
<img src="./example_outputs/4_ripple/2_wavelet_orig.png" height="400">
</div>
#### Power Spectrum Density
<div align="center">
<img src="./example_outputs/3_chirp/3_orig.png" height="400">
<img src="./example_outputs/3_chirp/3_psd_bandstop_filted%20(20%20-%2050%20Hz).png" height="400">
</div>
#### Phase-Amplitude Coupling
<div align="center">
<img src="./example_outputs/modulation_index_calculation_with_ezDSP_and_Tensorpac.png" height="400">
</div>
## Quick Start
``` python
import ezdsp as ed
# Parameters
SRC_FS = 1024 # Source sampling frequency
TGT_FS = 512 # Target sampling frequency
FREQS_HZ = [10, 30, 100] # Frequencies in Hz
LOW_HZ = 20 # Low frequency for bandpass filter
HIGH_HZ = 50 # High frequency for bandpass filter
SIGMA = 10 # Sigma for Gaussian filter
SIG_TYPES = [
"uniform",
"gauss",
"periodic",
"chirp",
"ripple",
"meg",
"tensorpac",
] # Available signal types
# Demo Signal
xx, tt, fs = ed.demo_sig(
t_sec=T_SEC, fs=SRC_FS, freqs_hz=FREQS_HZ, sig_type="chirp"
)
# xx is either of torch.tensor (on cpu / cuda), numpy.ndarray, or pd.DataFrame.
# Normalization
xx_norm = ed.norm.z(xx)
xx_minmax = ed.norm.minmax(xx)
# Resampling
xx_resampled = ed.resample(xx, fs, TGT_FS)
# Noise addition
xx_gauss = ed.add_noise.gauss(xx)
xx_white = ed.add_noise.white(xx)
xx_pink = ed.add_noise.pink(xx)
xx_brown = ed.add_noise.brown(xx)
# Filtering
xx_filted_bandpass = ed.filt.bandpass(xx, fs, low_hz=LOW_HZ, high_hz=HIGH_HZ)
xx_filted_bandstop = ed.filt.bandstop(xx, fs, low_hz=LOW_HZ, high_hz=HIGH_HZ)
xx_filted_gauss = ed.filt.gauss(xx, sigma=SIGMA)
# Hilbert Transformation
phase, amplitude = ed.hilbert(xx) # or envelope
# Wavelet Transformation
wavelet_coef, wavelet_freqs = ed.wavelet(xx, fs)
# Power Spetrum Density
psd, psd_freqs = ed.psd(xx, fs)
# Phase-Amplitude Coupling
pac, freqs_pha, freqs_amp = ed.pac(x_3d, fs) # This function is computationally demanding. Please monitor the RAM/VRAM usage.
```
# Alias
[`mngs.dsp`](https://github.com/ywatanabe1989/mngs/src/mngs/dsp/) has the same functionalities.
# Contact
Yusuke Watanabe (ywata1989@gmail.com).
Raw data
{
"_id": null,
"home_page": "https://github.com/ywatanabe1989/ezdsp",
"name": "ezdsp",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": null,
"keywords": "numpytorch, pandas, LFP, EEG, MEG",
"author": "ywatanabe1989",
"author_email": "ywata1989@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/e8/7f/823afc937fdce6f233b7d58e3f9649f9617fa7de495d738e4657864e869b/ezdsp-1.3.0.tar.gz",
"platform": null,
"description": "\n\n\n# ezDSP: Easy Digital Signal Processing\n\nezDSP is a digital signal processing toolbox written in PyTorch (`./src/ezdsp/nn/`). However, ezDSP not only processes torch.tensors (CPU & GPU) but also handles numpy.ndarray and pd.DataFrame, enabling a consistent and intensive workflow.\n\n## Installation\n```bash\n$ pip install ezdsp\n$ python ./example.py # ./example_outputs/ will be generated.\n```\n\n## Samples\n#### Normalization, Resampling, Noise Addition, Filtering, Hilbert Transformation\n<div align=\"center\">\n <img src=\"./example_outputs/3_chirp/1_signals.png\" height=\"400\">\n</div>\n\n#### Wavelet Transformation\n<div align=\"center\">\n <img src=\"./example_outputs/4_ripple/2_wavelet_orig.png\" height=\"400\">\n</div>\n\n#### Power Spectrum Density\n<div align=\"center\">\n <img src=\"./example_outputs/3_chirp/3_orig.png\" height=\"400\">\n <img src=\"./example_outputs/3_chirp/3_psd_bandstop_filted%20(20%20-%2050%20Hz).png\" height=\"400\">\n</div>\n\n#### Phase-Amplitude Coupling\n<div align=\"center\">\n <img src=\"./example_outputs/modulation_index_calculation_with_ezDSP_and_Tensorpac.png\" height=\"400\">\n</div>\n\n## Quick Start\n``` python\nimport ezdsp as ed\n\n# Parameters\nSRC_FS = 1024 # Source sampling frequency\nTGT_FS = 512 # Target sampling frequency\nFREQS_HZ = [10, 30, 100] # Frequencies in Hz\nLOW_HZ = 20 # Low frequency for bandpass filter\nHIGH_HZ = 50 # High frequency for bandpass filter\nSIGMA = 10 # Sigma for Gaussian filter\nSIG_TYPES = [\n \"uniform\",\n \"gauss\",\n \"periodic\",\n \"chirp\",\n \"ripple\",\n \"meg\",\n \"tensorpac\",\n] # Available signal types\n\n\n# Demo Signal\nxx, tt, fs = ed.demo_sig(\n t_sec=T_SEC, fs=SRC_FS, freqs_hz=FREQS_HZ, sig_type=\"chirp\"\n)\n# xx is either of torch.tensor (on cpu / cuda), numpy.ndarray, or pd.DataFrame.\n\n# Normalization\nxx_norm = ed.norm.z(xx)\nxx_minmax = ed.norm.minmax(xx)\n\n# Resampling\nxx_resampled = ed.resample(xx, fs, TGT_FS)\n\n# Noise addition\nxx_gauss = ed.add_noise.gauss(xx)\nxx_white = ed.add_noise.white(xx)\nxx_pink = ed.add_noise.pink(xx)\nxx_brown = ed.add_noise.brown(xx)\n\n# Filtering\nxx_filted_bandpass = ed.filt.bandpass(xx, fs, low_hz=LOW_HZ, high_hz=HIGH_HZ)\nxx_filted_bandstop = ed.filt.bandstop(xx, fs, low_hz=LOW_HZ, high_hz=HIGH_HZ)\nxx_filted_gauss = ed.filt.gauss(xx, sigma=SIGMA)\n\n# Hilbert Transformation\nphase, amplitude = ed.hilbert(xx) # or envelope\n\n# Wavelet Transformation\nwavelet_coef, wavelet_freqs = ed.wavelet(xx, fs)\n\n# Power Spetrum Density\npsd, psd_freqs = ed.psd(xx, fs)\n\n# Phase-Amplitude Coupling\npac, freqs_pha, freqs_amp = ed.pac(x_3d, fs) # This function is computationally demanding. Please monitor the RAM/VRAM usage.\n```\n\n# Alias\n[`mngs.dsp`](https://github.com/ywatanabe1989/mngs/src/mngs/dsp/) has the same functionalities.\n\n# Contact\nYusuke Watanabe (ywata1989@gmail.com).\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Easy digital signal processing",
"version": "1.3.0",
"project_urls": {
"Homepage": "https://github.com/ywatanabe1989/ezdsp"
},
"split_keywords": [
"numpytorch",
" pandas",
" lfp",
" eeg",
" meg"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "c8ed278df2dfe0d490e63d78ac451f09c7efe932c5a79caac45c8a734f4237f6",
"md5": "fad0f39885fbb5dba529b436c6a95ba6",
"sha256": "ec95448ad67041693baa82ba13a0b1958953f08d339f0524f60675a7d83a1c48"
},
"downloads": -1,
"filename": "ezdsp-1.3.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "fad0f39885fbb5dba529b436c6a95ba6",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.6",
"size": 32114,
"upload_time": "2024-04-08T15:30:30",
"upload_time_iso_8601": "2024-04-08T15:30:30.275768Z",
"url": "https://files.pythonhosted.org/packages/c8/ed/278df2dfe0d490e63d78ac451f09c7efe932c5a79caac45c8a734f4237f6/ezdsp-1.3.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "e87f823afc937fdce6f233b7d58e3f9649f9617fa7de495d738e4657864e869b",
"md5": "c13279757fd03ae73e83727df7437288",
"sha256": "5a08533ba57dae2f05b4ddaf883d984fcb9bd3dea52a3fd4a841491a3fda8080"
},
"downloads": -1,
"filename": "ezdsp-1.3.0.tar.gz",
"has_sig": false,
"md5_digest": "c13279757fd03ae73e83727df7437288",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 24123,
"upload_time": "2024-04-08T15:30:32",
"upload_time_iso_8601": "2024-04-08T15:30:32.253538Z",
"url": "https://files.pythonhosted.org/packages/e8/7f/823afc937fdce6f233b7d58e3f9649f9617fa7de495d738e4657864e869b/ezdsp-1.3.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-04-08 15:30:32",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "ywatanabe1989",
"github_project": "ezdsp",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [
{
"name": "julius",
"specs": []
},
{
"name": "matplotlib",
"specs": []
},
{
"name": "mne",
"specs": []
},
{
"name": "numpy",
"specs": []
},
{
"name": "numpy_fn",
"specs": []
},
{
"name": "pandas",
"specs": []
},
{
"name": "ripple_detection",
"specs": []
},
{
"name": "scipy",
"specs": []
},
{
"name": "tensorpac",
"specs": []
},
{
"name": "torch",
"specs": []
},
{
"name": "torch_fn",
"specs": []
},
{
"name": "torchaudio",
"specs": []
},
{
"name": "torchsummary",
"specs": []
},
{
"name": "mngs",
"specs": [
[
"==",
"1.2.3"
]
]
}
],
"lcname": "ezdsp"
}
JSON
