# Polars Loess
This is a loess local regression (locally estimated scatterplot smoothing) implementation in Rust for Polars.
**This is an early release. There is room for improvement in terms of performance, memory efficiency and feature set.
I value feedback from the community to see where to go next.**
## Installation
```bash
pip install polars-loess
```
## API
```python
loess(
'x-column', # Name of the existing x-values
'y-column', # Name of the existing y-values
'new-x-column', # Name of the new x values. y-values are interpolated for these x-values using loess
# Optional float to specify the fraction of the data used in each local regression.
# Exactly one of frac or points must be specified.
frac=None,
# Optional integer to specify the number of points used in each local regression.
# Exactly one of frac or points must be specified.
points=None,
# Optional integer to specify the degree of the polynomial used in each local regression.
# Default is 1.
degree=None,
)
```
## Example
```python
import polars as pl
from polars_loess import loess
df = pl.DataFrame({
'time': [
0.5578196, 2.0217271, 2.5773252, 3.4140288, 4.3014084,
4.7448394, 5.1073781, 6.5411662, 6.7216176, 7.2600583,
8.1335874, 9.1224379, 11.9296663, 12.3797674, 13.2728619,
14.2767453, 15.3731026, 15.6476637, 18.5605355, 18.5866354
],
'price': [
18.63654, 103.49646, 150.35391, 190.51031, 208.70115,
213.71135, 228.49353, 233.55387, 234.55054, 223.89225,
227.68339, 223.91982, 168.01999, 164.95750, 152.61107,
160.78742, 168.55567, 152.42658, 221.70702, 222.69040,
],
})
result = df.with_columns(loess = loess('time', 'price', 'time', frac=0.5))
print(result)
```
Another example can be found in `run.py`. The result looks like.
## Acknowledgements
* The loess implementation is based on https://github.com/joaofig/loess-rs
* The Python bindings are based on https://marcogorelli.github.io/polars-plugins-tutorial/
Raw data
{
"_id": null,
"home_page": null,
"name": "polars-loess",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "polars, loess, regression, smoothing, data",
"author": null,
"author_email": "Frank Sauerburger <frank@sauerburger.com>",
"download_url": "https://files.pythonhosted.org/packages/05/2a/ef7f8601878d88af75186cfa125e1cd59ddb3837cb95727468cbe4f2bd73/polars_loess-0.1.1.tar.gz",
"platform": null,
"description": "# Polars Loess\n\nThis is a loess local regression (locally estimated scatterplot smoothing) implementation in Rust for Polars.\n\n**This is an early release. There is room for improvement in terms of performance, memory efficiency and feature set.\nI value feedback from the community to see where to go next.**\n\n## Installation\n\n```bash\npip install polars-loess\n```\n\n## API\n\n```python\nloess(\n 'x-column', # Name of the existing x-values\n 'y-column', # Name of the existing y-values\n 'new-x-column', # Name of the new x values. y-values are interpolated for these x-values using loess\n \n # Optional float to specify the fraction of the data used in each local regression.\n # Exactly one of frac or points must be specified.\n frac=None,\n \n # Optional integer to specify the number of points used in each local regression.\n # Exactly one of frac or points must be specified.\n points=None,\n \n # Optional integer to specify the degree of the polynomial used in each local regression.\n # Default is 1.\n degree=None,\n)\n```\n\n\n## Example\n\n```python\nimport polars as pl\nfrom polars_loess import loess\n\ndf = pl.DataFrame({\n 'time': [\n 0.5578196, 2.0217271, 2.5773252, 3.4140288, 4.3014084,\n 4.7448394, 5.1073781, 6.5411662, 6.7216176, 7.2600583,\n 8.1335874, 9.1224379, 11.9296663, 12.3797674, 13.2728619,\n 14.2767453, 15.3731026, 15.6476637, 18.5605355, 18.5866354\n ],\n 'price': [\n 18.63654, 103.49646, 150.35391, 190.51031, 208.70115,\n 213.71135, 228.49353, 233.55387, 234.55054, 223.89225,\n 227.68339, 223.91982, 168.01999, 164.95750, 152.61107,\n 160.78742, 168.55567, 152.42658, 221.70702, 222.69040,\n ],\n})\nresult = df.with_columns(loess = loess('time', 'price', 'time', frac=0.5))\nprint(result)\n```\n\nAnother example can be found in `run.py`. The result looks like.\n\n\n\n## Acknowledgements\n\n\n* The loess implementation is based on https://github.com/joaofig/loess-rs\n* The Python bindings are based on https://marcogorelli.github.io/polars-plugins-tutorial/\n\n",
"bugtrack_url": null,
"license": null,
"summary": "LOESS algorithm for Polars",
"version": "0.1.1",
"project_urls": {
"Homepage": "https://gitlab.sauerburger.com/frank/polars-loess"
},
"split_keywords": [
"polars",
" loess",
" regression",
" smoothing",
" data"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "7ce70baedf5563dd3cfe846b66626728b7004e89c97534d08aa9a7994c029a08",
"md5": "293865b34d814f5db98876748bb4dcb6",
"sha256": "89e7d8e225414c3d8a01b8174204d54905522463211caab35b6ac7d73a1f0748"
},
"downloads": -1,
"filename": "polars_loess-0.1.1-cp38-abi3-macosx_11_0_arm64.whl",
"has_sig": false,
"md5_digest": "293865b34d814f5db98876748bb4dcb6",
"packagetype": "bdist_wheel",
"python_version": "cp38",
"requires_python": ">=3.9",
"size": 3192109,
"upload_time": "2024-10-06T15:34:34",
"upload_time_iso_8601": "2024-10-06T15:34:34.350386Z",
"url": "https://files.pythonhosted.org/packages/7c/e7/0baedf5563dd3cfe846b66626728b7004e89c97534d08aa9a7994c029a08/polars_loess-0.1.1-cp38-abi3-macosx_11_0_arm64.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "071ec3c26d9369ec8145da2f33478683fcc2cb829263ee6c517a7f3fe3deb66c",
"md5": "76d14bbf6c7b0fb454a53b51cfa35a1e",
"sha256": "3d5648bbeab1bc4112a7bb85d6d6c49d8b19935ac0f6ed6f0232e0b000546961"
},
"downloads": -1,
"filename": "polars_loess-0.1.1-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"has_sig": false,
"md5_digest": "76d14bbf6c7b0fb454a53b51cfa35a1e",
"packagetype": "bdist_wheel",
"python_version": "cp38",
"requires_python": ">=3.9",
"size": 3946714,
"upload_time": "2024-10-06T15:27:37",
"upload_time_iso_8601": "2024-10-06T15:27:37.792304Z",
"url": "https://files.pythonhosted.org/packages/07/1e/c3c26d9369ec8145da2f33478683fcc2cb829263ee6c517a7f3fe3deb66c/polars_loess-0.1.1-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "052aef7f8601878d88af75186cfa125e1cd59ddb3837cb95727468cbe4f2bd73",
"md5": "7d665994b599472462c6ddfa10cf6dc7",
"sha256": "36be59dd253e0764827dc813d5fef7b52a09c14ec4764e7bbc45732b2d634475"
},
"downloads": -1,
"filename": "polars_loess-0.1.1.tar.gz",
"has_sig": false,
"md5_digest": "7d665994b599472462c6ddfa10cf6dc7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 209623,
"upload_time": "2024-10-06T15:27:40",
"upload_time_iso_8601": "2024-10-06T15:27:40.654283Z",
"url": "https://files.pythonhosted.org/packages/05/2a/ef7f8601878d88af75186cfa125e1cd59ddb3837cb95727468cbe4f2bd73/polars_loess-0.1.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-10-06 15:27:40",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "polars-loess"
}
JSON