Failing rubberband baseline case

[rguliev]

Description

I got a case, where 'rubberband' baseline fails.
However, I tried a simpler implementation from here and it worked.

Your Setup

• Operating System: Windows 11
• Python version: 3.12.4
• pybaselines version: 1.1.0

Example Code

rubberband_example.csv

import numpy as np
import pybaselines

# Load the data
xy = np.genfromtxt("rubberband_example.csv")
x = xy[:,0]
y = xy[:,1]

# Try the baseline (raises an error)
# IndexError: arrays used as indices must be of integer (or boolean) type
bl = pybaselines.Baseline(x_data=x)
bl.rubberband(y)[0]

Working example:

import scipy
import numpy as np

def rubberband(x: np.array, y: np.array) -> np.ndarray:
    """
    Based on code from Stack Exchange (Author: Roman Kiselev)
    Source: https://dsp.stackexchange.com/a/30723/69444 (CC BY-SA 3.0)
    """
    v = scipy.spatial.ConvexHull(np.array([x, y]).T).vertices
    v = np.roll(v, -v.argmin())
    v = v[: (v.argmax() + 1)]
    bl = np.interp(x, x[v], y[v])

    return bl

rubberband(x,y)

Error Message

  File ~\AppData\Local\Programs\Python\Python312\Lib\site-packages\pybaselines\_algorithm_setup.py:266 in inner
    baseline, params = func(self, y, *args, **kwargs)

  File ~\AppData\Local\Programs\Python\Python312\Lib\site-packages\pybaselines\classification.py:905 in rubberband
    mask[np.unique(total_vertices)] = True

IndexError: arrays used as indices must be of integer (or boolean) type

[derb12]

Thanks for the bug report! This was caused by an incorrect handling of indices within rubberband. I fixed it on the development branch and will try to publish a new release to pypi in a week for two.