Correct sampto behaviour using rdsamp and rdann?

まず、同じ sampto 値とインデックス 100 のサンプルを使用して、提供された Python の再現コードを wfdb.rdsamp と wfdb.rdann で実行します。各エントリポイントが上限をどのように処理するかを追跡します。完了の条件は、それぞれの範囲セマンティクスに一貫性があり、返されたアノテーションのインデックス指定で IndexError が発生しないことです。

When reading the signal with rdsamp and the corresponding annotations with rdann using the same sampto value results in an IndexError when accessing the signal array it includes an annotation outside the array.

In the example below, to read the first 100 entries, sampto must be 100. If using the same value for the annotations up to and including 100 are read, rather than stopping at 99 (i.e., the 100th entry in a 0-indexed array).

Is this the expected behaviour?

Using: python 3.9.18 with wfdb 4.1.2

import numpy as np
import wfdb

# create test data
n = 250
test_sig = np.zeros((n, 1))
test_sig[:, 0] = np.linspace(0, 1, n)
print(f'original shape: {test_sig.shape}')

# write to file
wfdb.wrsamp('test', fs = 250, units=['mV'], sig_name=['test'], p_signal=test_sig, fmt=['16'])

# create annotation and write
wfdb.Annotation(
    record_name='test', 
    extension='atr',
    sample=np.array([0, 99, 100, 249]), 
    symbol=['N', 'N', 'N', 'N']
).wrann()

samp_from = 0
samp_to = 100

# read values
values, _ = wfdb.rdsamp('test', sampfrom=samp_from, sampto=samp_to)
print(f'sampto shape: {values.shape}')


atr_1 = wfdb.rdann('test', extension='atr', sampfrom=samp_from, sampto=samp_to, shift_samps=True)
print(f'samples: {atr_1.sample}')
# IndexError
# print(f'values at samples: {values[atr_1.sample, :]}')

samp_to_minus_1 = samp_to - 1
atr_2 = wfdb.rdann('test', extension='atr', sampfrom=samp_from, sampto=samp_to_minus_1, shift_samps=True)
print(f'samples: {atr_2.sample}')
# No IndexError
# print(f'values at samples: {values[atr_2.sample, :]}')