rajas / ecgtk Goto Github PK

Dear Raja

I'm using your code to read bw.dat record from PhysioNet nstdb database:

data, info = wt.rdsamp('/home/user/physionet/nstdb/bw')

The info is correct

{'signal_count': 2, 'signal_names': ['noise1', 'noise2'], 'samp_freq': 360.0, 'file_format': ['212', '212'], 'samp_count': 650000, 'first_values': [-29.0, 25.0], 'units': ['mV', 'mV'], 'zero_values': [0.0, 0.0], 'gains': [0.0, 0.0]}

but the data contains -Inf, Inf values for entire 3rd and 4th columns

Sincerely Yuriy

Using the rdann wfdbtools.py code to read annotation file from mitdb created with OSEA/EP Ltd Open Source Annotation software, I came accross the following issue:

Reading an annotation type 59 (SKIP) the annot_time is calculated with the lines (246:247):
annot_time.append(arr[i+2, 0] + (arr[i+2, 1] << 8) +
(arr[i+1, 0] << 16) + arr[i+1, 1] << 24)

Final bitshift is has lower precedence than additions and is therefore applied last. In my particular case, the incorrect time resulted in an empty summary.

Code should read:
annot_time.append(arr[i+2, 0] + (arr[i+2, 1] << 8) +
(arr[i+1, 0] << 16) + (arr[i+1, 1] << 24))