Multi-frequency data with plot_wfdb

[bemoody]

Currently, wfdb.plot_wfdb requires the input record to have a uniform d_signal or p_signal, i.e. it must be loaded using smooth_frames=True. If the record is loaded with smooth_frames=False then plot_wfdb complains "The record has no signal to plot".

This makes it impossible to visualize some signals (sample-data/wave_4, for example.)

More generally, being able to plot heterogeneous signals on the same time axis, with each signal at its native sampling frequency is useful for understanding the nature of the data you're working with.

[bemoody]

Also notice that plot_wfdb doesn't understand that the time scale of annotations may not be the same as the time scale of the signals. Example:

>>> os.system('sqrs -r sample-data/100')
>>> wfdb.plot_wfdb(wfdb.rdrecord('sample-data/100', sampto=2000), wfdb.rdann('sample-data/100','qrs',sampto=2000))

Here, the signals are sampled at 360 Hz, but the annotations have a tick frequency of 250 Hz. plot_wfdb assumes one annotation tick = one sample, which is wrong and can be misleading.