gemsciencetools / hmtk Goto Github PK

When using Numpy v 1.8.1 the plot_observed_recurrence gives the following error

/home/gw/Documents/hmtk/hmtk/hmtk/plotting/seismicity/catalogue_plots.py in plot_observed_recurrence(catalogue, completeness, dmag, end_year, filename, filetype, dpi)
319 completeness,
320 dmag,
--> 321 end_year)
322 plt.figure(figsize=DEFAULT_SIZE)
323 plt.semilogy(recurrence[:, 0], recurrence[:, 1], 'bo')

/home/gw/Documents/hmtk/hmtk/hmtk/plotting/seismicity/catalogue_plots.py in get_completeness_adjusted_table(catalogue, completeness, dmag, end_year)
288 if iloc == n_comp - 1:
289 print temp_rates, obs_idx
--> 290 obs_rates[obs_idx] = temp_rates
291 else:
292 obs_rates[obs_idx[:-1]] = temp_rates
ValueError: NumPy boolean array indexing assignment cannot assign 25 input values to the 26 output values where the mask is true

The documentation of this methods says that one of the two input parameters is a nhlib object. It needs to be updated. See https://github.com/GEMScienceTools/hmtk/blob/master/hmtk/seismicity/selector.py#L166

A setup.py must be added to allow easy installation of the toolkit without the need to change the PYTHONPATH, like:

pip install -e hmtk/

By chance I stumbled on this:
In several functions you use the numpy.arange method.
The way you use it affects e.g.
the plot_depth_histogram method in
hmtk/plotting/seismicity/catalogue_plots.py
where you exclude the largest depth in the catalogue from the plot.

Instead of
numpy.arange(0.,max_depth,bin_width)
I'd do something like:
numpy.arange(0.,max_depth+bin_width,bin_width)

I saw several other lines where this might cause problems

this function returns wrong observed rate values...

this line seems that should be

obs_idx = np.logical_and(mag_bins >= low_mag, mag_bins < high_mag + dmag)