Silcamgui crashes when trying to export a summary figure with no particles

[davidrwilliamson]

Describe the bug
If you try to export a summary montage and load a -STATS.csv file that has no particles, you get the following errorL

0it [00:00, ?it/s]
/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py:124: RuntimeWarning: invalid value encountered in true_divide
vd = tpvol / sv # micro-litres / litre
/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py:140: RuntimeWarning: invalid value encountered in double_scalars
nc = np.sum(count) / sv
/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py:696: RuntimeWarning: invalid value encountered in true_divide
nd = np.float64(nd) / sample_volume # nc per size bin per litre
/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py:701: RuntimeWarning: invalid value encountered in less
nd[nd<0] = np.nan # and nan impossible values!
/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py:198: RuntimeWarning: invalid value encountered in greater
ind = np.argwhere(nd>0)
Traceback (most recent call last):
File "/home/dave/PycharmProjects/PySilCam/pysilcam/silcamgui/silcamgui.py", line 458, in export_summary_figure
self.configfile, monitor=False)
File "/home/dave/PycharmProjects/PySilCam/pysilcam/plotting.py", line 281, in summarise_fancy_stats
settings.PostProcess, oilgas=oilgas)
File "/home/dave/PycharmProjects/PySilCam/pysilcam/postprocess.py", line 199, in nc_vc_from_stats
nd[ind[0]] = np.nan
IndexError: index 0 is out of bounds for axis 0 with size 0

Process finished with exit code 134 (interrupted by signal 6: SIGABRT)

Pysilcam will quit or freeze after this error.

To Reproduce
Use the attached file to try exporting a summary figure from.
run_D20210829T100138-STATS.csv

Expected behavior
Should probably present a message saying that the loaded stats file has no particles, and continue without crashing.

Additional context
Exporting a summary time series works OK and exports an empty time series.

[davidrwilliamson]

Low priority bug, you should rarely load a stats file with no particles. I'll submit a pull request to fix it by adding a 0-length check when I get around to it.

[emlynjdavies]

what would you like the behavior to be in this case? if there are no particles, there is anyway nothing to plot.

a warning message and error catching so you don't have to reload the GUI is smart but what is supposed to happen afterwards?

[davidrwilliamson]

Yes, I think exactly that would be good. It can just warn that there are no particles, not create a plot and return you to the interface so you can continue. I believe it currently creates empty oil and gas plots and then freezes.