np.hstack should not be used in _conclude

[yuxuanzhuang]

Expected behaviour

self._results should be reshaped into a sensible new variable with different n_blocks.

Actual behaviour

if using np.hstack, self._results have to be in the same dimensions.

Code to reproduce the behaviour

from pmda.contacts import Contacts

universe = mda.Universe(PSF, DCD)
sel_basic = "(resname ARG LYS) and (name NH* NZ)"
sel_acidic = "(resname ASP GLU) and (name OE* OD*)"
acidic = universe.select_atoms(sel_acidic)
basic = universe.select_atoms(sel_basic)

ana1 = Contacts((acidic, basic), (acidic, basic), radius=6.0)

ana1.run(n_blocks=3)

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 33 and the array at index 2 has size 32

Fix

• change np.hstack to np.concatenate
• add tests for the case that blocks don't have the same size. (e.g. for DCD file (98 frames), set n_blocks=3)

Sidenote

It seems that this error has been raised multiple times. I guess we should

• avoid mentioning np.hstack in the _conclude documents
• add thorough tests for these uneven block conditions.

Currently version of MDAnalysis:

(run python -c "import MDAnalysis as mda; print(mda.__version__)") 2.0.0-dev (run python -c "import pmda; print(pmda.__version__)") '0.3.0+17.g13fa3b5.dirty' (run python -c "import dask; print(dask.__version__)") 2.19.0