ccp4io/ccp4io_adaptbx present but not configured

[Anthchirp]

>>> import libtbx.load_env
>>> libtbx.env.has_module("ccp4io")
False
>>> libtbx.env.has_module("ccp4io_adaptbx")
False
>>> import ccp4io_adaptbx
>>> ccp4io_adaptbx       
<module 'ccp4io_adaptbx' from '(PREFIX)/lib/python3.8/site-packages/ccp4io_adaptbx/__init__.py'>

however

$ ls -d (PREFIX)/lib/python3.8/site-packages/ccp4io*
lib/python3.8/site-packages/ccp4io  lib/python3.8/site-packages/ccp4io_adaptbx

Does it make sense for the packages to be both present and not configured?

Environment (conda list):

$ conda list
...
cctbx-base                2020.10          py38hc79ae55_0    conda-forge
...

[bkpoon]

So, that part of has not been fully fleshed out. I can add extra logic for the has_module function to return True for installed modules. You can check the dist_path function for now.

import libtbx.load_env

e = libtbx.env

e.dist_path('ccp4io')

The idea will be that software that uses cctbx-base will not need to modify the environment object at all. As long as the modules are installed in the correct place, the environment object will think it is configured.

[Anthchirp]

I noticed this because of a failing test, but the relevant line is in iotbx:
https://github.com/cctbx/cctbx_project/blob/master/iotbx/mtz/extract_from_symmetry_lib.py#L7

[bkpoon]

I'll add the extra logic for completeness to the 2020.11 release which should be starting soon.

[Anthchirp]

Thank you. There is no rush on this