…nge :-)

Just a few errors in the tests in test_somd.py and test_generalunit.py

I suspect these may be because I haven't yet pulled in any changed from
Sire devel to feat_web. I'll get around to this next week :-)

… feature-web

to cope with the new search code in Sire

…t the new version of Sire

…o run shlex.split properly on windows. Debugging unit errors

…he same version as core

… command lines. Switched out shlex.split for this new function

… feature-web

…int statements

… feature-web

…es_num() as this is loading

a system from a pickled file, and the residue number is not correct.

… edge case bug

caused by using both Sire.Mol and sire.legacy.Mol.

Have some small edge cases to deal with now

FAILED test/Process/test_somd.py::test_pert_file[test/input/morphs/morph01.pickle-test/input/morphs/morph01.pert] -
ModuleNotFoundErr...
FAILED test/Process/test_somd.py::test_pert_res_num - ModuleNotFoundError: No module named 'Sire'
FAILED
test/Sandpit/Exscientia/Process/test_somd.py::test_pert_file[test/Sandpit/Exscientia/input/morphs/morph01.pickle-test/Sandpit/Exscientia/input/morphs/morph01.pert]
FAILED test/Sandpit/Exscientia/Process/test_somd.py::test_pert_res_num - ModuleNotFoundError: No module named 'Sire'
FAILED test/Sandpit/Exscientia/_SireWrappers/test_molecule.py::test_makeCompatibleWith - ValueError: 'Invalid search query:
'element F'
FAILED test/_SireWrappers/test_molecule.py::test_makeCompatibleWith - ValueError: 'Invalid search query: 'element F'

The test_somd failures are because we now can't unpickle a file that expects to load Sire instead of sire

The 'element F' failure is unexpected and will be debugged later

…turned zero results

would not raise an exception. I have fixed this for Molecule, but it will likely need
fixing for others.

I've marked the pickle tests in test_somd.py as XFAIL, as we will need to create a
new pickle object for this to work with the new version of Sire.

Otherwise, everything now passes.

…dows.

Updated to the same requirements system as sire, specifying all requirements
in `requirements.txt` and `requirements_build.txt`. These are parsed
by `update_recipe.py` to create the conda recipe.

I have moved the sire version requirement into requirements.txt so that
it is easier to specify the sire version on which a pull request or
push depends. This way we don't need to be playing around with the
tags or anything. This also lets us control the range of supported
sire versions more easily, e.g. not having to use the major number
as the pin.

I have also updated the build process so that we don't need to
clone the sire or biosimspace repositories again - for the tests
we just copy the tests from the already-cloned repo.

Note that this won't build an run now because it depends
on sire >= 2023.0.0, < 2023.1.0

… feature-web

… feature-web

…into feature-web

…into feature-kcombu

…ames

…even on Windows)

…pace into feature-web

allow the user to be explicit about what they are iterating over
when iterating over a search result

… is returned.

Added in the `.bonds()` function to show how to iterate over bonds (although
we do need to add a BSS.Bond object).

Fixed (I think) the perturbable search in the Exscientia sandpit

…orgotten print debug

…an be

properly handled as part of the return from a search query

… feature-web

… on another box

[ci skip]

[ci skip]