Closed sadielbartholomew closed 4 years ago
@sadielbartholomew Many thanks for the PR. Code distribution is something I'm really need to gain more knowledge about. The fixes look reasonable. I can also confirm that I made a conda syntax error regarding the 2nd point.
The installation instructions are largely fine, but I have noticed a few issues whilst conducting my review for openjournals/joss-reviews#2407, & it is more efficient for me to correct these as suggestions via a Pull Request than opening an Issue to try to explain them. Namely:
conda create
commands in the docs under 'Install from conda (experimental)' are both broken due to the syntax but work if one component is moved (see below);Detail on (1)
Following the conda installation commands exactly as in the README, I am able to successfully install IPART and (it then runs the tests etc. as working) with
pip install -e .
, but the alternative provided,python setup.py develop
, does not work for me using either the Python 2 or 3 environments set up according to the previous commands. When I run it I hit the error:(the same error arises in both the Python 2 and 3 environment cases). Something about the
cdutil
package means that the command is not sufficient. There may be a way to adapt it to make it work but sincepip install -e .
works fine anyway I suggest in this PR to just state to use that instead. Though feel free to address this issue in another way if you wish (in which case I can remove it from the PR).Detail on (3)
Note the conda error raised for the first such command as-is, fixed when the
ipart
package specification is moved before the-c
. (It may be that the current commands work on certain conda versions, perhaps that you were using when testing these commands, but I believe the working example here is the standard syntax that should be guaranteed to work across any version & is therefore safer):