Fix hvg subsetting bug

[VladimirShitov]

Changelog

This line in scvi integration component is supposed to subset highly-variable genes: adata = adata[:,adata.var['var_input']].copy()

However, this will not happen for the following reasons:

1. Instead of "var_input" par["var_input"] should be used
2. Even with proper key, adata.var[key] will return a column from adata.var and not the list of selected genes as expected

This PR fixes the issue

Checklist before requesting a review

• [x] I have performed a self-review of my code

• [ ] Conforms to the Contributor's guide

• Check the correct box. Does this PR contain:

  • [ ] Breaking changes
  • [ ] New functionality
  • [ ] Major changes
  • [ ] Minor changes
  • [x] Bug fixes

• [ ] Proposed changes are described in the CHANGELOG.md

• [ ] CI tests succeed!

[DriesSchaumont]

@VladimirShitov Good catch! Would you be able to implement tests?

[VladimirShitov]

Added test and fixed small comments. The function for HVG subsetting is now moved to a separate utils script, as it is used in other components as well. The problem is that the test pipeline fails because there is "No space left on device", so I'm not 100% sure that importing utils works correctly

[DriesSchaumont]

@VladimirShitov The disk issue seemed like a temporary one (?). I am now getting this:

[VladimirShitov]

Yep, this is a problem from my side now 😅 Fixing

[VladimirShitov]

Fixed it :) @DriesSchaumont , can you check?

[DriesSchaumont]

@VladimirShitov Could you adress the comment from Robrecht?

[VladimirShitov]

Done, please check :)

[DriesSchaumont]

LGTM! @rcannood If you would like to have a look