Add STACAS as new method component

[JGarnica22]

Describe your changes

Add methods/stacas new metric component.

Add non-supervised version of STACAS tool for integration of single-cell transcriptomics data.
This functionality enables correction of batch effects while preserving biological variability without requiring prior cell type annotations.

More information at:

• Describing publication: Andreatta et al., 2024
• Implementation in R package: STACAS

Checklist before requesting a review

• I have performed a self-review of my code

• Check the correct box. Does this PR contain:

  • Breaking changes
  • New functionality
  • Major changes
  • Minor changes
  • Bug fixes

• Proposed changes are described in the CHANGELOG.md

• CI Tests succeed and look good!

[rcannood]

@JGarnica22 Could you merge upstream/main into this PR so the github actions can run? 🙇

[JGarnica22]

@JGarnica22 Could you merge upstream/main into this PR so the github actions can run? 🙇

thanks, done!

[JGarnica22]

Hi,

It looks like the check failure is related to reticulate and its Python discovery/installation mechanism.

From the logs, the STACAS package installs fine, but when the test is run, reticulate tries to provision its own Python:

Error in python_config_impl(python) : 
  Error running '/root/.cache/R/reticulate/uv/cache/builds-v0/.../bin/python': No such file or directory

This suggests that the temporary Python environment created by reticulate is not persisted or is being erased/moved between build/test phases inside the container.

I found this same problem when I run the test locally but I could solve it by not installing dependencies via remotes::install_github(). Then everything works fine locally, and reticulate is not broken.

remotes::install_github("carmonalab/STACAS@2.2.0", dependencies = FALSE)

This makes me think the problem is container-specific:

• In the base image (openproblems/base_r:1.0.0), reticulate works fine until dependencies are installed.
• Installing with dependencies = TRUE seems to trigger a rebuild/relocation of reticulate’s Python, which later cannot be found during the test stage.

Would you recommend a preferred way forward here?

• Pinning/using a system-level Python inside the container so reticulate doesn’t try to manage its own?
• Or caching/persisting .cache/R/reticulate between steps?

Thanks for the guidance!

[JGarnica22]

Hi, thanks for your support, I merged the changes.

[JGarnica22]

Added changes into a new PR: