Integration with BiocPy/SummarizedExperiment

Hi developers,

Awesome effort to port tximport to Python. On a similar vein, I've been implementing some of the core R/Bioconductor datastructures in Python - BiocPy. I would love to find ways to collaborate, especially interfacing with SummarizedExperiment objects.

Please let me know if you have any questions.

Hi,

thanks for your interest in pytximport & the idea. So far, we had no plans to integrate with SummarizedExperiment, but I believe that it could be of interest and should be fairly straightforward to do so, so we definitely welcome pull requests.

Based on my understanding of SummarizedExperiment, you would just have to add a couple lines of code to support it as another output_format/output_type (format being file extension when saving to disk, type being return type of the function). Relevant code parts are in core/_tximport.py and start here:

pytximport/pytximport/core/_tximport.py

Line 443 in d386a3a

if output_format == "h5ad" and output_type != "anndata":

Since we would like to keep the install size of pytximport somewhat limited, it would be best if SummarizedExperiment was an optional dependency with adequate error handling when it's not installed. Ideally, a PR should also support use through the CLI version of pytximport, the relevant lines start here:

pytximport/pytximport/_cli.py

Line 132 in d386a3a

@click.option(

Before merging any changes into main, unit tests would also have to be implemented but we are happy to help add them on our side if you prefer not to touch that part of the codebase.

If you don't wish to implement any PR yourself but would still like to see SummarizedExperiment supported, please let us know & we will discuss internally whether we should add it to our roadmap.

Best,
Malte