location of maintained code

[daler]

@bgruening do you think it would be useful to add this to bioconda-utils? I was about to update the bioconda docs to point to this repo, but then realized it would be helpful if it came already bundled as a CLI utility like bioconda-utils conda-forge-r pkgs.txt.

Then again it might be useful to keep this repo independent of bioconda-utils since it should track conda-forge policy rather than bioconda.

Thoughts?

[bgruening]

Currently, its just a major hack and I had hoped to upstream it. But now with cb3 - its not worth anymore. So not sure what to do with it in the meantime.

@MathiasHaudgaard has worked on a python port. Maybe this is cleaner?

That said, I'm fine with moving it to bioconda-utils and make it more useable, discover-able in short term.

[jdblischak]

4 years later this hack is still the documented method for generating a recipe for an R package for submission to conda-forge. I had hoped that it would be replaced by grayskull, but I no longer think we should keep waiting (conda/grayskull#7).

The bioconda-utils port is a step in the right direction (e.g. easier to install). However I'm not sure it is the best fit for conda-forge for multiple reasons:

1. Not all conda-forge contributors use (or are familiar with) bioconda. Requiring them to use another channel is an added barrier, especially if they are new to conda-forge
2. conda-forge and bioconda have separate requirements for their recipe submissions. Having a dedicated conda-forge tool would allow the tools to specialize without conflict
3. Using bioconda-utils requires installing it and all its dependencies in the base env (Update docs for newbie developers (+favicon) bioconda/bioconda-utils#145). This is inconvenient if all you want to do is submit a recipe to conda-forge. I just attempted to install bioconda-utils in a separate env to see if maybe the limited capacity of recipe cleaning would be supported, but it failed (even running bioconda-utils --version failed)

Here's my proposal:

1. Let's create a pip-installable Python package to replace the scripts in this repo. It will call conda skeleton and then perform some post-processing
2. Add the new package to conda-forge
3. Update the conda-forge docs to use this new package for creating R recipes
4. Archive this repo

Thoughts?

[jdblischak]

The other main alternative would be to upstream all of these post-processing edits as optional flags to conda skeleton cran. Again, since grayskull is not going to support CRAN recipes any time soon, this now seems worth the effort. I started this by adding a flag to remove the instructional comments. Feedback welcome!

conda/conda-build#4302

[jdblischak]

I'm going ahead with the plan to upstream our preferences to the CRAN skeleton in conda-build. My PR for --no-comments was merged and released, so I've updated our helper script in PR #53.