search

linnabrown / run_dbcan

Run_dbcan V4, using genomes/metagenomes/proteomes of any assembled organisms (prokaryotes, fungi, plants, animals, viruses) to search for CAZymes.
http://bcb.unl.edu/dbCAN2
GNU General Public License v3.0
138 stars 40 forks source link

Improve packaging and distribution #18

Closed tesujimath closed 5 years ago

tesujimath commented 5 years ago

This software is currently unpackaged. This is what is making it difficult for users. The solution is not more detailed manual instructions, but to package it, so that users may easily install it from e.g. pip or conda.

In order to do this, there are some improvements that need to be made, for example, separating the code from the data, not writing output files into the directory where the scripts live, etc. All of this may be solved by packaging it.

Since the dependencies (diamond, hmmer, prodigal, graggenescan) are available via conda on the bioconda channel, that only leaves signalp to worry about, which greatly simplifies the installation. So I am recommending a conda packaging for dbcan, including the Hotpep scripts.

I am happy to contribute what is required. It will involve some cleaning up of the software. For this repo, this means at least a setup.py file. And then a recipe in bioconda.

Will you consider such a pull request?

linnabrown commented 5 years ago

Thank you for your suggestion. Could you provide us with some manuals of creating bioconda package?

tesujimath commented 5 years ago

I intend to do this for you, and contribute a pull request.

In the meantime, the documentation is here: https://bioconda.github.io/contributor/index.html