Support for conda environments

In order to increase the number of people who will use rhea (and the citations we will get) I propose to support both singularity containers and conda environments. The user can decide whether to use singularity (--use-singularity) or conda (--use-conda) when running snakemake. This means that tests need to be run twice which should not be a problem. The majority of the tools are available in the bioconda channel. We will only need to add our custom tools (e.g. TIN score calculation in bioconda). Let me know what you think.

added Discuss label

I highly agree, in MAPP we support both versions. Honestly, not every aspiring computational scientist has access to Singularity... This might also solve the problem of mac support, assuming bioconda packages are available for macOS as well.

Although this step might require some tedious work to make sure that you generate conda envs with exactly the same package versions as containerized.

For the TIN score calculation I think we can use:
https://anaconda.org/bioconda/rseqc

changed title from support for conda environments to Support for conda environments

added Future label

Yes, definitely nice-to-have. In a similar vein, we could also try to set up the project for execution in the cloud. In both cases though (especially with Conda), maintenance will become more of an issue then.

Oh, only just saw #102 (closed) re: cloud execution. I like.

removed Future label

changed milestone to %pipeline

assigned to @gypas

@gypas @bakma please take a look

@bakma Do you have a guide on how to upload packages in anaconda or do you follow another approach to support both conda and docker images in your pipelines?

If you would like to build and upload a conda package to the Anaconda Cloud it's quite easy. You need to have conda-build which creates a package based on a YAML file. Of course, you need to have a channel you will later upload the code to - set up an account @ anaconda cloud. Please take a look at the following general tutorial: https://docs.anaconda.com/anaconda-cloud/user-guide/tasks/work-with-packages/

As an example take a look at the following repo, which I have uploaded a conda package myself long time ago, as I was playing with a toy example of an R package with some really bad sorting functions. All I had to do is to add this file: https://github.com/AngryMaciek/badsoRts/blob/master/recipe/meta.yaml

added To Do label

added Doing label and removed To Do label

For the custom tools we should upload them to bioconda.

mentioned in issue #150 (closed)

removed Discuss label

mentioned in merge request !89 (merged)

Fixed in !89 (merged)