deployment of toy/examplary dataset

[durr1602]

Hi! I have a question related to the deployment of a toy dataset. For my workflow, I have put all files in config/ so it is automatically deployed. I then realized what I was aiming for (confirm installation by running the workflow on the toy dataset, then run the workflow again for an actual dataset) required to either re-deploy or manually remove the results/ and logs/ folders to make sure there wouldn't be any incompatibility with the user's actual dataset. If a --directory flag is specified, then I can isolate the toy dataset from any actual one, but now config/config.yaml should be located in said directory (which isn't deployed). So, one idea would be to package the toy dataset with its config to a tarball than add something like && curl after the snakedeploy command line. I'm just curious to know what people "usually" do, if this is something that has been discussed, etc.

Any help would be greatly appreciated :)

[durr1602]

two other possibilities:

1. instead of && curl, add a few lines after deployment to create toy/config/config.yaml (toy being the --directory for a run on the toy dataset)
2. keep the toy dataset in config/ but specify the configfile on the command line. Since --directory toy --configfile config/config.yaml works (contrasting with configfile behavior in common.smk/Snakefile which is relative to the working directory), I can simply instruct the user to either specify the flag or prepare config/config.yaml under the directory specified by --directory.