I was overjoyed a few years ago to learn that I could make a Makefile executable with a shebang:
#!/usr/bin/env make -f
That way I could invoke my workflow with ./my-pipeline instead of make -f my-pipeline.
I've since began favoring Snakemake over GNU Make for analysis workflows. Is there a simple way to implement similar behavior with Snakemake?
Depending on your platform and whether or not you need to pass additional arguments, you might be able to do this easily, or do something hacky to make it work.
If
#!/usr/bin/env make -f
works for you, that's fantastic because it implies you can pass arbitrarily many parameters via the shebang (e.g. if you are on MacOS this probably works). So you should just be able to say:
#!/usr/bin/env snakemake -s
The trick here is that the filename of the script you are currently executing will get passed as the last argument to whatever you call in the shebang line, so this line will run snakemake using the file itself as input, whatever it is called (i.e. it doesn't have to be called Snakefile).
For me, that doesn't work because I can only pass one argument via the shebang line on my Linux system. If I know the path to snakemake, and I do not need to ever specify additional command line arguments, I can still make it work like this:
#!/usr/local/bin/snakemake -s
If you need to do anything more complicated, or cross-platform, this might be a good place to start: How to use multiple arguments for awk with a shebang (i.e. #!)?
