estimate for segmentation.py memory usage

[mr-c]

In the CWL tool description for segmentation.py we need memory sizing hints in mebibytes (2**20).

We can provide this number using a bit of JavaScript.

The following numbers are available for the formula:

• number of inputs JSONs (inputs.bins.length) and the size of each (inputs.bins[N].size)
• Number of cells per file from --cells-per-file (inputs.cells_per_file)
• the size of the --fasta file (inputs.pangenome_sequence.size)
• and the number of CPU cores passed to --parallel-cores (runtime.cores, if that matters)

[mr-c]

Any progress on this @dimatr or others?

[dimatr]

still waiting the deNBI cloud to come up - will test big inputs there to get an impression

[mr-c]

Okay, didn't know that was a blocker. You can also run tests on https://biohackathon.curii.com :-)