Multiqc-config files will run into conflict when multiple instances of QuaC is run

Currently multiqc config file configs/multiqc_config.yaml gets auto-created by QuaC as part of the pipeline, based on QuaC-Watch thresholds used by user. This will result in conflict when multiple instances of QuaC is run. For example, imagine a scenario where one QuaC instance is run with WGS-specific QuaC-Watch configs and another instance with exome-specific QuaC-Watch configs.

Attempted using python's NamedTemporaryFile module but this didn't work out as we don't want the file to pre-exist. Use of this module will create a temp file when called.

mentioned in commit 9cf8f117

Attempted using uuid module (inspired here) to create a random string unique to a single instance of QuaC. This failed as well because it turns out snakemake recreates a new uuid for every job it triggers, which means that filename would not match between that in parent snakemake log and that in each individual unit job it triggers.

PS- This explains why unit jobs run by snakemake (ie. cluster jobs here) always have stdout which I expected only in the parent snakemake logs. In each unit job, snakemake essentially reruns the pipeline from scratch using options --allowed-rules, --wait-for-files, etc. See the jobscript submitted by snakemake for each unit jobs to learn more about this.

So any implementation that produces random string as (part of the) filename would fail if snakemake sends unit jobs to cluster. It needs to be supplied to snakemake via CLI for such implementation to work.

mentioned in commit a80f9225

mentioned in merge request !2 (merged)

closed with merge request !2 (merged)

mentioned in commit 28bbc54c