Rewriting of uploaded workflow.cfg configurable

[sverhoeven]

Is your feature request related to a problem? Please describe.
Now at

haddock3-webapp/app/models/applicaton.server.ts

Lines 57 to 77 in 811af37

	async function rewriteConfig(config_body: string) {
	const table = parse(dedupWorkflow(config_body), { bigint: false });
	table.run_dir = JOB_OUTPUT_DIR;
	table.mode = "local";
	table.postprocess = true;
	table.clean = true;
	delete table.ncores;
	delete table.batch_type;
	delete table.queue;
	delete table.queue_limit;
	delete table.concat;
	delete table.self_contained;
	delete table.cns_exec;
	// eslint-disable-next-line @typescript-eslint/no-explicit-any
	return stringify(table as any, {
	newline: "\n",
	indent: 2,
	integer: Number.MAX_SAFE_INTEGER,
	});
	}

the uploaded workflow.cfg is rewritten, removing all execution related parameters. This causes haddock3 job to use 4 cpu cores. This does not use all the cores available.

Describe the solution you'd like
It would be nice if the execution section could be overwritten somehow in a configurable way.
For example with an env var to a file like HADDOCK3_EXECUTATION_SECTION=./haddock3.exec.toml with

ncores = 16

Describe alternatives you've considered

1. Change code to adapt to instance (local dev vs production).
2. Reconfigure bartender so it can run multiple haddock3 each with 4 cores. However I want to make a single job go as quickly as possible. Having jobs queued is ok for me.

Additional context
Using infrastructure efficiency also depends on bartender config.

[sverhoeven]

Since #69 the HADDOCK3_NCORES env var sets the ncorers parameter.