Specify (short) input names in call to MultiQC

[tomkinsc]

The MultiQC output currently names samples using adapted versions of their full file paths. On machines where the inputs to MultiQC are stored within deep hierarchies (i.e. staged PAPI instances); this results in a MultiQC report where the input (sample) names are so long that they prevent the actual sample name from being displayed.

The call to MultiQC should be adapted to use shorter names—probably the basename of the input files stored within the input-filenames.txt created in the body of task MultiQC. The task should honor values passed via the sample_names input of the task MultiQC, overriding the auto-shortened sample names with those specified in the task input (if present).

Care will be likely be needed to ensure correct sample name ordering if the MultiQC report is extended to include other inputs in addition to those from FastQC.

[tomkinsc]

Or we can simple omit --dirs from the MultiQC invocation, and provide a boolean to optionally enable display of full paths.

[tomkinsc]

Changes in PR #467 to address this.