Support three assembly modes: short, long, and hybrid

[dfornika]

Depending on what data is available at any moment, we may need to perform analysis with only short (illumina) reads, only long (Oxford Nanopore aka ONT) reads, or a hybrid analysis using both short and long reads.

The basic strategy for keeping this organized will be:

• Always perform a short-read analysis on all illumina runs. Store the analysis to <config_analysis_output_dir>/<illumina_run_id>/short
• Always perform a long-read analysis on all nanopore runs. Store the analysis to <config_analysis_output_dir>/<nanopore_run_id>/long
• Perform a hybrid analysis for all samples from the illumina run that have corresponding nanopore samples available. Store the analysis to <config_analysis_output_dir>/<illumina_run_id>/hybrid
• Check on every scan to see if more nanopore samples have been made available for each illumina run, and add them to the hybrid analysis.

[dfornika]

Note: we can skip the BCCDC-PHL/taxon-abundance pipeline on the hybrid analyses, since we'll already have the short reads and the long reads classified in the short and long analyses, respectively.