Workaround for passing run_spp_nodups.R a filename ending in .gz
for -savn argument.  The script calls gzip internally, which bails
out since the file already has a gz extension.  Instead, rename
file without gz extension and then run gzip.

Add .bam for result and .narrowPeak.gz for bed file types.

Group bed peak results

See merge request !3

Add conf.py entry for the number of result file lines to use for
each document in the bed and peak collections.

Instead of one result file line per document, add them in batches
to reduce the overall number of documents in the collections.
Each document now contains a `result_lines` entry, which is a list
of dictionaries, one result file line per list element.

Added sphinx style documentation to peak_call_generator.  Removed another alignment fragment that was copied over.

Added sphinx style documentation for the align generator.  Adjusted genome loading to use helper functions instead of direct structure access.

Fixed bowtie2 genome prefix argument in align_generator.  Removed double extension error from bwa loading.

Added peak call generator.  Adjusted gem_callpeak yaml removing redundant genome_dir.  Updated peak_call module yaml accordingly, changed some names in macs2 file names.

Updated output files to maintain all jobs up to the point of their creation.  Updated all generator modules accordingly.

Updated marker tracking between steps to maintain all markers. Added remove_duplicates generator & adjusted workflow accordingly.

Adjusted align generator to use new construct_outputs function.  _add_align in workflow changed accordingly.

Created a helper function to construct outputs for individual generator jobs.  Added helper function to print parse results for each run.  Adjusted output file tracking.