release 53.0

• a wrapper object npg_pipeline::product to represent a product
• use products attribute to drive p4_stage1, seq_alignment and autoqc
• create composition.json files to guide archiving
• p4 params files for seq_alignment moved from no_cal/laneN to no_cal
  (changes run folder structure when merging lanes)
• cluster_count and seqchksum_comparator checks now done at run level instead
  of lane level
• upfront definition of all products
• generic runfolder scaffolding for any products
• since the top-level qc directory is no longer required, the tileviz
  directory is moved to the analysis directory
• reshuffle of roles in npg_pipeline::roles:
  npg_pipeline::roles::business::base merged into npg_pipeline::base;
  npg_pipeline::roles::business::flag_options moved to
  npg_pipeline::base::options, a number of pipeline options from other
  modules moved to this role;
  npg_pipeline::roles::accessors moved to npg_pipeline::base::config;
  helper functions moved to a new role - npg_pipeline::function::util
• ref_adapter_pre_exec_string method renamed to repos_pre_exec_string
• metadata_cache_dir method, formerly in npg_pipeline::roles::business::base,
  removed; npg_pipeline::function::p4_stage1_analysis module, the only user
  of this function, switched to use the relevant accessor from the
  npg_pipeline::runfolder_scaffold role
• minor changes for bcfstats qc check
• executor type (lsf or wr) can be specified in the configuration file
• wr executor:
  set per-job priority;
  increase priority for p4 stage 1 job and its predecessors;
  set priority of status and start-stop jobs to zero so that
  they are executed immediately, but still within dependencies
  and memory constraints;
  map queues to arbitrary wr options, in particular, a special queue
  for p4_stage1 maps to a specific cloud host flavour
• correction of build method for rpt_list attribute in product
• make bam_cluster_count_check pipeline job dependent on
  qc_spatial_filter (in function_list_central.json)
• archival daemon - limit number of simultaneously archived NovaSeq runs
• wr executor - explicitly propagate pipeline's environment to jobs
• illumina archiver job:
  exclude discontinued verbose attribute and paths that are not needed
  for the minimal work this loader is doing now;
  remove LSF preexec requesting that the job is a unique runner since
  db queries are much simpler now
• change signature of the autoqc archival job in line with extended
  functionality of the autoqc db loader (ability to find JSON files
  in the run folder)
• change components_as_products method of npg_pipeline::product to
  return a list with one item when there is only one component in
  the composition (instead of an empty list)
• tileviz index file with links to lane-level tileviz reports is created
• seq_alignment supports HISAT2 aligner for RNA libraries
• explicit iRODS destination collection is set for iRODS loaders,
  /seq/illumina/runs/RUN_ID for NovaSeq runs and /seq/RUN_ID
  for the rest
• explicitly use iRODS loader from an 'old' dated directory for
  old style runfolders
• a new function, archive_run_data_to_irods, to publish run-level non-product data to iRODS
• modify run_data_to_irods_archiver module to ensure the interop files go to a dedicated directory
• additional tags for NovaSeq in dbic_fixtures