Upcoming & To Do

Future Plans

The bycon package had been implemented specifically to a) drive and test run features during Beacon protocol development, while b) serving the Progenetix use case. This led to some overall complexity - keep Progenetix running w/ existing features while testing bycon and also implement the whole Beacon code base while not necessarily making use of all.

With a feature-rich but overly complex bycon package fulfilling those requirements ongoing work mainly targets:

• disentanglement of non-Beacon code and Progenetix specific configuration from the package
• simplification of configuration and processing pipelines to the emerging practices of a "matured" Beacon v2.n protocol

Bugs & TODO

• create a class which runs all requests against the "data" entry types (i.e. g_variants, biosamples, individuals, analyses, runs) and simplify beacon.py to a pure remapper(?)
• add non-CNV variants to standard plots
• clean & reduce handovers, e.g. do not provide handovers for sample variants (all variants for a sample can be downloaded through the REST path & sample id as /biosamples/{id}/g_variants, w/ optional &output=pgxseg etc.)
• consistant test suite as set of URLs & script for running over them / checkin the responses (e.g. testMode but also w/ queries...)
• flattening of the Progenetix database models w/ mapping methods for standard Beacon model output (& beyond, like Phenopackets)
  • using the "datatable_mappings" paradigm as now already implemented for the ByconVariant class
• move the "services" collection to a generic _byconServicesDB
  • no problem for querybuffer and beaconinfo which can be generated ad hoc
  • other affected collections would be genes and geolocs which need to be moved => _
  • special Progenetix project collections would have to be considered separately (e.g. w/ hard-coded progenetix db?): publications and ontologymaps => hard coded to progenetix
• split installation method into separate parts for beaconServer and services (the latter then in byconaut)
• add method to subset samples for multi-histogram generation
• byconaut: option for summary histogram over? under? samplesplot
• script for auto-generation of parameter documentation
• fix filter queries for correct no-match:
  • query type indicates that the filter is "collationed"
  • query_generation looks for term in collations to perform term expansion
  • BUG: if term is not found -> currently term is not added to final query object -> query is only based on other parameters
• disentangle general configurations, resources (which stay with the package) and instance-specific ones and load them from their appropriate locations
  • beaconServer and services scripts need to (over-) load from configs within their directories
  • this is in principle already possible, just need disentanglement etc.
• Data: fix publication data for consequent inclusion of ISO3166alpha2 codes
• fix publications.py default filters "PMID" and "genomes:>0"