looper report is intended to create a static html report. looper table provides "Project/Sample statistics and table output generator".
Currently (in Looper 1.5.0), there are two Reporter classes (Reporter and ReporterOld) and two Tabulator classes (Tabulator and TabulatorOld). It appears that with the release of Looper 1.4.0, looper report and table were to only be used when pipestat is configured.
However, looper will currently still use the ReporterOld and TabulatorOld classes if pipestat is not configured.
We've also recently moved the report generation to pipestat.summarize such that looper report (when configured to use pipestat) will call Reporter class which uses pipestat.sumamrize to generate the html report. However, looper table and Tabulator Class are still completely contained within looper.
Some questions:
Do we still want looper table and should we also move the creating of the stats table generation to pipestat and have looper table call something like pipestat.table?
I noticed looper will only use the newer Reporter and Tabulator classes if: 1. pipestat configuration is given in the config file AND 2. we use --project as an argument in the CLI. Otherwise, looper uses the older classes. Is there a reason for requiring these to run at the project level?
Thoughts on path forward:
[x] ensure looper table works with the current versions of pipestat v0.5.0 (currently it appears to be broken).
[x] move stat table generation to pipestat?
[x] remove TabulatorOld, ReporterOld, CheckerOld give the user a warning if they attempt to use report or table without configuring pipestat.
[x] remove html report generation code from looper completely (its still present because ReporterOld is still using it).
[x] write tests to confirm looper can be configured to use pipestat and use the above functions without issue.
In 2020,
looper summarize
was split intoreport
andtable
functions: https://github.com/pepkit/looper/pull/255looper report
is intended to create a static html report.looper table
provides "Project/Sample statistics and table output generator".Currently (in Looper 1.5.0), there are two Reporter classes (
Reporter
andReporterOld
) and two Tabulator classes (Tabulator
andTabulatorOld
). It appears that with the release of Looper 1.4.0,looper report
andtable
were to only be used when pipestat is configured.However,
looper
will currently still use theReporterOld
andTabulatorOld
classes if pipestat is not configured.We've also recently moved the report generation to
pipestat.summarize
such thatlooper report
(when configured to use pipestat) will callReporter
class which uses pipestat.sumamrize to generate the html report. However,looper table
and Tabulator Class are still completely contained within looper.Some questions:
looper table
and should we also move the creating of the stats table generation to pipestat and have looper table call something likepipestat.table
?--project
as an argument in the CLI. Otherwise, looper uses the older classes. Is there a reason for requiring these to run at the project level?Thoughts on path forward:
looper table
works with the current versions of pipestat v0.5.0 (currently it appears to be broken).report
ortable
without configuring pipestat.