The stages a run passes through, and how far along each one is
Source:R/pipeline.R
pipeline_stages.Rdrun_taupatch() announces each stage as it starts. This turns those
announcements into a position on a progress bar, so a caller watching a long
run can tell a covariate download from a model fit.
Details
The fractions are where a stage begins, and they are weighted by how long each takes rather than spread evenly: fetching covariates is most of a real run and labelling patches is instant, so an even split would sit at 40% for twenty minutes and then race through the rest.
Matched on the message rather than signalled through a condition class, because the messages are the pipeline's existing interface to anyone watching it and a second channel would be a second thing to keep in step.
Examples
pipeline_stages()
#> pattern label
#> 1 ^Loading zooplankton data Reading the station database
#> 2 ^Fetching environmental covariates Downloading covariates from Copernicus
#> 3 ^Preparing covariates before the join Resampling covariates before the join
#> 4 ^Fetching bathymetry Downloading bathymetry
#> 5 ^Deriving covariates Computing derived covariates
#> 6 ^Matching covariates to stations Matching covariates to stations
#> 7 ^Attaching climate indices Attaching climate indices
#> 8 ^Labeling patches Labelling high-abundance patches
#> 9 ^Jackknifing covariates Testing covariates by jackknife
#> 10 ^Fitting model Fitting and cross-validating the model
#> 11 ^Projecting monthly suitability Projecting monthly maps
#> 12 ^Output written to Writing output
#> at
#> 1 0.02
#> 2 0.06
#> 3 0.55
#> 4 0.60
#> 5 0.64
#> 6 0.72
#> 7 0.76
#> 8 0.78
#> 9 0.79
#> 10 0.86
#> 11 0.94
#> 12 0.99