Statement of need
The CL-PFU database uses many sources of input data spanning hundreds of data files. When creating the CL-PFU database, matrices are created according to the PSUT framework, final energy data are extended to the useful stage, and energy conversion chains are converted to exergy quantifications. A way to specify all calculation steps is needed. Furthermore, the calculations take many hours to complete, so what is done should stay done during debugging. Calculation pipelines are needed, ones that clearly identify all steps in database creation and ones that can be resumed where needed.
This package (PFUPipeline) provides functions to create several products of the CL-PFU database. The primary objective of PFUPipeline is to create data frames of RUVY matrices in matsindf format. A targets pipeline provides helpful dependency management for the calculations.
Quick start
At the RStudio console, type
library(targets) # to load the targets package
tar_visnetwork() # to see a directed acyclic graph of the calculations that will take place
tar_make_future(workers = 2) # to execute the calculations (or `workers = 8`, if you have enough cores)Accessing targets
A list of targets can be found with PFUPipeline::target_names. A list of target meanings can be found with ?PFUPipeline::target_names.
targets::tar_read(<<target>>) pulls the value of a target out of the targets cache. (<<target>> should be an unquoted symbol such as Specified.)
Fresh start
targets::tar_destroy() invalidates the targets cache and forces reanalysis of everything. Reanalyzing everything may take a while.
More Information
For information about the targets package, see the targets manual.
For documentation on the PFUPipeline package, see https://EnergyEconomyDecoupling.github.io/PFUPipeline/.
History
The CL-PFU database is supported by an EPSRC fellowship awarded to Paul Brockway of Leeds University. A goal of the fellowship is building a world database of country-specific primary, final, and useful exergy for 1960–2019.