Implement Memoise in WSP generator

[rburghol]

The following functions will cache nhd results. These are preferable to reduce timeouts. We should put them in a function file, then call that file (later to be migrated to vahydro or hydroools). Ex: memo_get_nhdplus()

# set up a caching function for NWIS
library("memoise")
dir <- Sys.getenv("USGS_cache_dir")
if (dir == "") {
  dir <- tools::R_user_dir(package="dataRetrieval", which="cache")
}
if (!dir.exists(dir)) {
  dir.create(dir, recursive=TRUE)
}
# maybe don't use this just yet.
#unlink(dir, recursive = TRUE, force = TRUE)
db <- memoise::cache_filesystem(dir)
one_day <- 24*60^2
one_year <- 365 * one_day
memo_readNWISdv <- memoise::memoise(dataRetrieval::readNWISdv, ~memoise::timeout(one_day), cache = db)
memo_readNWISsite <- memoise::memoise(dataRetrieval::readNWISsite, ~memoise::timeout(one_day), cache = db)
memo_get_nhdplus <- memoise::memoise(nhdplusTools::get_nhdplus, ~memoise::timeout(one_year), cache = db)
memo_get_UT <- memoise::memoise(nhdplusTools::get_UT, ~memoise::timeout(one_year), cache = db)
memo_plot_nhdplus <- memoise::memoise(nhdplusTools::plot_nhdplus, ~memoise::timeout(one_year), cache = db)
memo_navigate_nldi <- memoise::memoise(nhdplusTools::navigate_nldi, ~memoise::timeout(one_year), cache = db)

[megpritch]

Memoise is Working in WSP_Regional_Summaries.Rmd!

• fn_get_memo_nhdplus.R - Located in the functions folder, but right now it doesn't define or require a function call. It currently caches data as stated in the code above, and this script is sourced at the top of the Rmd with the rest of the function and config files.
• If the cached data exists when the Rmd goes to load NHD data, then the memo'd function memo_plot_nhdplus is applied.

Time Saved:

• Both BigSandy_UpperTennessee_1 and Shenandoah_1 now render in under a minute.
• Shenandoah_1 rendered in 78s the first time (before data was cached), and rendered in 54s the second time around. New files can be seen in the caching directory on my computer as well.
• BigSandy_UpperTennessee_1 rendered in 57s once NHD was cached.

Next Steps

• Now that I better understand what is going on with the memoise functions--do we want to cache the entirety of VA's NHD data while running the Dataframe_Generator.Rmd so that in the WSP_Regional_Summaries.Rmd we're able to quickly render many different origins in a row? Rather than only saving time if you're re-running the same origin.