read_abares_trade() returns all character strings as factors that are sorted by "Trade_flow", "Fiscal_year", "Month", "Trade_code", "State", "Australian_port", "Overseas_location".read_abares_trade().read_abares_trade() no longer drops the leading "0" in the "Trade_codes" column.soil_thickness functions are now topsoil_thickness functions, e.g., read_soil_thickness_stars() is now read_topsoil_thickness_stars().
This is to clarify that the data is for topsoil only and not all soil layers, thanks, @obr-soil!
get_ functions are now integrated into read_, there is no need to call any get_ functions separately or use piping.
Caching is no longer possible between sessions to simplify the package's maintenance and CRAN-proof it. I would encourage users to look into using {targets} or other methods of managing their data in the workflow.
When fetching AGFD, users can filter data by climate year or climate/price year depending on the data and only work with that year or years.
This functionality still requires ALL of the AGFD to be fetched, so this isn't faster, but any of the read_agfd functions are faster since they are only reading a smaller portion of data into the active R session, thanks @potterzot!.
All read_ functions now support importing local files and parsing them.
Users may now download the data using methods other than an R session and import local data.
Additional datasets are now serviced:
Users can set options for {read.abares} including:
Improved documentation including:
@source field that points to the file being provided@references field that points to references for the data.Improved testing:
Files are more reliably downloaded rather than timing out for some users, thanks to @obrsoil for the help troubleshooting this issue, painful as it was.
skimr::skim() is used in the vignette to display the AGFD {data.table} formatted data rather than just using head().
{bit64} has been added to the Suggested packages to help users avoid warning messages when working with data in the console via {data.table}, thanks, @potterzot.
Alternative installation instructions using {remotes} are provided in the README for users that may prefer or may not use {pak}, thanks, @econpotter.
The topsoil thickness map now correctly displays proper continuous values rather than classes, thanks, @obr-soil.
The geospatial vignette examples run more quickly, except for the AAGIS examples, which are still slow due to large downloads, but all local operations in R now run much more quickly, thanks to @econpotter.
Rename functions that both download and read files into active R session from get_ to read_ to avoid confusion with functions that only fetch data and have separate read_ functions
Adds new function, print_agfd_nc_file_format() to provide details on the AGFD NetCDF files' contents
Uses Geopackages for {sf} objects rather than .Rds, faster and smaller file sizes when caching
Checks and corrects the geometries of the AAGIS Regions shapefile upon import and applies to the cached object if applicable
Improved documentation
@source field that points to the file being provided@references field that points to references for the dataCode linting thanks to {flint}
Use {httr2} to handle downloads
Use {brio} to write downloads to disk
Use {httptest2} to help test downloads
Gracefully handle errors when AGFD zip files are corrupted on download, provide the user with an informative message and remove corrupted download
Tests are run in parallel for quicker testing
{sf} operations are now quiet when reading data where possible
No longer checks the length of a Boolean vector when checking the number of files in the cache before proceeding with removing them
Fixes bugs in get_agfd() when creating the directories for saving the downloaded file
Fixes bug in get_aagis_regions() when creating the cached object file
Fixes "URL" field in DESCRIPTION file (@mpadge https://github.com/ropensci/read.abares/issues/1)