Submitted to CRAN 11/10/2020
Thanks to @jsta for updating the README and helping with several other under the hood fixes.
There are two overall major updates with this release. (1) All functions should work with all temporal periods and all variables; previously some functions only worked with daily data and not all variables were able to be downloaded from the prism website. (2) A new API was implemented that results in many functions being deprecated in favor of the updated naming convention. This change was intended to provide consistent names for functions that apply to different steps in the work flow implemented in this package. The details of these changes are:
prism_*_dl_dir()) that set and check the prism download directory. These are hopefully easier to remember than using the base R
getOption()functions and the prism option variable name “prism.path”.
check_path(), which is deprecated and will be removed in the next release.
prism_archive_subset(). This makes it much easier to get the data for a specific type/temporal period from the prism archive. (#69)
ls_prism_data(), which will be removed in a future release.
prism_archive_ls(), which now always returns only folder names as a vector, instead of a data.frame that could have between 1 and 3 columns.
pd_to_file()) to the vector returned by
del_early_prov(), which will be removed in a future release. It also now works with all time steps and prompts user to select which folders will be removed before removing them (when R is in interactive mode). (#89)
check_corrupt(), which will be removed in a future release. It also now works with time steps other than daily and it gains a
download_corruptargument that controls whether corrupt files are automatically re-downloaded.
prism_archive_pd()both return vectors of prism data folder names, i.e., prism data, i.e.,
pd. There are a number of functions that act on the prism data. These are the
prism_md()will be removed in a future release. It is replaced by:
pd_to_file()) and get the type (parameter) of the prism data (
keep_pre81_monthsparameter. This lets the user determine if all of the monthly and annual data are kept, since the download includes all 12 months + the annual data for years before 1981. If this is
TRUEthen all monthly data are kept, instead of only those that were specified in the current call to
keep_pre81_monthsparameter in previous downloads, the download functions now check that pre-1981 data does not exist before downloading it. To do this,
pre81_monthsparameter, which allows the functions to know which months were requested for downloading. (#81)
prism_check()is deprecated and will be no longer be exported in the next release.
get_prism_normals()will now error if neither monthly nor annual data are specified to be downloaded and will download monthly and annual data simultaneously if asked to do so. (#77)
get_prism_*()functions are documented in same help page. (#79)
get_prism_station_md()) now reports a warning if not all requested dates exist in the metadata data frame. (#87 related). It also now works for monthly and normals; not solely daily prism data.
pd_get_md()was added to parse .info.txt metadata, by converting an existing internal function. (#88)
del_early_prov()) now invisibly returns the folders that it removes.
prism_image()) invisibly returns the
ggobject it creates. It also shows the units for the prism variable in the fill legend. (#99)
get_prism_monthlys()can now download 1981 data. (@sdtaylor #59, #63)
get_prism_annual()can now download pre 1981 data by itself. (@rabutler #64)
get_prism_dailys()now correctly sets the progress bar.
get_prism_dailys()works with only the
datesparameter specified. (@rabutler #66)
Changed acquisition method to use prism webservice. Prior to this change the FTP acquisition method often caused the server to time out when download request volume became too high.
Changed method of metadata extraction. Originally we parsed XML metadata to get information about raster files. However the XML for historical data is particularly sparse. The new method relies on parsing file names instead. While more universal, it may be less stable