Skip to content

Latest commit

 

History

History
123 lines (88 loc) · 9.53 KB

NEWS.md

File metadata and controls

123 lines (88 loc) · 9.53 KB

prism 0.2.1.9000 (development version)

  • Removed global assignment inside prism_webservice()
  • Documented return value for pd_stack(), prism_set_dl_dir() ,pd_to_file(), pd_get_type() & get_prism_*() functions.
  • Added examples to documentation for pd_to_file(), prism_archive_clean(), prism_archive_verify(), prism_set_dl_dir(), & pd_get_*() functions.
  • Added a better error message when the PRISM webservice is down (not returning status 200). (#122)

prism 0.2.1

Released October 17, 2023

Minor Enhancements

  • Added service parameter to get_prism_dailys(), get_prism_monthlys(), and get_prism_annual(), so user can provide subscription based URLs, instead of default public available 4km data. (@adamlilith)
  • Changed defaults in get data functions (#104)
    • get_prism_monthlys()
      • mon now defaults to 1:12, so default now downloads all months for the specified years
      • years no longer has a default of NULL, so user will understand it has to be provided
    • get_prism_annual()
      • years no longer has a default of NULL, so user will understand it has to be provided
    • Minor updates to documentation of mon and years parameters

Bug Fixes and Clean Up

  • Fixing CRAN notes
    • Removed LazyData and LazyLoad from Description to fix CRAN note.
    • Removed lubridate from imports to fix CRAN note.
  • Removed vignette that depended on too many other libraries; did not want to add them all to Suggests and also relied on external shape file. This may be linked to from README in the future.
  • Updated error message in pd_image() when input has length of 0.
  • Fixed documentation for get_prism_normals() to reference the new period (1991-2020). (#111)
  • Updated error message in get_prism_dailys() if a date before Jan. 1, 1981 is provided.
  • Deprecated check parameter in get_prism_dailys() as it does not exist in monthly or annual functions. (#116)
  • Removed purrr from imports by switching one call from purrr:map() to lapply().

prism 0.2.0

Published December 5, 2020

Thanks to @jsta for updating the README and helping with several other under the hood fixes.

Breaking Changes

  • prism_webservice() is no longer exported as it is wrapped by get_prism_*() functions, and requires a correctly specified url. It can still be called with prism:::prism_webservice() if users really need it. (#83)
  • pr_parse() is no longer exported. Use pd_get_name() or pd_get_date() instead.

Major Updates

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:

  • Users can now download vpdmin, vpdmax, and tdmean variables for 30-year normals, daily, monthly, and annual data. (#68)
  • There are now several functions (prism_*_dl_dir()) that set and check the prism download directory. These are hopefully easier to remember than using the base R options() and getOption() functions and the prism option variable name "prism.path".
    • prism_check_dl_dir() replaces check_path(), which is deprecated and will be removed in the next release.
  • The prism data are downloaded to this directory and then referred to as the "prism archive". These are the prism_archive_*() functions.
    • New function: prism_archive_subset(). This makes it much easier to get the data for a specific type/temporal period from the prism archive. (#69)
    • prism_archive_ls() replaces ls_prism_data(), which will be removed in a future release.
      • The return type and options changed from ls_prism_data() to 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.
      • The previous behavior can be achieved by applying new functions (pd_get_name() and pd_to_file()) to the vector returned by prism_archive_ls().
    • prism_archive_clean() replaces 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)
    • prism_archive_verify() replaces 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_corrupt argument that controls whether corrupt files are automatically re-downloaded.
  • prism_archive_ls() and 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 pd_*() functions.
    • pd_image() replaces prism_image().
    • pd_plot_slice() replaces prism_slice().
    • pd_stack() replaces prism_stack().
    • pd_get_station_md() replaces get_prism_station_md(). (#88)
    • prism_md() will be removed in a future release. It is replaced by:
      • pd_get_name(), which is equivalent to prism_md(f, FALSE).
      • pd_get_date(), which is equivalent to prism_md(f, TRUE).
    • Two new functions were added that convert the prism data to a full absolute path (pd_to_file()) and get the type (parameter) of the prism data (pd_get_type()).

Other Changes

  • The way pre 1981 data is handled has been updated.
    • get_prism_annual() and get_prism_monthlys() gain a keep_pre81_months parameter. 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 TRUE then all monthly data are kept, instead of only those that were specified in the current call to get_prism_*(). (#82)
    • Because pre-1981 data might already have been downloaded based on keep_pre81_months parameter in previous downloads, the download functions now check that pre-1981 data does not exist before downloading it. To do this, prism_webservice() and prism_check() gain a pre81_months parameter, which allows the functions to know which months were requested for downloading. (#81)
  • The prism web service only allows a user to download the same data twice in a 24-hour period. The download functions now report when the user has exceeded the allowable number of attempts to download the same file (in one day). If a user tries to download the same file more than two times in one day, the message is posted as a warning and the returned text file is saved in the prism archive. This is not posted as an error so that if a query of multiple files runs into this issue, it does not abort the full query. Another warning posts if the unzipped folder is empty. (#80)
  • 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)
  • all get_prism_*() functions are documented in same help page. (#79)
  • Help pages for non-exported functions have been removed.
  • More tests were added. Up to 50% coverage now. Some tests are only run locally to ensure prism download limits are not exceeded.
  • pd_get_station_md() (formerly 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)
  • prism_archive_clean() (formerly del_early_prov()) now invisibly returns the folders that it removes.
  • pd_image() (formerly prism_image()) invisibly returns the gg object it creates. It also shows the units for the prism variable in the fill legend. (#99)

prism 0.1.0

Minor changes

  • New functions
    • del_early_prov() searches the download folder for duplicated PRISM data and keeps only the newest version.
    • get_prism_station_md() extracts metadata from daily PRISM data.
  • get_prism_dailys() gains a check parameter that allows the user to specify how prism files are checked.

Bug fixes

  • 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.
  • fixed bug in gen_dates() so that get_prism_dailys() works with only the dates parameter specified. (@rabutler #66)

Under the hood

  • added internal gen_dates() function for determining the specified dates (either from minDate and maxDate. or dates) used by the get_prism_*() functions.
  • added tests for gen_dates() .

prism 0.0.7

Changes

  • 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

Bug fixes

  • Fixed FTP time out error by switching to webservice

  • Fixed prism_stack by adjusting to new metadata extraction method