Changelog • harp

harp v0.2.2

define_domain() has been improved so that it can take a proj string, e.g. "+proj=lcc +lon_0=15 +lat_0=63.3 +lat_1=63.3 +lat_2=63.3 +R=6371000" as the proj argument.
geo_reproject() has been added to enable data frames with latitude and longitude columns to be expressed in projection coordinates.

New parameters added to OBSOUL reading.
Partial fix of bug that resulted in errors when extracting domain information from WRF output files - care should be taken when reading data from WRF files as this is not fully tested.

Fixed a bug that caused errors in check_obs_against_fcst() for deterministic forecasts
Fixed bug that meant lag_forecast() and shift_forecast() did not accept the harp >= v0.2.0 column naming convention

Adds functionality to generate time sequences with seq_secs(), seq_mins(), seq_hours() and seq_days()

Fix a bug whereby read_point_obs() failed when not all accumulation periods are available in the SQLite file
Add “ws” to list of known parameters in parse_harp_parameter()

When no projection string is available, get the projection information from CF compliant attributes
Get dimension names from variables. In general this means that dimension names no longer need to be set via netcdf_opts()

Major updates in all harp packages.
Introduction of new package {harpCore} which provides classes and functionalities that are common to all harp packages, including methods for geographic transformations via geo_<.> functions and fast neighbourhood smoothing of gridded fields via the nbhd_smooth() function.
A lot of internal changes to make things run more smoothly.

Arguments start_date, end_date, and by to read functions are deprecated and replaced by dttm, to be used together with seq_dttm() to generate a sequence of date-time strings. Note that you can still use the old arguments, but will be periodically warned to change.
New column names in outputs to read functions

bind_fcst() is deprecated. bind() should be used instead.
get_filenames() is deprecated. generate_filenames() should be used instead.
msub() is deprecated. psub() should be used instead.
read_det_interpolate() and read_eps_interpolate() are defunct. read_forecast(..., transformation = "interpolate") should be used instead.
read_obs_convert() is defunct. read_obs(..., transformation = "...") should be used instead.

Parameters are defined via the internal data list harp_params. This list includes parameter name substitutions for different file formats and the possibility to apply a function to specific parameters at read time.
The parameter list can be added to or modified with add_param_def() and modify_param_def() respectively.
All data read in by harp read functions attain a class - the print method for each class tells you what type of dataset it is. One exception is for read_obs() and read_point_obs(), which both return basic tibbles (data frames).
use_grib_stepRange() is added as a helper to select grib messages via grib_opts(param_find = list(param = use_grib_stepRange(...))). This is useful for determining between accumulated and instantaneous variables with the same grib shortName. {lead_time} can be used to take the current lead time.

Default groupings argument changed from "leadtime" to "lead_time" for consistency with changes in {harpIO}. Scripts that use verification functions with "leadtime", "validdate" and "fcdate" as values to groupings should be changed to "lead_time", "valid_dttm" and "fcst_dttm" respectively.
Verification outputs now use the column name fcst_model instead of mname so that there is consistency throughout harp.
Attributes for verification outputs have changed to include all forecast dates, all stations and all groupings used in the verification. Any scripts that make use of these attributes should be updated to reflect the new attributes.
scale_point_forecast() and scale_point_obs() are deprecated. scale_param() should be used instead.
gather_members() and spread_members() are deprecated. pivot_members() should be used instead.
first_validdate() and last_validdate() are deprecated. unique_valid_dttm() should be used instead.
pull_stations() is deprecated. unique_stations() should be used instead.
bootstrap_score(), pooled_bootstrap_score() and bind_bootstrap_score() are defunct. bootstrap_verify() and bind_point_verif() should be used instead.

Verification functions have gained new progress bars and are generally less verbose in what they are doing, restricting messages to only progress with computing scores for different verification groups.
New verification score hexbin. This gives a data frame of what is essentially a heat map of forecast - observation value pairs.
New class, attributes and print method for verification function outputs.
jitter_fcst() now accepts vectorized functions so should be a lot faster.

plot_scatter() is soft deprecated. Since the data are computed in verification functions, scatter (or more accurately hexbin) plots can now be made with plot_point_verif(..., score = hexbin)
plot_station_eps() is defunct. plot_station_ts() should be used instead.

Plot functions are backwards compatible so that verification data produced by old versions of {harpPoint} with leadtime, mname, fcdate, validdate included in the column names will still work.
geom_georaster(), geom_geocontour() and geom_geocontour_filled() are new functions for plotting gridded data using ggplot. An upscaling option is added to enable faster plotting of rasters and calculation / smoothing of contours.
Map data for plots projected to the domain of a gridded data field can be retrieved using get_map().
The shiny app for point verification can now select from different time axes and will recognise and plot vertical profile verifications.
The shiny app for point verification has gained two new options: full_dir_navigation controls whether a modal is opened for selecting data directories (the old behaviour), or simply populating a dropdown selector. theme allows you to control the overall appearance of the app and you can choose between “dark”, “light” and “white”.