ForestElementsR 3.0.0

Major overhaul of the species coding system.

• Species codings may now be hierarchical: a species can appear both as an individual code and inside one or more coarser group codes of the same coding (a laminar family of species sets). Coding tables gained a level column (0 = finest leaf, higher numbers = coarser groups). bavrn_state and tum_wwk_long are now hierarchical.
• Codings may carry non-tree codes — categories without a tree-species equivalent, such as the Bavarian state code 99 (“Strauch”, shrub). Whether a code is a tree code is derived from its link to the species_master_table, so there is no separate flag to keep in sync. New exported helpers fe_species_is_tree() and fe_species_non_tree_codes() report this.
  • fe_stand() and its spatial children now reject non-tree species codes with a clear error.
  • Casting a non-tree code into another coding yields NA (with a message) instead of raising an error.
• Casting into a hierarchical coding resolves each species to the finest code that contains it. A cast that loses information (aggregating finer codes into a coarser group) is now signalled with a message instead of a warning; forward-ambiguous casts still raise an error. A small documented dataset, species_cast_overrides, declares deliberate target codes for otherwise forward-ambiguous casts.
• New exported, CSV-driven builders for maintaining the codings and the master table: master_template_csv(), master_table_from_csv(), coding_template_from_master(), coding_table_from_template(), and cast_overrides_from_csv(). Coding tables are now stored in a canonical row order (level, then code, then master key), and the builder rejects inconsistent or conflicting entries early.
• Extended the species data: new Bavarian state forest species and codes, tum_wwk_long filled out with grouping levels and now covering all master species, and additional master species (e.g. the genera Cedrus and Corylus).
• Added two cast overrides for bavrn_state -> tum_wwk_short: the oak group code 70 (“Eiche (Gruppe)”) now resolves to 6 (oak), and the “other hardwood” code 80 (“Sonstige Laubhölzer”) to 8 (other hardwood). Both casts were forward-ambiguous (70 straddles {6, 8} via red oak, 80 straddles {8, 9} hard/soft broadleaf) and therefore errored. This unblocks downstream casting of BaySF inventory data carrying these group codes.
• New exported helper fe_species_get_field_table() returns a compact, field-usable lookup table for a coding: each code exactly once, with the coding’s own species/group names (all three languages), plus the level and is_tree columns, in canonical order. Rendering it into a printable document is left to downstream packages.
• The “Tree Species Codings” vignette was updated to cover the overhauled system: hierarchical codings and the level column, non-tree codes and is_tree, the finest-node cast and the cast-override mechanism, the cast information-loss message (formerly a warning), the new field table, and a rewritten “Information for developers” part documenting the CSV-driven builder workflow.

Other news

• Exported the object validation helpers has_required_names, has_required_types_or_classes, has_no_missing_values, and is_distinct (previously internal). They remain marked with @keywords internal, but are now part of the public interface so that dependent packages (e.g. FeNEU) can call them with :: instead of relying on the non-exported ::: mechanism, which is not permitted on CRAN.
• Promoted wood_density_tum_wwk_short from an internal object (in sysdata.rda) to a documented, exported dataset. It is now accessible as ForestElementsR::wood_density_tum_wwk_short.

ForestElementsR 2.2.0

• Implemented S3 function get_center() that retrieves the coordinates of a meaningful center point for fe_stand childs that carry spatial information.
• The yield table related functions ytable_age_slice, ytable_max_slice, site_index, ytable_lookup, si_to_mai_age, si_to_mai_max, and stocking_level were made more safe. These functions are not designed for working with input variable vectors of length > 1. However, under certain circumstances such input would trigger a warning, but produce dubious results. Now, any attempt to provide input vectors with more than one element to these functions results in an error.
• Edited all three vignettes.

ForestElementsR 2.1.0

• Implemented automated tests for stand_sums_static()
• Revised function stand_sums_static(); new version of v_gri() allowed to simplify the code.
• Modified the important function v_gri() to make it more tolerant. In case of missing height values, it does not stop anymore but returns NA and issues a warning.

ForestElementsR 2.0.1

Revised CRAN submission

We have addressed the issues raised by Konstanze Lauseker in her review of the initial submission:

• Added references to the description field of the DESCRIPTION file.
  
• Removed the use of ::: even in internal documentation.
  
• Replaced \dontrun{} with try() for examples that intentionally generate errors.
  
• Substituted options(warn = -1) with suppressWarnings() in the automated tests.

ForestElementsR 2.0.0

Initial CRAN submission.