test: modularize test suite, fix S7 checks, and add comprehensive coverage (PR 7)

[aitorvv]

Description

This Pull Request implements PR 7 from the development roadmap: a modular and comprehensive test suite for cidree-silviculture.

• Modularized the monolithic test suite (replacing test-dendrometry.R with dedicated files like test-stand-density.R, test-sample-size.R, etc.).
• Fixed S7 object inspection checks using proper extraction logic (S7::prop()).
• Corrected various package-level calculations and validation issues (e.g., dominant height diameter validation and basal area multipliers).
• Calibrated test expectations for LiDAR metrics (lid_lhdi) and thinning functions.

Verification

devtools::load_all("cidree-silviculture")

# 1. Verify bugfix in LHDI diversity index precision:
lidar_heights <- c(1.1, 1.2, 1.3, 1.6, 1.7, 2.1, 2.2, 2.3)
lid_lhdi(lidar_heights) # Expected to run without warnings and return exactly: 0.7144

# 2. Verify Sample Size module (Stratified) doesn't abort with correct syntax:
df_dummy <- data.frame(
  strata_id = rep(c("A", "B"), each = 4),
  vol = c(10, 12, 15, 11, 20, 22, 25, 23),
  area = rep(c(5, 10), each = 4)
)
sample_strat <- silv_sample_size_stratified(
  data = df_dummy, x = vol, strata = strata_id, total_area = area, 
  plot_size = 500, method = "prop", quiet = TRUE
)
str(sample_strat) # Must return a data.frame with proportional sampling sizes

# 3. Verify Thinning Simulation object extraction via S7
df_plot <- inventory_samples |> 
  dplyr::filter(plot_id == 8) |>
  dplyr::count(species, dclass = silv_tree_dclass(diameter)) |>
  dplyr::mutate(ntrees_ha = silv_density_ntrees_ha(n, plot_size = 10))

thinning_result <- silv_treatment_thinning(
  df_plot, dclass, dclass, ntrees_ha, thinning = "below", perc = 0.3, .groups = "species"
)
names(S7::prop(thinning_result, "data")) # Must include: "ntrees_ha_extract", "ntrees_ha_remain"