Changelog

Source: NEWS.md

rharv 0.3.0

Release polish and fixes from a pre-release review.

• Fixed rh_size_for(vary = "demand") for targets below 100%: the search bound was clipped at the zero-deficit maximum, silently understating the result (also affected rh_iso_curve()/rh_resource_curve() solving for demand at sub-100 levels).

• rh_size_for(vary = "capacity") now clamps a fixed initial to the trial capacity (it used to error) and accepts a per-day demand vector; it also validates that base contains the required parameters.

• rh_guarantee_curve() returns NA for infeasible points instead of aborting the whole frontier; rh_guaranteed_capacity() detects supply-limited (structurally infeasible) problems early with a clear message.

• rh_plot_resource_curve() falls back to facets when a resource is entirely unreachable; rh_plot_monthly_balance() labels months correctly for partial-year simulations; rh_plot_design_curve() gives a friendly error when grid attributes were lost; rh_series_spread() tolerates all-NA groups; the step solver rejects non-positive steps.

• Shiny explorer: debounced recomputation of the trade-off grid and guarantee values, fast-mode now also applies to the guarantee values, and the initial storage slider stays consistent with the capacity slider.

• The demo notebook (rh_demo()) was rewritten as an annotated, paper-style document (in Portuguese): every section states the model and its equations, runs the code, and explains how to read the result, covering the YAS and YBS operating rules, the mass-balance check, the standard versus legacy attendance algebra, and a reading guide for each chart.

• rh_simulate() documentation now names the two overflow_timing rules by their literature names (YBS = after_demand, YAS = before_demand).

• pkgdown now renders equations properly (MathJax enabled).

rharv 0.2.0

Analysis tools, decision plots and an interactive app.

• rh_sweep() (one parameter) and rh_grid() (two parameters) compute performance metrics over a range of designs, with a fast day-of-year climatology mode; metrics are selectable ("attendance_pct", "reliability_pct", …).
• rh_plot_tradeoff() draws the area x capacity surface with iso-guarantee contours, and rh_plot_levers() compares the levers (more roof, more tank or less demand).
• rh_guarantee_curve() traces the zero-deficit frontier, rh_plot_syr() the storage-yield-reliability curves and rh_plot_design_curve() the dimensionless, site-transferable version.
• rh_size_for() sizes any lever to an arbitrary guarantee level; rh_iso_curve() with rh_plot_iso() draws iso-guarantee curves relating two levers (optionally faceted by a third); rh_resource_curve() with rh_plot_resource_curve() plots the required capacity and area against the guarantee level on a dual axis, the stage-area-volume analogue.
• Diagnostic plots rh_plot_behaviour(), rh_plot_monthly_balance() and rh_plot_failure_calendar().
• Variability and scenario helpers: rh_series_spread() with rh_plot_spread(), rh_seasonal_demand() (monthly demand profile), rh_scenarios_from_years() (dry-to-wet or wet-to-dry stress series) and rh_compare().
• rh_explore() and rh_app() launch a Shiny explorer to vary parameters and watch the behaviour, metrics, trade-off surface and guarantee values update live.
• New analysis-and-sizing vignette, and a full-feature demo notebook shipped with the package: open it in your editor with rh_demo().

rharv 0.1.0

Initial release.

• rh_simulate() runs the daily reservoir water-balance simulation, returning an rharv_sim object with print(), summary(), as.data.frame() and autoplot() methods.
• rh_available_volume() computes the captured rainwater volume from precipitation, area, runoff and efficiency (Vdisp = P * A * C * eta).
• rh_metrics() and accessors (rh_attendance(), rh_deficit_total(), rh_overflow_total()).
• Guarantee-based sizing: rh_guaranteed_demand(), rh_guaranteed_capacity(), rh_required_area(), with bisection, optimize (base R stats::optimize) and step solvers.
• rh_daily_demand() and plotting helpers rh_plot_climatology()/autoplot().
• Example datasets precip_pi and areas_pi, and the ifpb-pi-case-study vignette (a worked example).