ec50estimator 1.0.0

• Added plot_EC50_curves() for plotting raw dose-response observations and fitted drc curves across isolates, strata, and one or more candidate models.
• estimate_EC50() and ec50_multimodel() now store plotting metadata and fitted models, allowing plot_EC50_curves(fit) without repeating the formula, data, grouping columns, or model functions.
• Added ec50_estimates(), ec50_metadata(), fitted_models(), and curve_data() helpers for working directly with fitted EC50 objects.
• Added workflow helpers for model selection, fit quality, failed fits, prediction, reporting, data checks, and residual diagnostics.
• Reorganized pkgdown and vignettes around the user workflow: check data, fit models, diagnose fits, select models, plot curves, predict, and report.
• Added a new getting-started vignette and rewrote the single-model and multimodel vignettes to emphasize fit objects as the main interface.
• Removed row-name warnings from predict_ec50() and residual_data().
• Added clearer validation for formula, response, and dose inputs, including stable empty outputs when no groups can be fitted.
• Reset row names in extractor outputs and preserved model-selection columns for all-failed multimodel fits.
• Made quiet = TRUE suppress informational messages from internal drc fitting calls.
• Added a clear error when plot_EC50_curves() is asked for models = "best" without a fitted ec50_multimodel object.
• Modernized EC50 estimation internals with shared validation, clearer errors, and warnings for failed isolate-level fits.
• Kept runtime dependencies focused on drc and ggplot2.
• Removed vignette-only dependencies on cowplot and ggridges.
• Added support for type = "relative" and type = "absolute" in estimate_EC50().
• Added automated tests for grouped estimates, argument validation, absolute effective-dose levels, and multimodel output.
• Converted the example dataset to standard package data format.
• Refreshed the vignette, README, and pkgdown configuration.