adressed review comments on metric coverage.PMU plot label and docs

Author: ayushman1210

modules/benchmark/R/metric_Coverage.R

@@ -25,14 +25,10 @@ metric_Coverage <- function(dat, ...) {
   PEcAn.logger::logger.info("Metric: Prediction Interval Coverage")
 
   valid <- !is.na(dat$obvs) & !is.na(q_low) & !is.na(q_high)
-  
-  if ("obvs_sd" %in% names(dat)) {
-    valid <- valid & !is.na(dat$obvs_sd)
-    covered <- (dat$obvs[valid] - dat$obvs_sd[valid]) <= q_high[valid] &
-               (dat$obvs[valid] + dat$obvs_sd[valid]) >= q_low[valid]
-  } else {
-    covered <- dat$obvs[valid] >= q_low[valid] & dat$obvs[valid] <= q_high[valid]
+  if (!any(valid)) {
+    return(NA_real_)
   }
-  
+
+  covered <- dat$obvs[valid] >= q_low[valid] & dat$obvs[valid] <= q_high[valid]
   return(mean(covered))
 }

modules/benchmark/R/metric_timeseries_plot.R

@@ -50,33 +50,30 @@ metric_timeseries_plot <- function(metric_dat, var, unit = NULL, filename = NA,
     sharpness <- mean(sub_dat$model_q95 - sub_dat$model_q05, na.rm = TRUE)
     bias <- mean(sub_dat$model - sub_dat$obvs, na.rm = TRUE)
 
+    pmu_val_str <- "N/A"
     pass_str <- "N/A"
     if ("obs_se" %in% colnames(sub_dat) && "obs_n" %in% colnames(sub_dat)) {
       valid_pmu <- !is.na(sub_dat$obs_se) & !is.na(sub_dat$obs_n)
       if (any(valid_pmu)) {
         se2_n <- (sub_dat$obs_se[valid_pmu]^2) * sub_dat$obs_n[valid_pmu]
         pooled_var <- sum(se2_n) / sum(sub_dat$obs_n[valid_pmu])
         pmu <- sqrt(pooled_var)
+        pmu_val_str <- sprintf("%.2f", pmu)
 
-        passes_validation <- (coverage_pct >= 90) && (abs(bias) < pmu)
+        passes_validation <- (!is.na(coverage_pct) && coverage_pct >= 90) && (abs(bias) < pmu)
         pass_str <- ifelse(passes_validation, "PASS", "FAIL")
       }
     }
 
     data.frame(
       site = sub_dat$site[1],
       variable = sub_dat$variable[1],
-      label = sprintf("Coverage: %.1f%%\nSharpness: %.2f\nBias: %.2f\nPMU: %s", coverage_pct, sharpness, bias, pass_str)
+      label = sprintf("Coverage: %.1f%%\nSharpness: %.2f\nBias: %.2f\nPMU: %s\nStatus: %s", coverage_pct, sharpness, bias, pmu_val_str, pass_str)
     )
   }))
 
   # Determine if model "passes" at each point based on available intervals
-  if ("obvs_sd" %in% colnames(metric_dat)) {
-    metric_dat$Pass <- (metric_dat$obvs - metric_dat$obvs_sd) <= metric_dat$model_q95 &
-                       (metric_dat$obvs + metric_dat$obvs_sd) >= metric_dat$model_q05
-  } else {
-    metric_dat$Pass <- metric_dat$obvs >= metric_dat$model_q05 & metric_dat$obvs <= metric_dat$model_q95
-  }
+  metric_dat$Pass <- metric_dat$obvs >= metric_dat$model_q05 & metric_dat$obvs <= metric_dat$model_q95
 
   if (any(!is.na(metric_dat$Pass))) {
     metric_dat$Observation_Status <- ifelse(metric_dat$Pass, "Observed (Pass)", "Observed (Fail)")

modules/benchmark/R/run_benchmark.R

@@ -152,7 +152,7 @@ register_metric("COVERAGE", function(dat) {
 #'
 #' @param aligned data.frame with columns: model, obvs, time
 #' @param metrics character vector of metric names
-#' @return data.frame with columns: metric, value
+#' @return data.frame in wide format with columns `Site` and each requested metric column per site.
 compute_metrics <- function(aligned, metrics = c("RMSE", "MAE", "R2")) {
   # Treat data as one group if no site column
   if (!"site" %in% colnames(aligned)) {