add configuration example for yellowtail

data-raw/configuration_yellowtail.R

@@ -0,0 +1,90 @@
+# TODO list
+# * Only pull data once per species and then combine the pull results with the
+#   configuration matrix again
+# * Fix how format_data returns an object without the nwfscSurvey class instead
+#   it is class(data) > [1] "tbl_df" "tbl" "data.frame"
+# * for vessel_year, might want a different level scaling things might not have
+#   to give this to grid
+configuration <- tibble::as_tibble(read.csv(
+  file.path("data-raw", "configuration.csv")
+))
+
+configuration <- configuration |>
+  dplyr::filter(species == "yellowtail rockfish")
+
+# Change the formula to add year : region interaction
+configuration$formula <- "catch_weight ~ 0 + fyear*region + pass_scaled"
+configuration$knots <- 400
+configuration$spatiotemporal2 <- "off" # based on EW's initial modeling, these worked best
+
+data <- configuration |>
+  # Row by row ... do stuff then ungroup
+  dplyr::rowwise() |>
+  # Pull the data based on the function found in fxn column
+  dplyr::mutate(
+    data_raw = list(format_data(eval(parse(text = fxn)))),
+    data_filtered = list(data_raw |>
+      dplyr::filter(
+        depth <= min_depth, depth >= max_depth,
+        latitude >= min_latitude, latitude <= max_latitude,
+        year >= min_year, year <= max_year
+      ) |>
+        dplyr::mutate(region = ifelse(latitude > 40.1666667, "N", "S")))
+  ) |>
+  dplyr::ungroup()
+
+# Confirm no data in the south in 2007:
+dplyr::filter(data$data_filtered[[1]], catch_weight > 0) |>
+  dplyr::group_by(region, year) |>
+  dplyr::summarise(n = n())
+
+# Find variables that aren't identifiable for presence-absence model
+lm <- lm(formula = as.formula(configuration$formula),
+         data = data$data_filtered[[1]])
+not_identifiable <- names(which(is.na(coef(lm))))
+# Find variables that aren't identifiable for positive model
+lm_pos <- lm(formula = as.formula(configuration$formula),
+         data = dplyr::filter(data$data_filtered[[1]], catch_weight>0))
+pos_not_identifiable <- names(which(is.na(coef(lm_pos))))
+
+
+.map <- names(coef(lm))
+.map[names(coef(lm)) %in% not_identifiable] <- NA
+.map <- factor(.map)
+.start <- rep(0, length(coef(lm)))
+.start[names(coef(lm)) %in% not_identifiable] <- -20
+
+.map_pos <- names(coef(lm_pos))
+.map_pos[names(coef(lm_pos)) %in% pos_not_identifiable] <- NA
+.map_pos <- factor(.map)
+.start_pos <- rep(0, length(coef(lm_pos)))
+.start_pos[names(coef(lm_pos)) %in% pos_not_identifiable] <- -20
+
+best <- data |>
+  dplyr::mutate(
+    # Evaluate the call in family
+    family = purrr::map(family, .f = ~ eval(parse(text = .x))),
+    # Run the model on each row in data
+    results = purrr::pmap(
+      .l = list(
+        data = data_filtered,
+        formula = formula,
+        family = family,
+        anisotropy = anisotropy,
+        n_knots = knots,
+        share_range = FALSE,
+        spatiotemporal = purrr::map2(spatiotemporal1, spatiotemporal2, list),
+        sdmtmb_control = list(
+          sdmTMB::sdmTMBcontrol(
+            map = list(b_j = .map, b_j2 = .map_pos),
+            start = list(b_j = .start, b_j2 = .start_pos)
+          )
+        )
+      ),
+      .f = indexwc::run_sdmtmb
+    )
+  )
+
+
+
+