Add fitting results on full dataset

Signed-off-by: Liang Zhang <[email protected]>

_scripts/analyze_factors.R

@@ -7,6 +7,25 @@ tar_option_set(
 )
 setup_parallel_plan()
 
+targets_cfa <- tarchetypes::tar_map(
+  list(type = rlang::syms(c("efa", "chc", "cognition"))),
+  tar_target(
+    config_dims,
+    config_dims_alternatives |>
+      select(manifest, latent = type)
+  ),
+  tarchetypes::tar_map(
+    hypers_model,
+    tar_fit_cfa(
+      config_dims,
+      indices_cogstruct_games_censored,
+      theory = theory,
+      missing = "ml",
+      tar_post_fit = "gof"
+    )
+  )
+)
+
 list(
   tarchetypes::tar_file_read(
     indices_cogstruct_games_censored,
@@ -86,5 +105,40 @@ list(
         list_rbind(),
       fit_bootstrap
     )
+  ),
+  tarchetypes::tar_file_read(
+    config_dims_theory,
+    "config/game_dims_theory.csv",
+    read = read_csv(!!.x, col_types = cols(game_id = "I"))
+  ),
+  tar_target(
+    config_dims_alternatives,
+    config_dims_theory |>
+      unite(manifest, game_name_abbr, index_name, sep = ".") |>
+      inner_join(rename(model_efa, efa = latent), by = "manifest")
+  ),
+  tar_target(
+    file_config_dims,
+    config_dims_alternatives |>
+      writexl::write_xlsx("_output/config_dims.xlsx"),
+    format = "file"
+  ),
+  targets_cfa,
+  tarchetypes::tar_combine(
+    gofs,
+    zutils::select_list(targets_cfa, starts_with("gof")),
+    command = list(!!!.x) |>
+      map(as_tibble) |>
+      bind_rows(.id = ".id") |>
+      zutils::separate_wider_dsv(
+        ".id",
+        c(names(hypers_model), "type"),
+        patterns = c(".+?", ".+"),
+        prefix = "gof"
+      )
+  ),
+  tar_target(
+    scores,
+    extract_latent_scores(fit_bf_efa, indices_cogstruct_games_censored)
   )
 )

_scripts/predict_phenotypes.R

@@ -9,7 +9,8 @@ setup_parallel_plan()
 
 config_neural <- prepare_config_neural(
   xcpd == "gsr",
-  run == "full"
+  run == "full",
+  atlas == "4S256Parcels"
 )
 hypers_cpm <- hypers_cpm |>
   dplyr::filter(
@@ -105,7 +106,7 @@ cpm_branches_perms <- tarchetypes::tar_map(
 list(
   tarchetypes::tar_file_read(
     scores,
-    path_obj_from_proj("scores", "cognition_structure"),
+    path_obj_from_proj("scores", "analyze_factors"),
     read = qs::qread(!!.x)
   ),
   tar_prepare_neural_data(config_neural),