Fixing some return linters

Author: davidrsch

app/logic/determine_iterations.R

@@ -6,17 +6,24 @@ box::use(
 )
 
 #' @export
-determine_iterations <- function(transformations, scales,
-                                 horizon, inp_amount, lstm, epoch, tests, ...) {
+determine_iterations <- function(
+  transformations,
+  scales,
+  horizon,
+  inp_amount,
+  lstm,
+  epoch,
+  tests,
+  ...
+) {
   inp_amount <- str_split(inp_amount, ",")[[1]]
   lstm <- str_split(lstm, ",")[[1]]
   parameters <- c(as.list(environment()), list(...))
-  max_options <- max(unlist(map(parameters, length))
-  )
-  iterations <- parameters |>
+  max_options <- max(unlist(map(parameters, length)))
+  # iterations
+  parameters |>
     map(rep, length.out = max_options) |>
     as_tibble() |>
     expand.grid() |>
     distinct()
-  return(iterations)
 }

app/logic/model.R

@@ -1,7 +1,15 @@
 box::use(
   dplyr[across, bind_cols, contains, lead, left_join, matches, mutate, n],
   dplyr[pull, rename, select, slice],
-  keras3[compile, fit, keras_input, keras_model, layer_dense, layer_lstm, layer_reshape],
+  keras3[
+    compile,
+    fit,
+    keras_input,
+    keras_model,
+    layer_dense,
+    layer_lstm,
+    layer_reshape
+  ],
   purrr[map, reduce],
   stats[na.omit, predict],
   tibble[as_tibble, tibble],
@@ -12,9 +20,13 @@ box::use(
 #' @export
 multiple_lead <- function(data, n) {
   data |>
-    mutate(across(matches(names(data)), \(x) lead(x, n = n), .names = "lead_{col}_n{n}")) |>
+    mutate(across(
+      matches(names(data)),
+      \(x) lead(x, n = n),
+      .names = "lead_{col}_n{n}"
+    )) |>
     mutate(index = dplyr::row_number(), .before = value) |>
-    select(- value)
+    select(-value)
 }
 
 # - Inputs. Gets the data, the amount of historical data to use,
@@ -34,10 +46,10 @@ get_input_vector <- function(data, inp_amount, horizon) {
       select(-index) |>
       na.omit()
   }
-  inputs <- inputs |>
+  # inputs
+  inputs |>
     as.matrix() |>
     array(c(dim(inputs), 1))
-  return(inputs)
 }
 
 # - Outputs. Gets the data, the amount of historical data to use,
@@ -57,10 +69,10 @@ get_output_vector <- function(data, inp_amount, horizon) {
       select(-index) |>
       na.omit()
   }
-  outputs <- outputs |>
+  # outputs
+  outputs |>
     as.matrix() |>
     array(c(dim(outputs), 1))
-  return(outputs)
 }
 
 # Getting model. Receive the historical data, the temporal horizon and
@@ -85,7 +97,6 @@ get_model <- function(inp_amount, horizon, lstm) {
 # - Iterate along the samples to train and obtain predictions.
 #' @export
 fit_predict <- function(model, input_vec, output_vec, epoch) {
-
   predictions <- rep(NA, ((dim(output_vec)[1] - 1) * dim(output_vec)[2])) |>
     matrix(ncol = dim(output_vec)[2]) |>
     array(c((dim(output_vec)[1] - 1), dim(output_vec)[2], dim(output_vec)[3]))
@@ -112,15 +123,16 @@ slide_columns <- function(data, column) {
   c(
     rep(NA, column - 1),
     pull(data, column),
-    rep(NA, dim(data)[2] - column))
+    rep(NA, dim(data)[2] - column)
+  )
 }
 
 # Extract predictions as a tibble,
 # - If temporal horizon is 1 stay as it is
 # - Else find mean value of each time point along the predictions
 #' @export
 twod_predictions <- function(predictions_3d) {
-  predictions <- predictions_3d[, , 1] |>
+  predictions <- predictions_3d[,, 1] |>
     as_tibble(.name_repair = "unique")
 
   if (dim(predictions)[2] == 1) {
@@ -150,8 +162,8 @@ model_flow <- function(data, modeldata) {
   outputs <- get_output_vector(data, inp_amount, horizon)
   model <- get_model(inp_amount, horizon, lstm)
   predictions <- fit_predict(model, inputs, outputs, epoch)
-  predictions <- twod_predictions(predictions)
-  return(predictions)
+  # predictions
+  twod_predictions(predictions)
 }
 
 # Test the process of creating and obtaining predictions by the number
@@ -177,8 +189,8 @@ test_model_flow <- function(data, modeldata) {
     pred_tests[, i] <- model_flow(data, modeldata)
   }
 
-  pred_tests <- pred_tests |>
+  # pred_tests
+  pred_tests |>
     nest() |>
     rename(tests = data)
-  return(pred_tests)
 }

app/logic/process.R

@@ -1,6 +1,24 @@
 box::use(
-  dplyr[across, all_of, arrange, bind_cols, bind_rows, distinct, if_else, left_join],
-  dplyr[mutate, rename, row_number, select, slice, slice_tail, starts_with, summarize_all],
+  dplyr[
+    across,
+    all_of,
+    arrange,
+    bind_cols,
+    bind_rows,
+    distinct,
+    if_else,
+    left_join
+  ],
+  dplyr[
+    mutate,
+    rename,
+    row_number,
+    select,
+    slice,
+    slice_tail,
+    starts_with,
+    summarize_all
+  ],
   purrr[map],
   stats[na.omit],
   stringr[str_detect],
@@ -9,9 +27,9 @@ box::use(
 )
 
 box::use(
-  app/logic/model[test_model_flow],
-  app/logic/scale[get_actual_scales, get_all_scales],
-  app/logic/transform[get_actual_serie, get_all_transformations],
+  app / logic / model[test_model_flow],
+  app / logic / scale[get_actual_scales, get_all_scales],
+  app / logic / transform[get_actual_serie, get_all_transformations],
 )
 
 # Get data to work with by using the specified sequence and forecast
@@ -22,7 +40,6 @@ box::use(
 # "value"
 #' @export
 get_process_data <- function(data, sequence, forecast) {
-
   if (!any(is.null(sequence), sequence == "")) {
     new_data <- data[, c(sequence, forecast)]
     names(new_data) <- c("sequence", "value")
@@ -33,8 +50,8 @@ get_process_data <- function(data, sequence, forecast) {
     names(new_data) <- "value"
   }
 
-  new_data <- na.omit(new_data)
-  return(new_data)
+  # new_data
+  na.omit(new_data)
 }
 
 # Function to returns min and max value of a column if column exist
@@ -73,15 +90,15 @@ get_all_series <- function(data, transformations, scales) {
   se_min_max <- extract_mi_ma(data, "second")
 
   data <- get_all_scales(data, scales)
-  data <- list(
+  # data
+  list(
     data = data,
     first_diff = first_diff,
     second_diff = second_diff,
     ex_min_max = ex_min_max,
     fi_min_max = fi_min_max,
     se_min_max = se_min_max
   )
-  return(data)
 }
 
 # Obtain all the tests for all the model configuration (modeldata) for
@@ -97,28 +114,30 @@ get_predict <- function(name, data, modeldata) {
   predictions <- tibble(tests_results = rep(NA, dim(modeldata)[1]))
 
   for (i in seq_len(dim(modeldata)[1])) {
-    predictions[i, ] <- test_model_flow(data |> select(all_of(name)), modeldata[i, ])
+    predictions[i, ] <- test_model_flow(
+      data |> select(all_of(name)),
+      modeldata[i, ]
+    )
   }
 
   predictions <- predictions |>
     mutate(index = row_number()) |>
     select(index, tests_results)
   modeldata <- modeldata |>
     mutate(index = row_number())
-  predictions <- left_join(modeldata, predictions, by = "index") |>
+  # predictions
+  left_join(modeldata, predictions, by = "index") |>
     select(-index) |>
     nest() |>
     rename(model_data = data) |>
     mutate(data = name) |>
     select(data, model_data)
-  return(predictions)
 }
 
 # Function to compute squared error
 #' @export
 squared_error <- function(x, y) {
-  z <- (y - x)^2
-  return(z)
+  (y - x)^2
 }
 
 # Function to get RMSE from test results and give proper format for
@@ -153,7 +172,7 @@ get_results <- function(data, original) {
 
     rmse[i, ] <- bind_cols(tests_results, pred_original) |>
       mutate(across(starts_with("test_"), \(x) squared_error(x, value))) |>
-      select(- value) |>
+      select(-value) |>
       summarize_all(mean) |>
       mutate(across(starts_with("test_"), \(x) sqrt(x))) |>
       rowMeans()
@@ -179,16 +198,19 @@ get_results <- function(data, original) {
     data[i, "tests_results"] <- tests_results
   }
 
-  data <- bind_cols(data, rmse) |>
+  # data
+  bind_cols(data, rmse) |>
     mutate(
       transformations = str_detect(data, "value") |>
-        if_else("Original", if_else(str_detect(data, "first"), "First", "Second")),
+        if_else(
+          "Original",
+          if_else(str_detect(data, "first"), "First", "Second")
+        ),
       scales = str_detect(data, "z_o") |>
         if_else("0 to 1", if_else(str_detect(data, "m_p"), "-1 to 1", "Exact")),
       .after = data
     ) |>
     select(-c(data, tests))
-  return(data)
 }
 
 # Execute the process using the previously defined functions. Receive as
@@ -224,8 +246,13 @@ process <- function(data, sequence, forecast, iterations) {
     map(\(x) get_predict(name = x, new_data, iterations)) |>
     bind_rows()
 
-  predictions <- get_actual_scales(predictions, ex_min_max, fi_min_max, se_min_max)
+  predictions <- get_actual_scales(
+    predictions,
+    ex_min_max,
+    fi_min_max,
+    se_min_max
+  )
   predictions <- get_actual_serie(predictions, data, first_diff, second_diff)
-  predictions <- get_results(predictions, data)
-  return(predictions)
+  # predictions
+  get_results(predictions, data)
 }

app/logic/scale.R

@@ -9,14 +9,12 @@ box::use(
 # Get scaled series
 #' @export
 zero_to_one <- function(x) {
-  y <- rescale(x = x, to = c(0, 1), from = c(min(x), max(x)))
-  return(y)
+  rescale(x = x, to = c(0, 1), from = c(min(x), max(x)))
 }
 
 #' @export
 minus_to_plus <- function(x) {
-  y <- rescale(x = x, to = c(-1, 1), from = c(min(x), max(x)))
-  return(y)
+  rescale(x = x, to = c(-1, 1), from = c(min(x), max(x)))
 }
 
 # Apply specified scales to data.
@@ -44,9 +42,9 @@ get_all_scales <- function(data, scales) {
   }
   scaled <- list(exact, zero_one, minus_plus)
   no_null <- !unlist(map(scaled, is.null))
-  scaled <- scaled[no_null] |>
+  # scaled
+  scaled[no_null] |>
     bind_cols()
-  return(scaled)
 }
 
 # Rescale tests results to original scale. Receive as input the output
@@ -72,7 +70,8 @@ rescale_back <- function(data, serie, ex_min_max, fi_min_max, se_min_max) {
     t_min_max <- se_min_max
   }
 
-  data <- data |>
+  # data
+  data |>
     filter(data == serie) |>
     unnest(model_data) |>
     unnest(tests_results) |>
@@ -84,7 +83,6 @@ rescale_back <- function(data, serie, ex_min_max, fi_min_max, se_min_max) {
     ) |>
     nest(tests_results = starts_with("test_")) |>
     nest(model_data = c(horizon, inp_amount, lstm, epoch, tests, tests_results))
-  return(data)
 }
 
 # Execute the previous define function only on those tests results that need to

app/logic/test_id_datatables.R

@@ -6,7 +6,8 @@ box::use(
 # input the table name that should be pass to the testid
 #' @export
 test_id_datatables <- function(name) {
-  jsfunction <- paste0(
+  # jsfunction
+  paste0(
     "function(settings, json) {",
     # Add 'data-testid' to filters
     "$(this.api().table().header()).find('tr:eq(1) td').slice(1).each(function(index) {",
@@ -35,5 +36,4 @@ test_id_datatables <- function(name) {
     "}"
   ) |>
     JS()
-  return(jsfunction)
 }