output sub directory for every user facing function (#179)

* output sub directory for every user facing function * align names * input checks * update references to output directories in documentation --------- Co-authored-by: CJ Yetman <[email protected]>
RMI-PACTA · Oct 30, 2024 · 7f441a3 · 7f441a3
1 parent 15bcce5
commit 7f441a3
Show file tree

Hide file tree

Showing 14 changed files with 122 additions and 55 deletions.
diff --git a/R/analyse.R b/R/analyse.R
@@ -19,6 +19,19 @@
 #' @examples
 #' # TODO
 analyse <- function(config) {
+  config <- load_config(config)
+
+  output_analysis_dir <- get_output_analysis_dir(config)
+
+  stop_if_not_length(output_analysis_dir, 1L)
+  stop_if_not_inherits(output_analysis_dir, "character")
+
+  if (dir.exists(output_analysis_dir)) {
+    warning("Output directory `dir_analysis` already exists. The existing directory will be removed and replaced with the output of the current run.")
+    unlink(output_analysis_dir, recursive = TRUE)
+  }
+  dir.create(output_analysis_dir, recursive = TRUE, showWarnings = FALSE)
+
   run_pacta(config)
   run_aggregate_alignment_metric(config)
   plot_aggregate_loanbooks(config)

diff --git a/R/load_config.R b/R/load_config.R
@@ -22,36 +22,20 @@ get_abcd_sheet <- function(params) {
   params[["file_names"]][["sheet_abcd"]]
 }
 
-get_output_dir <- function(params) {
-  params[["directories"]][["dir_output"]]
-}
-
 get_output_prepare_dir <- function(params) {
-  file.path(
-    params[["directories"]][["dir_output"]],
-    "prepare_abcd"
-  )
+  params[["directories"]][["dir_prepared_abcd"]]
 }
 
 get_output_matched_loanbooks_dir <- function(params) {
-  file.path(
-    params[["directories"]][["dir_output"]],
-    "matched_loanbooks"
-  )
+  params[["directories"]][["dir_matched_loanbooks"]]
 }
 
 get_output_prio_diagnostics_dir <- function(params) {
-  file.path(
-    params[["directories"]][["dir_output"]],
-    "prioritized_loanbooks_and_diagnostics"
-  )
+  params[["directories"]][["dir_prioritized_loanbooks_and_diagnostics"]]
 }
 
 get_output_analysis_dir <- function(params) {
-  file.path(
-    params[["directories"]][["dir_output"]],
-    "analysis"
-  )
+  params[["directories"]][["dir_analysis"]]
 }
 
 get_scenario_tms_filename <- function(params) {

diff --git a/R/match_loanbooks.R b/R/match_loanbooks.R
@@ -21,9 +21,19 @@
 match_loanbooks <- function(config) {
   config <- load_config(config)
 
+  # input paths for match_loanbooks
   input_loanbooks_dir <- get_loanbook_dir(config)
   output_prepare_dir <- get_output_prepare_dir(config)
+
+  # output path for match_loanbooks
   output_matched_loanbooks_dir <- get_output_matched_loanbooks_dir(config)
+
+  if (dir.exists(output_matched_loanbooks_dir)) {
+    warning(
+      "Output directory `dir_matched_loanbooks` already exists. The existing directory will be removed and replaced with the output of the current run. NOTE: This includes any manually matched loan book files you may have placed in this directory."
+    )
+    unlink(output_matched_loanbooks_dir, recursive = TRUE)
+  }
   dir.create(output_matched_loanbooks_dir, recursive = TRUE, showWarnings = FALSE)
 
   matching_by_sector <- get_match_by_sector(config)

diff --git a/R/plot_aggregate_loanbooks.R b/R/plot_aggregate_loanbooks.R
@@ -35,9 +35,6 @@ plot_aggregate_loanbooks <- function(config) {
   by_group <- get_by_group(config)
   by_group <- check_and_prepare_by_group(by_group)
 
-  dir.create(output_analysis_aggregated_dir, recursive = TRUE, showWarnings = FALSE)
-
-
   # load required data----
 
   if (is.null(by_group)) {

diff --git a/R/prepare_abcd.R b/R/prepare_abcd.R
@@ -18,6 +18,19 @@
 #' @examples
 #' # TODO
 prepare_abcd <- function(config) {
+  config <- load_config(config)
+
+  output_prepare_dir <- get_output_prepare_dir(config)
+
+  stop_if_not_length(output_prepare_dir, 1L)
+  stop_if_not_inherits(output_prepare_dir, "character")
+
+  if (dir.exists(output_prepare_dir)) {
+    warning("Output directory `dir_prepared_abcd` already exists. The existing directory will be removed and replaced with the output of the current run.")
+    unlink(output_prepare_dir, recursive = TRUE)
+  }
+  dir.create(output_prepare_dir, recursive = TRUE, showWarnings = FALSE)
+
   remove_inactive_companies(config)
   prepare_sector_split(config)
 }
diff --git a/R/prepare_sector_split.R b/R/prepare_sector_split.R
@@ -33,12 +33,9 @@
 prepare_sector_split <- function(config) {
   config <- load_config(config)
 
+  # input/output paths for prepare_sector_split
   output_prepare_dir <- get_output_prepare_dir(config)
-
-  dir.create(output_prepare_dir, recursive = TRUE, showWarnings = FALSE)
-
   sector_split_path <- get_sector_split_path(config)
-
   advanced_company_indicators_path <- get_advanced_company_indicators_path(config)
   sheet_advanced_company_indicators <- get_advanced_company_indicators_sheet(config)
 

diff --git a/R/prioritise_and_diagnose.R b/R/prioritise_and_diagnose.R
@@ -19,6 +19,19 @@
 #' @examples
 #' # TODO
 prioritise_and_diagnose <- function(config) {
+  config <- load_config(config)
+
+  output_prio_diagnostics_dir <- get_output_prio_diagnostics_dir(config)
+
+  stop_if_not_length(output_prio_diagnostics_dir, 1L)
+  stop_if_not_inherits(output_prio_diagnostics_dir, "character")
+
+  if (dir.exists(output_prio_diagnostics_dir)) {
+    warning("Output directory `dir_prioritized_loanbooks_and_diagnostics` already exists. The existing directory will be removed and replaced with the output of the current run.")
+    unlink(output_prio_diagnostics_dir, recursive = TRUE)
+  }
+  dir.create(output_prio_diagnostics_dir, recursive = TRUE, showWarnings = FALSE)
+
   run_match_prioritize(config)
   run_calculate_match_success_rate(config)
   run_calculate_loanbook_coverage(config)

diff --git a/R/remove_inactive_companies.R b/R/remove_inactive_companies.R
@@ -1,13 +1,13 @@
 remove_inactive_companies <- function(config) {
   config <- load_config(config)
 
+  # input path for remove_inactive_companies
   path_abcd <- get_abcd_path(config)
   sheet_abcd <- get_abcd_sheet(config)
 
+  # output path for remove_inactive_companies
   output_prepare_dir <- get_output_prepare_dir(config)
 
-  dir.create(output_prepare_dir, recursive = TRUE, showWarnings = FALSE)
-
   remove_inactive_companies <- get_remove_inactive_companies(config)
   start_year <- get_start_year(config)
   time_frame <- get_time_frame(config)

diff --git a/R/run_calculate_loanbook_coverage.R b/R/run_calculate_loanbook_coverage.R
@@ -1,6 +1,7 @@
 run_calculate_loanbook_coverage <- function(config) {
   config <- load_config(config)
 
+  # input/output paths for loan book coverage
   output_prepare_dir <- get_output_prepare_dir(config)
   output_prio_diagnostics_dir <- get_output_prio_diagnostics_dir(config)
 

diff --git a/R/run_calculate_match_success_rate.R b/R/run_calculate_match_success_rate.R
@@ -1,6 +1,7 @@
 run_calculate_match_success_rate <- function(config) {
   config <- load_config(config)
 
+  # input/output paths for match success rate
   dir_input <- get_input_dir(config)
   input_loanbooks_dir <- get_loanbook_dir(config)
   output_prio_diagnostics_dir <- get_output_prio_diagnostics_dir(config)

diff --git a/R/run_match_prioritize.R b/R/run_match_prioritize.R
@@ -1,11 +1,12 @@
 run_match_prioritize <- function(config) {
   config <- load_config(config)
 
+  # input paths for match prioritize
   output_matched_loanbooks_dir <- get_output_matched_loanbooks_dir(config)
   output_prepare_dir <- get_output_prepare_dir(config)
 
+  # output path for match prioritize
   output_prio_diagnostics_dir <- get_output_prio_diagnostics_dir(config)
-  dir.create(output_prio_diagnostics_dir, recursive = TRUE, showWarnings = FALSE)
 
   match_prio_priority <- get_match_priority(config)
 

diff --git a/example.config.yml b/example.config.yml
@@ -1,7 +1,10 @@
 default:
   directories:
     dir_input: "path/to/input_folder"
-    dir_output: "path/to/output_folder"
+    dir_prepared_abcd: "path/to/prepared_abcd"
+    dir_matched_loanbooks: "path/to/matched_loanbooks"
+    dir_prioritized_loanbooks_and_diagnostics: "path/to/prioritized_loanbooks_and_diagnostics"
+    dir_analysis: "path/to/analysis"
   file_names:
     filename_scenario_tms: "scenarios_2022_tms.csv"
     filename_scenario_sda: "scenarios_2022_sda.csv"

diff --git a/vignettes/config_yml.Rmd b/vignettes/config_yml.Rmd
@@ -33,7 +33,10 @@ The `directories`{.yaml} section contains options to define locally accessible p
 ```yaml
   directories:
     dir_input: "~/Desktop/test/input"
-    dir_output: "~/Desktop/test/output"
+    dir_prepared_abcd: "~/Desktop/test/prepared_abcd"
+    dir_matched_loanbooks: "~/Desktop/test/matched_loanbooks"
+    dir_prioritized_loanbooks_and_diagnostics: "~/Desktop/test/prioritized_loanbooks_and_diagnostics"
+    dir_analysis: "~/Desktop/test/analysis"
 ```
 
 #### dir_input
@@ -44,12 +47,36 @@ The `directories`{.yaml} section contains options to define locally accessible p
     dir_input: "~/Desktop/test/input"
 ```
 
-#### dir_output
+#### dir_prepared_abcd
 
-`dir_output`{.yaml} is a path to a directory where the outputs should be saved. It must be a single string/character value, and it must refer to a valid, accessible, local directory. As an example:
+`dir_prepared_abcd`{.yaml} is a path to a directory where the outputs of the function `prepare_abcd()` should be saved. It must be a single string/character value, and it must refer to a valid, accessible, local directory. As an example:
 
 ```yaml
-    dir_output: "~/Desktop/test/output"
+    dir_prepared_abcd: "~/Desktop/test/prepared_abcd"
+```
+
+#### dir_matched_loanbooks
+
+`dir_matched_loanbooks`{.yaml} is a path to a directory where the outputs of the function `match_loanbooks()` should be saved. It must be a single string/character value, and it must refer to a valid, accessible, local directory. As an example:
+
+```yaml
+    dir_matched_loanbooks: "~/Desktop/test/matched_loanbooks"
+```
+
+#### dir_prioritized_loanbooks_and_diagnostics
+
+`dir_prioritized_loanbooks_and_diagnostics`{.yaml} is a path to a directory where the outputs of the function `prioritise_and_diagnose()` should be saved. It must be a single string/character value, and it must refer to a valid, accessible, local directory. As an example:
+
+```yaml
+    dir_prioritized_loanbooks_and_diagnostics: "~/Desktop/test/prioritized_loanbooks_and_diagnostics"
+```
+
+#### dir_analysis
+
+`dir_analysis`{.yaml} is a path to a directory where the outputs of the function `analyse()` should be saved. It must be a single string/character value, and it must refer to a valid, accessible, local directory. As an example:
+
+```yaml
+    dir_analysis: "~/Desktop/test/analysis"
 ```
 
 ## file_names: