test_hubert.R

files <- list.files(
  "_SharedFolder_GlobalES/Dashboard/Canada/json_files",
  full.names = TRUE
)

# Parse only non-empty files, handling errors gracefully
json_list <- lapply(files, function(file) {
  tryCatch(
    rjson::fromJSON(file = file),
    error = function(e) NULL # Return NULL if parsing fails
  )
})

unique_source_id_values <- unique(unlist(sapply(json_list, function(json) if (!is.null(json)) names(json$source_id) else NULL)))

unique_source_id_values <- unique_source_id_values[grepl("ces", unique_source_id_values)]

# Create a dataframe with the required structure
result <- do.call(rbind, lapply(unique_source_id_values, function(survey_name) {
  data.frame(
    variable = basename(files), # File names
    ces = survey_name, # Survey name
    cleaned = sapply(json_list, function(json) {
      if (!is.null(json) && "source_id" %in% names(json) && 
          survey_name %in% names(json$source_id) && 
          "cleaned" %in% names(json$source_id[[survey_name]])) {
        json$source_id[[survey_name]]$cleaned
      } else {
        NA
      }
    }),
    stringsAsFactors = FALSE
  )
}))

library(ggplot2)
library(dplyr)

# Convertir les données en format approprié pour ggplot2
result_long <- result %>%
  mutate(
    cleaned_color = case_when(
      cleaned == 1 ~ "Cleaned",
      cleaned == 0 ~ "À faire",
      is.na(cleaned) ~ "NA"
    )
  )

# Créer le graphique
ggplot(result_long, aes(x = ces, y = variable, fill = cleaned_color)) +
  geom_tile(color = "white") + # Ajoute des contours blancs pour les cases
  scale_fill_manual(
    values = c("Cleaned" = "lightgreen", "À faire" = "red", "NA" = "gray"),
    name = "Cleaned"
  ) +
  theme_minimal() +
  labs(
    x = "Année (CES)",
    y = "Fichiers JSON",
    title = "Statut des données 'Cleaned' par année et fichier"
  ) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))