global.R

# ==============================================================================
# Polyphenol Estimation Pipeline - Shiny Application
# Global Configuration and Database Loading
# ==============================================================================
#
# PIPELINE DEVELOPED BY:
#   Stephanie M.G. Wilson
#   University of California, Davis
#   Contact: smgwilson@ucdavis.edu
#   Repository: https://github.com/SWi1/polyphenol_pipeline/
#   License: MIT
#
# SHINY APP DEVELOPED BY:
#   Richard Stoker
#   United States Department of Agriculture - Agricultural Research Service
#   Contact: Richard.Stoker@usda.gov
#   License: CC0 (Public Domain)
#
# DESCRIPTION:
#   This file loads required packages, reference databases, and defines
#   helper functions used throughout the Shiny application.
#
# VERSION: 0.1 Alpha
# DATE: November 2025
#
# ==============================================================================

# Load required packages
suppressPackageStartupMessages({
  library(shiny)
  library(bslib)
  library(shinyWidgets)
  library(DT)
  library(plotly)
  library(tidyverse)
  library(readxl)
  library(vroom)
  library(rmarkdown)
  library(zip)
})

# ------------------------------------------------------------------------------
# Application Constants
# ------------------------------------------------------------------------------
APP_TITLE <- "Polyphenol Estimation Pipeline"
APP_VERSION <- "0.1 Alpha"
PRIMARY_COLOR <- "#6f42c1"
SECONDARY_COLOR <- "#5c2d91"

# ------------------------------------------------------------------------------
# File Paths - Self-contained within shiny_app directory
# ------------------------------------------------------------------------------
DATA_DIR <- "data"
DEMO_DIR <- "demo_data"

# Reference database paths
FDD_FILE <- file.path(DATA_DIR, "FDA_FDD_All_Records_v_3.1.xlsx")
FOODB_CONTENT_FILE <- file.path(DATA_DIR, "FooDB_polyphenol_content_with_dbPUPsubstrates_Aug25.csv")
CLASS_TAX_FILE <- file.path(DATA_DIR, "FooDB_polyphenol_list_3072.csv")
MAPPING_FILE <- file.path(DATA_DIR, "FDA_FooDB_Mapping_Nov_2025.csv")
EUGENOL_FILE <- file.path(DATA_DIR, "FooDB_Eugenol_Content_Final.csv")
DII_SUBCLASSES_FILE <- file.path(DATA_DIR, "FooDB_DII_polyphenol_list.csv")

# Demo data
DEMO_DATA_FILE <- file.path(DEMO_DIR, "VVKAJ_Items.csv")

# ------------------------------------------------------------------------------
# Load Reference Databases
# These files are loaded once at startup for performance
# ------------------------------------------------------------------------------
message("Loading reference databases...")

# FDA Food Disaggregation Database V3.1
if (file.exists(FDD_FILE)) {
  FDD_V3 <- read_xlsx(FDD_FILE) %>%
    rename(
      latest_survey = "Latest Survey",
      wweia_food_code = "WWEIA Food Code",
      wweia_food_description = "WWEIA Food Description",
      fdd_ingredient = "Basic Ingredient Description",
      ingredient_percent = "Ingredient Percent"
    ) %>%
    select(wweia_food_code, wweia_food_description, fdd_ingredient, ingredient_percent) %>%
    mutate(wweia_food_code = as.integer(wweia_food_code))
  message("  - FDA FDD database loaded")
} else {
  FDD_V3 <- NULL
  warning("FDA FDD database not found at: ", FDD_FILE)
}

# FooDB polyphenol content
if (file.exists(FOODB_CONTENT_FILE)) {
  FooDB_mg_100g <- vroom(FOODB_CONTENT_FILE, show_col_types = FALSE) %>%
    distinct(food_id, compound_public_id, .keep_all = TRUE) %>%
    select(-c(food_public_id, food_name)) %>%
    relocate(orig_content_avg, .before = citation) %>%
    filter(!is.na(orig_content_avg_RFadj))
  message("  - FooDB polyphenol content loaded")
} else {
  FooDB_mg_100g <- NULL
  warning("FooDB polyphenol content file not found")
}

# FooDB polyphenol class taxonomy
if (file.exists(CLASS_TAX_FILE)) {
  class_tax <- vroom(CLASS_TAX_FILE, show_col_types = FALSE) %>%
    select(c(compound_public_id, class))
  message("  - FooDB class taxonomy loaded")
} else {
  class_tax <- NULL
  warning("FooDB class taxonomy file not found")
}

# FDA-FooDB Mapping
if (file.exists(MAPPING_FILE)) {
  fdd_foodb_mapping <- vroom(MAPPING_FILE, show_col_types = FALSE) %>%
    select(-c(method, score))
  message("  - FDA-FooDB mapping loaded")
} else {
  fdd_foodb_mapping <- NULL
  warning("FDA-FooDB mapping file not found")
}

# FooDB Eugenol Content (for DII calculation)
if (file.exists(EUGENOL_FILE)) {
  FooDB_eugenol <- vroom(EUGENOL_FILE, show_col_types = FALSE)
  message("  - FooDB eugenol content loaded")
} else {
  FooDB_eugenol <- NULL
  warning("FooDB eugenol content file not found")
}

# FooDB DII polyphenol subclasses
if (file.exists(DII_SUBCLASSES_FILE)) {
  FooDB_DII_subclasses <- vroom(DII_SUBCLASSES_FILE, show_col_types = FALSE)
  message("  - FooDB DII subclasses loaded")
} else {
  FooDB_DII_subclasses <- NULL
  warning("FooDB DII subclasses file not found")
}

message("Database loading complete.")

# ------------------------------------------------------------------------------
# Helper Functions
# ------------------------------------------------------------------------------

# Check if all required databases are loaded
databases_ready <- function() {
  all(!is.null(FDD_V3),
      !is.null(FooDB_mg_100g),
      !is.null(class_tax),
      !is.null(fdd_foodb_mapping))
}

# Validate ASA24 dietary input data
validate_asa24_data <- function(data) {
  required_cols <- c("UserName", "FoodCode", "RecallNo", "FoodAmt")
  missing <- setdiff(required_cols, names(data))
  if (length(missing) > 0) {
    return(list(valid = FALSE, message = paste("Missing required columns:", paste(missing, collapse = ", "))))
  }

  recalls_per_user <- data %>%
    group_by(UserName) %>%
    summarise(n_recalls = n_distinct(RecallNo), .groups = "drop")

  if (max(recalls_per_user$n_recalls, na.rm = TRUE) < 2) {
    return(list(valid = FALSE, message = "Data must contain multiple recalls per participant."))
  }

  return(list(valid = TRUE, message = "Data validation passed."))
}

# Validate NHANES dietary input data
validate_nhanes_data <- function(data) {
  required_cols <- c("SEQN", "DRXIFDCD", "RecallNo")
  missing <- setdiff(required_cols, names(data))
  if (length(missing) > 0) {
    return(list(valid = FALSE, message = paste("Missing required columns:", paste(missing, collapse = ", "))))
  }
  return(list(valid = TRUE, message = "Data validation passed."))
}

# Apply coffee/tea brewing adjustment
# This adjustment accounts for the brewing process where polyphenols from
# coffee or tea are extracted into water
apply_brewing_adjustment <- function(fdd_data) {
  fdd_data %>%
    group_by(wweia_food_code) %>%
    mutate(
      has_tea = any(str_detect(fdd_ingredient, regex("Tea", ignore_case = TRUE))),
      has_coffee = any(str_detect(fdd_ingredient, regex("Coffee", ignore_case = TRUE))),
      has_water = any(str_detect(fdd_ingredient, regex("Water", ignore_case = TRUE))),
      brewing_adjustment_total = case_when(
        has_tea & has_water ~ sum(
          ingredient_percent[str_detect(fdd_ingredient, regex("Tea|Water", ignore_case = TRUE))],
          na.rm = TRUE),
        has_coffee & has_water ~ sum(
          ingredient_percent[str_detect(fdd_ingredient, regex("Coffee|Water", ignore_case = TRUE))],
          na.rm = TRUE),
        TRUE ~ NA_real_),
      brewing_adjustment_percentage = if_else(
        str_detect(fdd_ingredient, regex("Coffee|Tea", ignore_case = TRUE)),
        brewing_adjustment_total,
        NA_real_)
    ) %>%
    select(-c(has_tea, has_coffee, has_water, brewing_adjustment_total)) %>%
    ungroup()
}

# Color palette for visualizations
get_viz_colors <- function(n) {
  if (n <= 8) {
    return(c("#6f42c1", "#28a745", "#fd7e14", "#17a2b8", "#dc3545",
             "#ffc107", "#6610f2", "#e83e8c")[1:n])
  }
  colorRampPalette(c("#6f42c1", "#28a745", "#fd7e14", "#17a2b8"))(n)
}