library(rvest)
library(dplyr)
library(stringr)
library(readr)
library(httr)

#' Scrape JMA Data (Daily, Hourly, 10-Minute, Monthly)
#'
#' Fetches weather data for a specific station, year, month, and optionally day.
#'
#' @param block_no Station Block Number (ID).
#' @param year Year (numeric).
#' @param month Month (numeric).
#' @param day Day (numeric, required for Hourly/10-Minute resolutions).
#' @param prec_no Prefecture Number (ID).
#' @param type Station type ("s1" or "a1").
#' @param resolution Resolution ("Daily", "Hourly", "10 Minutes", "Monthly")
#'
#' @return A data frame containing the data, or NULL if failed.
get_jma_data <- function(block_no, year, month, day = NULL, prec_no, type = "s1", resolution = "Daily") {
  # Map resolution to URL part
  res_code <- "daily"
  if (resolution == "Hourly") res_code <- "hourly"
  if (resolution == "10 Minutes" || resolution == "10-Minute") res_code <- "10min"
  if (resolution == "Monthly") res_code <- "monthly"

  # URL construction - include day for hourly/10min
  if (resolution %in% c("Daily", "Monthly")) {
    url <- sprintf(
      "https://www.data.jma.go.jp/obd/stats/etrn/view/%s_%s.php?prec_no=%s&block_no=%s&year=%d&month=%d&day=&view=",
      res_code, type, prec_no, block_no, year, month
    )
  } else {
    # Hourly and 10-minute require day parameter
    if (is.null(day)) {
      warning("Hourly/10-Minute resolution requires a day parameter")
      return(NULL)
    }
    url <- sprintf(
      "https://www.data.jma.go.jp/obd/stats/etrn/view/%s_%s.php?prec_no=%s&block_no=%s&year=%d&month=%d&day=%d&view=",
      res_code, type, prec_no, block_no, year, month, day
    )
  }

  message(sprintf("Downloading %s data from: %s", resolution, url))

  message(sprintf("Downloading %s data from: %s", resolution, url))
  
  # Use httr::GET with timeout to prevent hanging
  page <- tryCatch(
    {
      resp <- httr::GET(url, httr::timeout(10))
      if (httr::status_code(resp) != 200) {
        warning(sprintf("Failed to download data: HTTP %s", httr::status_code(resp)))
        return(NULL)
      }
      read_html(resp)
    },
    error = function(e) {
      warning("Failed to download data: ", e)
      return(NULL)
    }
  )

  if (is.null(page)) {
    return(NULL)
  }

  # Find the weather data table (monthly tables are shorter)
  tables <- html_table(page, fill = TRUE, header = FALSE)
  weather_table <- NULL
  min_rows <- ifelse(resolution == "Monthly", 10, 20)
  for (t in tables) {
    if (nrow(t) > min_rows && ncol(t) > 3) {
      weather_table <- t
      break
    }
  }

  if (is.null(weather_table)) {
    return(NULL)
  }
  if (nrow(weather_table) <= 3) {
    return(NULL)
  }

  if (nrow(weather_table) <= 3) {
    return(NULL)
  }

  # Column Selection based on resolution and station type
  cols <- NULL
  col_names <- NULL
  time_col_name <- "Day" # Default
  df <- NULL

  # Skip header rows (row 1 is main header, row 2 might be sub-header/units)
  # For simple parsing, we rely on Row 1 for keywords.
  if (resolution == "Monthly") {
      # Monthly tables might have 1 or 2 header rows. Row 1 usually contains the variable names.
      data_rows <- weather_table[2:nrow(weather_table), ]
  } else {
      data_rows <- weather_table[2:nrow(weather_table), ]
  }

    # Dynamic Column Parsing based on Header Names
    # Header is always in row 1 for these tables, but Row 2 has sub-headers (Mean/Max/Min) for Monthly
    header_row <- as.character(weather_table[1, ])
    header_row_2 <- if (nrow(weather_table) >= 2) as.character(weather_table[2, ]) else rep("", length(header_row))
    
    # Define mappings (Keyword -> English Name)
    col_map <- list()
    col_indices <- c()
    col_final_names <- c()
    
    # helper to find index matching both row 1 (main) and row 2 (sub) pattern
    find_col <- function(p1, p2 = NULL) {
        i1 <- grep(p1, header_row)
        if (length(i1) == 0) return(NULL)
        
        if (is.null(p2)) return(i1[1])
        
        # Check sub-headers at these indices
        matches <- i1[grep(p2, header_row_2[i1])]
        if (length(matches) > 0) return(matches[1])
        return(NULL)
    }

    # 1. Time / Day / Month
    if (resolution == "Monthly") {
      idx <- grep("月", header_row)
      if (length(idx) > 0) { col_indices <- c(col_indices, idx[1]); col_final_names <- c(col_final_names, "Month") }
    } else if (resolution == "Daily") {
      idx <- grep("^日", header_row)
      if (length(idx) > 0) { col_indices <- c(col_indices, idx[1]); col_final_names <- c(col_final_names, "Day") }
    } else if (resolution == "Hourly") {
      idx <- grep("^時", header_row)
      if (length(idx) > 0) { col_indices <- c(col_indices, idx[1]); col_final_names <- c(col_final_names, "Hour") }
    } else {
      idx <- grep("^時", header_row)
      if (length(idx) > 0) { col_indices <- c(col_indices, idx[1]); col_final_names <- c(col_final_names, "Time") }
    }
    
    # 2. Pressure
    if (resolution == "Monthly") {
        idx_sta <- find_col("気圧", "現地|平均")
        if (is.null(idx_sta)) {
             match <- grep("気圧", header_row)
             if (length(match) > 0) idx_sta <- match[1]
        }
        idx_sea <- find_col("気圧", "海面")
        if (is.null(idx_sea) && length(grep("気圧", header_row)) > 1) {
             match <- grep("気圧", header_row)
             idx_sea <- match[2]
        }
        if (!is.null(idx_sta)) { col_indices <- c(col_indices, idx_sta); col_final_names <- c(col_final_names, "Pressure") }
        if (!is.null(idx_sea)) { col_indices <- c(col_indices, idx_sea); col_final_names <- c(col_final_names, "Pressure_Sea_Level") }
        
        # Precip
        idx_pr <- find_col("降水量", "合計")
        if (is.null(idx_pr)) idx_pr <- grep("降水量", header_row)[1]
        if (!is.null(idx_pr)) { col_indices <- c(col_indices, idx_pr); col_final_names <- c(col_final_names, "Precipitation") }
        
        # Temp
        idx_tm <- find_col("気温", "平均")
        if (is.null(idx_tm)) idx_tm <- grep("気温", header_row)[1]
        idx_th <- find_col("気温", "最高")
        idx_tl <- find_col("気温", "最低")
        if (!is.null(idx_tm)) { col_indices <- c(col_indices, idx_tm); col_final_names <- c(col_final_names, "Temp_Mean") }
        if (!is.null(idx_th)) { col_indices <- c(col_indices, idx_th); col_final_names <- c(col_final_names, "Temp_Max") }
        if (!is.null(idx_tl)) { col_indices <- c(col_indices, idx_tl); col_final_names <- c(col_final_names, "Temp_Min") }
        
        # Humidity
        idx_hm <- find_col("湿度", "平均")
        if (is.null(idx_hm)) idx_hm <- grep("湿度", header_row)[1]
        idx_hl <- find_col("湿度", "最小")
        if (!is.null(idx_hm)) { col_indices <- c(col_indices, idx_hm); col_final_names <- c(col_final_names, "Humidity") }
        if (!is.null(idx_hl)) { col_indices <- c(col_indices, idx_hl); col_final_names <- c(col_final_names, "Humidity_Min") }
        
        # Wind
        idx_ws <- find_col("風速", "平均風速|平均") 
        if (is.null(idx_ws)) idx_ws <- grep("風速", header_row)[1]
        idx_ws_max <- find_col("風速", "最大風速")
        if (!is.null(idx_ws)) { col_indices <- c(col_indices, idx_ws); col_final_names <- c(col_final_names, "Wind_Speed") }
        if (!is.null(idx_ws_max)) { col_indices <- c(col_indices, idx_ws_max); col_final_names <- c(col_final_names, "Wind_Speed_Max") }
        
    } else if (resolution == "Daily") {
        # Pressure
        id_p_sta <- find_col("気圧", "現地")
        if (is.null(id_p_sta)) id_p_sta <- grep("気圧", header_row)[1]
        id_p_sea <- find_col("気圧", "海面")
        if (is.null(id_p_sea) && length(grep("気圧", header_row)) > 1) {
             # assumption: 2nd pressure col is Sea Level
             id_p_sea <- grep("気圧", header_row)[2]
        }
        if (!is.null(id_p_sta)) { col_indices <- c(col_indices, id_p_sta); col_final_names <- c(col_final_names, "Pressure") }
        if (!is.null(id_p_sea)) { col_indices <- c(col_indices, id_p_sea); col_final_names <- c(col_final_names, "Pressure_Sea_Level") }

        # Precip
        id_pr_tot <- find_col("降水量", "合計")
        if (is.null(id_pr_tot)) id_pr_tot <- grep("降水量", header_row)[1]
        id_pr_1h <- find_col("降水量", "最大1時間")
        if (is.null(id_pr_1h) && length(grep("降水量", header_row)) >= 2) id_pr_1h <- grep("降水量", header_row)[2]
        id_pr_10m <- find_col("降水量", "最大10分")
        if (is.null(id_pr_10m) && length(grep("降水量", header_row)) >= 3) id_pr_10m <- grep("降水量", header_row)[3]
        
        if (!is.null(id_pr_tot)) { col_indices <- c(col_indices, id_pr_tot); col_final_names <- c(col_final_names, "Precipitation") }
        if (!is.null(id_pr_1h)) { col_indices <- c(col_indices, id_pr_1h); col_final_names <- c(col_final_names, "Precipitation_Max_1h") }
        if (!is.null(id_pr_10m)) { col_indices <- c(col_indices, id_pr_10m); col_final_names <- c(col_final_names, "Precipitation_Max_10min") }

        # Temp
        # Positional Fallback: 1=Mean, 2=Max, 3=Min
        temp_cols <- grep("気温", header_row)
        if (length(temp_cols) > 0) {
             # Try explicit first
             id_tm <- find_col("気温", "平均")
             id_th <- find_col("気温", "最高")
             id_tl <- find_col("気温", "最低")
             
             # If explicit fails but we have 3 columns, assume Mean/Max/Min
             if (is.null(id_tm) && length(temp_cols) >= 1) id_tm <- temp_cols[1]
             if (is.null(id_th) && length(temp_cols) >= 2) id_th <- temp_cols[2]
             if (is.null(id_tl) && length(temp_cols) >= 3) id_tl <- temp_cols[3]
             
             if (!is.null(id_tm)) { col_indices <- c(col_indices, id_tm); col_final_names <- c(col_final_names, "Temp_Mean") }
             if (!is.null(id_th)) { col_indices <- c(col_indices, id_th); col_final_names <- c(col_final_names, "Temp_Max") }
             if (!is.null(id_tl)) { col_indices <- c(col_indices, id_tl); col_final_names <- c(col_final_names, "Temp_Min") }
        }
        
        # Humidity
        # Positional: 1=Mean, 2=Min
        hum_cols <- grep("湿度", header_row)
        if (length(hum_cols) > 0) {
            id_hm <- find_col("湿度", "平均")
            id_hl <- find_col("湿度", "最小")
            
            if (is.null(id_hm) && length(hum_cols) >= 1) id_hm <- hum_cols[1]
            if (is.null(id_hl) && length(hum_cols) >= 2) id_hl <- hum_cols[2]
            
            if (!is.null(id_hm)) { col_indices <- c(col_indices, id_hm); col_final_names <- c(col_final_names, "Humidity") }
            if (!is.null(id_hl)) { col_indices <- c(col_indices, id_hl); col_final_names <- c(col_final_names, "Humidity_Min") }
        }
        
        # Wind
        # Positional: 1=Mean, 2=Max, 3=MaxDir, 4=Gust, 5=GustDir
        wind_cols <- grep("風速", header_row)
        if (length(wind_cols) > 0) {
             id_ws <- find_col("風速", "平均風速")
             if (is.null(id_ws) && length(wind_cols) >= 1) id_ws <- wind_cols[1]
             
             id_ws_max <- find_col("風速", "最大風速")
             if (is.null(id_ws_max) && length(wind_cols) >= 2) id_ws_max <- wind_cols[2]
             
             id_ws_gust <- find_col("風速", "最大瞬間")
             if (is.null(id_ws_gust) && length(wind_cols) >= 4) id_ws_gust <- wind_cols[4]
             
             if (!is.null(id_ws)) { col_indices <- c(col_indices, id_ws); col_final_names <- c(col_final_names, "Wind_Speed") }
             if (!is.null(id_ws_max)) { col_indices <- c(col_indices, id_ws_max); col_final_names <- c(col_final_names, "Wind_Max_Speed") }
             if (!is.null(id_ws_gust)) { col_indices <- c(col_indices, id_ws_gust); col_final_names <- c(col_final_names, "Wind_Gust_Speed") }
        }
        
    } else {
        # Hourly/10-min
        
        # Pressure
        # Hourly usually has unique columns for Station and Sea Level
        idx_sta <- find_col("気圧", "現地")
        if (is.null(idx_sta)) {
             # fallback for simple hourly tables
             match <- grep("気圧", header_row)
             # If "気圧" but no subheader (unlikely for specific), or multiple
             if (length(match) > 0) idx_sta <- match[1]
        }
        
        idx_sea <- find_col("気圧", "海面")
        if (is.null(idx_sea) && length(grep("気圧", header_row)) > 1) {
             match <- grep("気圧", header_row)
             idx_sea <- match[2]
        }
        
        if (!is.null(idx_sta)) { col_indices <- c(col_indices, idx_sta); col_final_names <- c(col_final_names, "Pressure") }
        if (!is.null(idx_sea)) { col_indices <- c(col_indices, idx_sea); col_final_names <- c(col_final_names, "Pressure_Sea_Level") }
        
        # Precip
        idx_pr <- grep("降水量", header_row)
        if (length(idx_pr) > 0) { col_indices <- c(col_indices, idx_pr[1]); col_final_names <- c(col_final_names, "Precipitation") }
        
        # Temp
        idx_tm <- grep("気温", header_row)
        if (length(idx_tm) > 0) { col_indices <- c(col_indices, idx_tm[1]); col_final_names <- c(col_final_names, "Temperature") }
        
        # Humidity
        idx_rh <- grep("湿度", header_row)
        if (length(idx_rh) > 0) { col_indices <- c(col_indices, idx_rh[1]); col_final_names <- c(col_final_names, "Humidity") }

        # Wind Speed
        # Usually Row 2 has "風速" (Mean Speed)
        idx <- find_col("風速", "風速")
        # If Row 2 fails (some tables simple), fall back to Row 1 strict
        if (is.null(idx)) idx <- grep("風速", header_row)[1]
        
        if (!is.null(idx)) {
             col_indices <- c(col_indices, idx)
             col_final_names <- c(col_final_names, "Wind_Speed")
        }
        
        # Wind Direction
        # Row 2 explicitly "風向"
        idx_dir <- find_col("風向", "風向")
        # If Row 2 fails, try unique Row 1 match? 
        # But commonly Row 1 is "Wind Speed/Direction" merged.
        # Fallback: if header_row contains "風向" but NOT "風速" at that index?
        if (is.null(idx_dir)) {
             matches <- grep("風向", header_row)
             if (length(matches) > 0) idx_dir <- matches[length(matches)] # Take the *last* one if ambiguous? Unsafe.
        }
        
        if (!is.null(idx_dir)) {
             col_indices <- c(col_indices, idx_dir)
             col_final_names <- c(col_final_names, "Wind_Direction")
        }
    }

    # 7. Sunshine
    idx <- grep("日照", header_row)
    if (length(idx) > 0) { 
        col_indices <- c(col_indices, idx[1])
        suffix <- if (resolution == "10 Minutes" || resolution == "10-Minute") "_Minutes" else "_Hours"
        col_final_names <- c(col_final_names, paste0("Sunshine", suffix))
    }
    
    # 8. Snow
    if (resolution == "Monthly") {
         idx_fall <- find_col("雪", "降雪") 
         idx_depth <- find_col("雪", "最深積雪")
         idx_days <- grep("雪日数", header_row_2) # Snow days is usually separate or under Atm Phenomena
         
         if (!is.null(idx_fall)) { col_indices <- c(col_indices, idx_fall); col_final_names <- c(col_final_names, "Snowfall") }
         if (!is.null(idx_depth)) { col_indices <- c(col_indices, idx_depth); col_final_names <- c(col_final_names, "Snow_Depth") }
         if (length(idx_days) > 0) { col_indices <- c(col_indices, idx_days[1]); col_final_names <- c(col_final_names, "Snow_Days") }
         
         # Other Days
         idx_fog <- grep("霧日数", header_row_2)
         if (length(idx_fog) > 0) { col_indices <- c(col_indices, idx_fog[1]); col_final_names <- c(col_final_names, "Fog_Days") }
         
         idx_thunder <- grep("雷日数", header_row_2)
         if (length(idx_thunder) > 0) { col_indices <- c(col_indices, idx_thunder[1]); col_final_names <- c(col_final_names, "Thunder_Days") }
         
    } else {
         # Daily/Hourly/10-min
         # Hourly might have "雪" in Row 1, and "降雪"/"積雪" in Row 2
         idx_fall <- find_col("雪|降雪", "降雪")
         if (is.null(idx_fall)) idx_fall <- grep("降雪", header_row)[1] # Fallback for Daily simple
         
         if (!is.null(idx_fall) && !is.na(idx_fall)) { 
             col_indices <- c(col_indices, idx_fall)
             col_final_names <- c(col_final_names, "Snowfall") 
         }
         
         idx_depth <- find_col("雪|積雪", "積雪")
         if (is.null(idx_depth)) idx_depth <- grep("積雪", header_row)[1]
         
         if (!is.null(idx_depth) && !is.na(idx_depth)) { 
             col_indices <- c(col_indices, idx_depth) 
             col_final_names <- c(col_final_names, "Snow_Depth") 
         }
    }


    # 9. Additional Parameters (Dew Point, Vapor, Solar, Cloud, Visibility)
    # Usually only available in Hourly/10-min or specialized Daily tables
    
    # Dew Point
    idx <- grep("露点温度", header_row)
    if (length(idx) > 0) {
         col_indices <- c(col_indices, idx[1])
         col_final_names <- c(col_final_names, "Dew_Point")
    }

    # Vapor Pressure
    idx <- grep("蒸気圧", header_row)
    if (length(idx) > 0) {
         col_indices <- c(col_indices, idx[1])
         col_final_names <- c(col_final_names, "Vapor_Pressure")
    }

    # Global Solar Radiation
    idx <- grep("全天日射", header_row)
    if (length(idx) > 0) {
         col_indices <- c(col_indices, idx[1])
         col_final_names <- c(col_final_names, "Solar_Radiation")
    }

    # Cloud Cover
    idx <- grep("雲量", header_row)
    if (length(idx) > 0) {
         col_indices <- c(col_indices, idx[1])
         col_final_names <- c(col_final_names, "Cloud_Cover")
    }

    # Visibility
    idx <- grep("視程", header_row)
    if (length(idx) > 0) {
         col_indices <- c(col_indices, idx[1])
         col_final_names <- c(col_final_names, "Visibility")
    }

    # Extract Data (Skip row 1 header)
    data_rows <- weather_table[2:nrow(weather_table), ]
    
    # Deduplicate indices (just in case)
    # Keep strictly unique indices to avoid column duplication errors
    if (length(col_indices) > 0) {
        # Check for NAs matching
        valid_mask <- !is.na(col_indices)
        col_indices <- col_indices[valid_mask]
        col_final_names <- col_final_names[valid_mask]
        
        # Deduplication based on index
        # We need to keep the FIRST occurrence or iterate
        # actually, simply checking duplications:
        dupe_mask <- !duplicated(col_indices)
        col_indices <- col_indices[dupe_mask]
        col_final_names <- col_final_names[dupe_mask]
    }
    
    if (length(col_indices) == 0) return(NULL)
    
    df <- data_rows[, col_indices, drop = FALSE]
    colnames(df) <- col_final_names

  suppressWarnings({
    clean_numeric <- function(x) {
      if (all(is.na(x))) {
        return(x)
      }
      x <- as.character(x)
      x <- str_remove_all(x, "[\\]\\)\\\\\\u00A0]")
      x[x == "///" | x == "--" | x == "" | x == "×" | x == "×"] <- NA
      as.numeric(x)
    }

    # For Monthly, add Year/Month and clean
    if (resolution == "Monthly") {
      df$Month <- clean_numeric(df$Month)
      numeric_cols <- setdiff(names(df), "Month")
      for (col in numeric_cols) {
        df[[col]] <- clean_numeric(df[[col]])
      }
      
      # Filter out non-numeric months (e.g. sub-headers) BEFORE creating Date
      # Also filter to the specific requested month to avoid returning the whole year
      # (since get_jma_range_data calls this iteratively for each month)
      df <- df %>%
        filter(!is.na(Month)) %>%
        filter(Month == month) %>%
        mutate(
          Year = year,
          Date = as.Date(sprintf("%04d-%02d-01", Year, Month))
        ) %>%
        select(Year, Month, everything())
    } else if (resolution == "Daily") {
      df <- df %>%
        mutate(
          Year = year,
          Month = month,
          across(everything(), clean_numeric)
        ) %>%
        select(Year, Month, everything()) %>%
        filter(!is.na(Day))
    } else {
      # For Hourly/10-min, add Year/Month/Day
      df <- df %>%
        mutate(
          Year = year,
          Month = month,
          Day = day
        )

      # Clean numeric columns (except Time/Hour/Date AND Wind_Direction)
      numeric_cols <- setdiff(names(df), c("Time", "Hour", "Year", "Month", "Day", "Wind_Direction"))
      for (col in numeric_cols) {
        df[[col]] <- clean_numeric(df[[col]])
      }
      
      # Clean Wind_Direction specifically (keep as character)
      if ("Wind_Direction" %in% names(df)) {
          df$Wind_Direction <- as.character(df$Wind_Direction)
          df$Wind_Direction <- str_remove_all(df$Wind_Direction, "[\\]\\)\\\\\\u00A0]")
          df$Wind_Direction[df$Wind_Direction == "///" | df$Wind_Direction == "--" | df$Wind_Direction == "" | df$Wind_Direction == "×"] <- NA
          
          # Convert to Degrees
          wind_dir_map <- c(
            "北" = 360, "N" = 360,
            "北北東" = 22.5, "NNE" = 22.5,
            "北東" = 45, "NE" = 45,
            "東北東" = 67.5, "ENE" = 67.5,
            "東" = 90, "E" = 90,
            "東南東" = 112.5, "ESE" = 112.5,
            "南東" = 135, "SE" = 135,
            "南南東" = 157.5, "SSE" = 157.5,
            "南" = 180, "S" = 180,
            "南南西" = 202.5, "SSW" = 202.5,
            "南西" = 225, "SW" = 225,
            "西南西" = 247.5, "WSW" = 247.5,
            "西" = 270, "W" = 270,
            "西北西" = 292.5, "WNW" = 292.5,
            "北西" = 315, "NW" = 315,
            "北北西" = 337.5, "NNW" = 337.5,
            "静穏" = NA, "Calm" = NA
          )
          
          df$Wind_Direction_Deg <- wind_dir_map[df$Wind_Direction]
      }

      # For Hour column, extract just the number
      if ("Hour" %in% names(df)) {
        # Filter out sub-header rows (e.g. "時")
        df <- df %>% filter(Hour != "時")
        df$Hour <- as.numeric(as.character(df$Hour))
        df <- df %>% filter(!is.na(Hour))
      }
      
      # For Time column (10-min), filter standard sub-headers
      if ("Time" %in% names(df)) {
          df <- df %>% filter(Time != "時分")
      }

      # Reorder columns
      df <- df %>% select(Year, Month, Day, everything())
    }
  })

  return(df)
}


#' Scrape JMA data for a Date Range (Daily/Monthly by month, Hourly/10-min by day)
#'
#' @param block_no Station ID
#' @param start_date Date object
#' @param end_date Date object
#' @param prec_no Prefecture ID
#' @param type Station Type (s1 or a1)
#' @param resolution Resolution string ("Daily", "Hourly", "10 Minutes", "Monthly")
get_jma_range_data <- function(block_no, start_date, end_date, prec_no, type = "s1", resolution = "Daily") {
  results <- list()

  if (resolution %in% c("Daily", "Monthly")) {
    # For Daily/Monthly: iterate months
    dates <- seq(as.Date(format(start_date, "%Y-%m-01")),
      as.Date(format(end_date, "%Y-%m-01")),
      by = "month"
    )

    for (d in dates) {
      yr <- as.numeric(format(as.Date(d, origin = "1970-01-01"), "%Y"))
      mo <- as.numeric(format(as.Date(d, origin = "1970-01-01"), "%m"))

      df <- get_jma_data(block_no, yr, mo, day = NULL, prec_no, type, resolution)

      if (!is.null(df)) {
        if ("Date" %in% names(df)) {
          df_filtered <- df %>%
            filter(Date >= start_date & Date <= end_date) %>%
            select(-Date)
        } else if ("Day" %in% names(df)) {
          df_filtered <- df %>%
            mutate(Date = as.Date(sprintf("%04d-%02d-%02d", Year, Month, Day))) %>%
            filter(Date >= start_date & Date <= end_date) %>%
            select(-Date)
        } else {
          df_filtered <- df
        }
        results[[paste(yr, mo, sep = "_")]] <- df_filtered
      }
      Sys.sleep(0.1)
    }
  } else {
    # For Hourly/10-min: iterate each day in the range
    all_days <- seq(start_date, end_date, by = "day")

    for (d in all_days) {
      d_date <- as.Date(d, origin = "1970-01-01")
      yr <- as.numeric(format(d_date, "%Y"))
      mo <- as.numeric(format(d_date, "%m"))
      dy <- as.numeric(format(d_date, "%d"))

      df <- get_jma_data(block_no, yr, mo, day = dy, prec_no, type, resolution)

      if (!is.null(df)) {
        results[[paste(yr, mo, dy, sep = "_")]] <- df
      }
      Sys.sleep(0.1)
    }
  }

  if (length(results) == 0) {
    return(NULL)
  }
  bind_rows(results)
}