fun/plot_climat_obs_data.R

#' Plot CLIMAT Observational Data
#'
#' Creates 4 separate Plotly visualizations for observational climate data.
#' Returns a list of 4 plots that can be displayed in separate cards.
#'
#' @param df Data frame with columns: Date, MeanTemp, MaxTempAbs, MinTempAbs, Precipitation, SunshineDuration, etc.
#' @param station_name Title for the plot
#' @return A list of 4 plotly objects: temperature, precipitation, sunshine, pressure
#' @export
plot_climat_obs_data <- function(df, station_name = "") {
    library(plotly)
    library(dplyr)

    # --- Helper Functions (Local) ---

    ensure_temporal_continuity <- function(df, resolution = NULL) {
        if (is.null(df) || nrow(df) < 2) {
            return(df)
        }

        # Auto-detect resolution if not provided
        if (is.null(resolution)) {
            dt_diff <- as.numeric(difftime(df$Date[2], df$Date[1], units = "days"))
            if (!is.na(dt_diff)) {
                if (dt_diff > 25) {
                    resolution <- "monthly"
                } else if (dt_diff >= 0.9) {
                    resolution <- "daily"
                } else {
                    resolution <- "hourly"
                }
            } else {
                resolution <- "daily" # Fallback
            }
        }

        # Determine step
        if (resolution == "monthly") {
            step <- "1 month"
        } else if (resolution == "daily") {
            step <- "1 day"
        } else {
            step <- "1 hour"
        }

        full_time <- seq(from = min(df$Date), to = max(df$Date), by = step)
        complete_df <- data.frame(Date = full_time)
        df_out <- dplyr::left_join(complete_df, df, by = "Date")
        return(df_out)
    }

    split_into_chunks <- function(df, value_col) {
        if (!any(is.na(df[[value_col]]))) {
            return(list(df))
        }

        is_valid <- !is.na(df[[value_col]])
        r <- rle(is_valid)
        end_indices <- cumsum(r$lengths)
        start_indices <- c(1, head(end_indices, -1) + 1)

        chunks <- list()
        for (i in seq_along(r$values)) {
            if (r$values[i]) {
                chunks[[length(chunks) + 1]] <- df[start_indices[i]:end_indices[i], ]
            }
        }
        return(chunks)
    }

    empty_plot <- function(msg = "No data available") {
        plotly::plot_ly() %>%
            plotly::add_annotations(
                text = msg,
                showarrow = FALSE,
                xref = "paper", yref = "paper",
                x = 0.5, y = 0.5
            ) %>%
            plotly::layout(xaxis = list(visible = FALSE), yaxis = list(visible = FALSE))
    }

    if (is.null(df) || nrow(df) == 0) {
        return(list(
            temp = empty_plot(),
            precip = empty_plot(),
            sun = empty_plot(),
            pressure = empty_plot()
        ))
    }

    # Ensure Date is Date object
    df$Date <- as.Date(df$Date)

    # Ensure Continuity (Auto-detect)
    df <- ensure_temporal_continuity(df)

    # Calculate date range for display
    data_max <- max(df$Date, na.rm = TRUE)
    data_min <- min(df$Date, na.rm = TRUE)
    view_min <- data_max - (365 * 5)
    if (view_min < data_min) view_min <- data_min

    # Common x-axis configuration
    common_xaxis <- list(
        title = "",
        gridcolor = "rgba(200,200,200,0.3)",
        range = c(view_min, data_max),
        rangeselector = list(
            x = 1, y = 1.05, xanchor = "right",
            buttons = list(
                list(count = 1, label = "1y", step = "year", stepmode = "backward"),
                list(count = 5, label = "5y", step = "year", stepmode = "backward"),
                list(count = 10, label = "10y", step = "year", stepmode = "backward"),
                list(step = "all")
            )
        ),
        rangeslider = list(visible = FALSE),
        type = "date"
    )

    # --- Plot 1: Temperature ---
    has_temp <- "MeanTemp" %in% names(df) && any(!is.na(df$MeanTemp))
    p_temp <- plotly::plot_ly(type = "scatter", mode = "lines")

    if (has_temp) {
        # Absolute Max/Min (Filled Envelope)
        if ("MinTempAbs" %in% names(df) && "MaxTempAbs" %in% names(df)) {
            # We need valid pairs for envelope. If one is NA, we can't draw the range.
            # So we split based on one of them (assuming they usually go together).
            chunks <- split_into_chunks(df, "MaxTempAbs")

            for (i in seq_along(chunks)) {
                chunk <- chunks[[i]]
                if (nrow(chunk) < 2) next

                p_temp <- p_temp %>%
                    add_trace(
                        data = chunk,
                        x = ~Date, y = ~MinTempAbs, type = "scatter", mode = "lines",
                        line = list(color = "rgba(211, 47, 47, 0.1)", width = 1),
                        showlegend = FALSE, name = "Abs Min",
                        hovertemplate = "Abs Min: %{y:.1f}°C<extra></extra>",
                        connectgaps = FALSE,
                        legendgroup = "abs_range"
                    ) %>%
                    add_trace(
                        data = chunk,
                        x = ~Date, y = ~MaxTempAbs, type = "scatter", mode = "lines",
                        fill = "tonexty", fillcolor = "rgba(211, 47, 47, 0.1)",
                        line = list(color = "rgba(211, 47, 47, 0.1)", width = 1),
                        name = "Abs Max",
                        showlegend = (i == 1),
                        hovertemplate = "Abs Max: %{y:.1f}°C<extra></extra>",
                        connectgaps = FALSE,
                        legendgroup = "abs_range"
                    )
            }
        }

        # Mean Max/Min (Filled Envelope)
        if ("MeanMinTemp" %in% names(df) && "MeanMaxTemp" %in% names(df)) {
            chunks <- split_into_chunks(df, "MeanMaxTemp")

            for (i in seq_along(chunks)) {
                chunk <- chunks[[i]]
                if (nrow(chunk) < 2) next

                p_temp <- p_temp %>%
                    add_trace(
                        data = chunk,
                        x = ~Date, y = ~MeanMinTemp, type = "scatter", mode = "lines",
                        line = list(color = "rgba(211, 47, 47, 0.3)", width = 1),
                        showlegend = FALSE, name = "Mean Min",
                        hovertemplate = "Mean Min: %{y:.1f}°C<extra></extra>",
                        connectgaps = FALSE,
                        legendgroup = "mean_range"
                    ) %>%
                    add_trace(
                        data = chunk,
                        x = ~Date, y = ~MeanMaxTemp, type = "scatter", mode = "lines",
                        fill = "tonexty", fillcolor = "rgba(211, 47, 47, 0.3)",
                        line = list(color = "rgba(211, 47, 47, 0.3)", width = 1),
                        name = "Mean Max",
                        showlegend = (i == 1),
                        hovertemplate = "Mean Max: %{y:.1f}°C<extra></extra>",
                        connectgaps = FALSE,
                        legendgroup = "mean_range"
                    )
            }
        }

        # Main Mean Temp Line (Use full DF with NAs for simple line break)
        p_temp <- p_temp %>% add_lines(
            data = df,
            x = ~Date, y = ~MeanTemp, name = "Mean Temp",
            line = list(color = "#b71c1c", width = 2),
            hovertemplate = "Mean Temp: %{y:.1f}°C<extra></extra>",
            connectgaps = FALSE
        )
    } else {
        p_temp <- empty_plot("No temperature data")
    }

    p_temp <- p_temp %>%
        layout(
            title = list(text = "Temperature", font = list(size = 12), x = 0.05),
            yaxis = list(title = "Temp (°C)", gridcolor = "rgba(200,200,200,0.3)"),
            xaxis = common_xaxis,
            plot_bgcolor = "rgba(250,250,252,1)",
            hovermode = "x unified",
            legend = list(orientation = "h", x = 0.5, xanchor = "center", y = -0.15),
            margin = list(t = 50, r = 10, b = 60, l = 50)
        ) %>%
        config(displaylogo = FALSE)

    # --- Plot 2: Precipitation ---
    has_precip <- "Precipitation" %in% names(df) && any(!is.na(df$Precipitation))
    has_precip_days <- "PrecipDays" %in% names(df) && any(!is.na(df$PrecipDays))
    p_precip <- plotly::plot_ly()

    if (has_precip) {
        p_precip <- p_precip %>% add_bars(
            data = df,
            x = ~Date, y = ~Precipitation, name = "Precipitation",
            marker = list(color = "#0277bd"),
            hovertemplate = "Precip: %{y:.1f} mm<extra></extra>"
        )

        if (has_precip_days) {
            p_precip <- p_precip %>% add_trace(
                data = df,
                x = ~Date, y = ~PrecipDays, name = "Precip Days",
                yaxis = "y2", type = "scatter", mode = "markers",
                marker = list(color = "#388e3c", size = 5, line = list(color = "yellow", width = 1)),
                hovertemplate = "Days >1mm: %{y:.0f}<extra></extra>"
            )
        }
    } else {
        p_precip <- empty_plot("No precipitation data")
    }

    p_precip <- p_precip %>%
        layout(
            title = list(text = "Precipitation", font = list(size = 12), x = 0.05),
            yaxis = list(title = "Precip (mm)", rangemode = "tozero", gridcolor = "rgba(200,200,200,0.3)"),
            yaxis2 = list(title = "Days", overlaying = "y", side = "right", showgrid = FALSE, rangemode = "tozero"),
            xaxis = common_xaxis,
            plot_bgcolor = "rgba(250,250,252,1)",
            hovermode = "x unified",
            legend = list(orientation = "h", x = 0.5, xanchor = "center", y = -0.15),
            margin = list(t = 50, r = 50, b = 60, l = 50)
        ) %>%
        config(displaylogo = FALSE)

    # --- Plot 3: Sunshine Duration ---
    has_sun <- "SunshineDuration" %in% names(df) && any(!is.na(df$SunshineDuration))
    p_sun <- plotly::plot_ly()

    if (has_sun) {
        p_sun <- p_sun %>% add_bars(
            data = df,
            x = ~Date, y = ~SunshineDuration, name = "Sunshine",
            marker = list(color = "#fbc02d"),
            hovertemplate = "Sun: %{y:.1f} h<extra></extra>"
        )
    } else {
        p_sun <- empty_plot("No sunshine data")
    }

    p_sun <- p_sun %>%
        layout(
            title = list(text = "Sunshine Duration", font = list(size = 12), x = 0.05),
            yaxis = list(title = "Sun (h)", gridcolor = "rgba(200,200,200,0.3)"),
            xaxis = common_xaxis,
            plot_bgcolor = "rgba(250,250,252,1)",
            hovermode = "x unified",
            legend = list(orientation = "h", x = 0.5, xanchor = "center", y = -0.15),
            margin = list(t = 50, r = 10, b = 60, l = 50)
        ) %>%
        config(displaylogo = FALSE)

    # --- Plot 4: Pressure (MSLP + Vapour Pressure) ---
    has_vapour <- "VapourPressure" %in% names(df) && any(!is.na(df$VapourPressure))
    has_press <- "MeanSeaLevelPressure" %in% names(df) && any(!is.na(df$MeanSeaLevelPressure))
    p_pressure <- plotly::plot_ly()

    if (has_press) {
        p_pressure <- p_pressure %>%
            add_lines(
                data = df,
                x = ~Date, y = ~MeanSeaLevelPressure, name = "MSLP",
                line = list(color = "#7b1fa2", width = 2),
                hovertemplate = "MSLP: %{y:.1f} hPa<extra></extra>",
                connectgaps = FALSE
            ) %>%
            add_lines(
                x = ~Date, y = rep(1013.25, nrow(df)), name = "Std (1013.25)",
                line = list(color = "grey", dash = "dash", width = 1),
                hoverinfo = "skip",
                inherit = FALSE
            )
    }

    if (has_vapour) {
        p_pressure <- p_pressure %>% add_lines(
            data = df,
            x = ~Date, y = ~VapourPressure, name = "Vapour Pressure",
            yaxis = "y2", line = list(color = "#2e7d32", width = 2),
            hovertemplate = "VapPres: %{y:.1f} hPa<extra></extra>",
            connectgaps = FALSE
        )
    }

    if (!has_press && !has_vapour) {
        p_pressure <- empty_plot("No pressure data")
    }

    p_pressure <- p_pressure %>%
        layout(
            title = list(text = "Pressure", font = list(size = 12), x = 0.05),
            yaxis = list(title = "MSLP (hPa)", gridcolor = "rgba(200,200,200,0.3)"),
            yaxis2 = list(title = "VP (hPa)", overlaying = "y", side = "right", showgrid = FALSE),
            xaxis = common_xaxis,
            plot_bgcolor = "rgba(250,250,252,1)",
            hovermode = "x unified",
            legend = list(orientation = "h", x = 0.5, xanchor = "center", y = -0.15),
            margin = list(t = 50, r = 50, b = 60, l = 50)
        ) %>%
        config(displaylogo = FALSE)

    # Return list of 4 plots
    return(list(
        temp = p_temp,
        precip = p_precip,
        sun = p_sun,
        pressure = p_pressure
    ))
}