# Required packages
# library(ggplot2)
# library(dplyr)
raster_files_NEW <- list.files(merged_final,full.names = T, pattern = ".tif")

# Extracting the dates from vrt_list (assuming the format "YYYY-MM-DD.vrt" at the end)
no_cloud_dates <- as.Date(sapply(raster_files_NEW, function(x) {
  sub(".*/([0-9]{4}-[0-9]{2}-[0-9]{2})\\.tif", "\\1", x)
}))

# Generate a full sequence of dates in the range
start_date <- min(no_cloud_dates)
end_date <- max(no_cloud_dates)
all_dates <- seq(start_date, end_date, by = "day")

# Create a data frame marking no clouds (1) and clouds (0)
cloud_data <- data.frame(
  date = all_dates,
  cloud_status = ifelse(all_dates %in% no_cloud_dates, 1, 0)
)

# Plot the data
ggplot(cloud_data, aes(x = date, y = cloud_status)) +
  geom_point() +
  labs(x = "Date", y = "Cloud Status (1 = No Cloud, 0 = Cloud)") +
  scale_y_continuous(breaks = c(0, 1)) +
  theme_minimal()

# Updated ggplot code to display months on the x-axis
ggplot(cloud_data, aes(x = date, y = cloud_status)) +
  geom_point() +
  scale_x_date(date_labels = "%b", date_breaks = "1 month") +
  labs(x = "Month", y = "Cloud Status (1 = No Cloud, 0 = Cloud)") +
  scale_y_continuous(breaks = c(0, 1)) +
  theme_minimal()

# Group data by year and week
cloud_data <- cloud_data %>%
  mutate(week = isoweek(date), year = year(date)) %>%
  group_by(year, week) %>%
  summarise(no_cloud_days = sum(cloud_status == 1),
            cloud_days = sum(cloud_status == 0))

# 1. Show how many weeks per year have no images (clouds for all 7 days)
weeks_no_images <- cloud_data %>%
  filter(cloud_days == 7)

# Plot weeks with no images
ggplot(weeks_no_images, aes(x = week, y = year)) +
  geom_tile(fill = "red") +
  labs(x = "Week", y = "Year", title = "Weeks with No Images (Full Cloud Cover)") +
  theme_minimal()


# 2. Determine when most clouds are present (cloud_days > no_cloud_days)
weeks_most_clouds <- cloud_data %>%
  filter(cloud_days > no_cloud_days)

# Plot when most clouds are present
ggplot(weeks_most_clouds, aes(x = week, y = year)) +
  geom_tile(fill = "blue") +
  labs(x = "Week", y = "Year", title = "Weeks with Most Clouds") +
  theme_minimal()

# Group weeks by number of cloud days and count how many weeks had 0-7 cloud days
weeks_by_cloud_days <- cloud_data %>%
  group_by(cloud_days) %>%
  summarise(week_count = n())

# Display the summary
print(weeks_by_cloud_days)

# Optional: Plot the results to visualise how many weeks had 0-7 cloud days
ggplot(weeks_by_cloud_days, aes(x = cloud_days, y = week_count)) +
  geom_bar(stat = "identity", fill = "skyblue") +
  labs(x = "Number of Cloud Days (per week)", y = "Number of Weeks",
       title = "Distribution of Cloud Days per Week") +
  theme_minimal()

weeks_by_no_cloud_days <- cloud_data %>%
  mutate(no_cloud_days = 7 - cloud_days) %>%
  group_by(no_cloud_days) %>%
  summarise(week_count = n())

# Plot the results to visualise how many weeks had 0-7 cloud-free days
ggplot(weeks_by_no_cloud_days, aes(x = no_cloud_days, y = week_count)) +
  geom_bar(stat = "identity", fill = "#00A799") +
  geom_text(aes(label = week_count), vjust = -0.5, size = 4) +  # Add the count of weeks on top of bars
  labs(x = "Number of Cloud-Free Days (per week)", y = "Number of Weeks",
       title = "Distribution of Cloud-Free Days per Week") +
  theme_minimal()