import re
import os
import subprocess
from urllib.parse import urlparse

import numpy as np

from datalayers.datasources.tiff_layer import TiffLayer

class MalariaatlasTraveltimehc(TiffLayer):

    def __init__(self):
        super().__init__()
        self.precision = 3 # to small value for rounding
        self.format_suffix = "minutes"

    def download(self):
        urls = [
            "https://data.malariaatlas.org/geoserver/Accessibility/ows?service=CSW&version=2.0.1&request=DirectDownload&ResourceId=Accessibility:202001_Global_Motorized_Travel_Time_to_Healthcare"
        ]

        for url in urls:
            self._save_url_to_file(url)

            # Check if file is already unzipped
            a = urlparse(url)
            file_name = os.path.basename(a.path)

            if os.path.isfile(self.get_data_path() / "202001_Global_Motorized_Travel_Time_to_Healthcare_2019.tif"):
                self.logger.debug("File already unzipped.")
                return
            try:
                in_file = self.get_data_path() / file_name
                out_dir = self.get_data_path().as_posix()
                print(f'unzip {in_file} -d {out_dir}')
                subprocess.run(f'unzip {in_file} -d {out_dir}', shell=True,
                    capture_output=True, check=True)
            except subprocess.CalledProcessError as error:
                self.logger.warning("Could not unzip files: %s", error.stderr)

    def consume(self, file, band, shape):
        x = re.search(r'_([0-9]+)\.tif$', os.path.basename(file))
        year = int(x[1])

        self.rows.append({
            'year':          year,
            'shape_id':      shape.id,
            "value":         np.nanmean(band),
            "value_min":     np.nanmin(band),
            "value_max":     np.nanmax(band),
            "value_std_dev": np.nanstd(band)
        })