data.py

import numpy as np
import netCDF4 as nc
import requests
import os

def download(url=None, cached_etag=None):
    try:
        if url is None:
            base = b'https://www.ncei.noaa.gov/data/noaa-global-surface-temperature/v5/access/gridded/'
            resp = requests.get(base)
            start = resp.content.find(b'href="NOAAGlobalTemp')
            end = resp.content.find(b'>', start)
            fname = resp.content[start+6:end-1]
            url = base+fname
        if not os.path.exists(fname):
            resp = requests.head(url, allow_redirects=True)
            if resp.status_code == 200:
                headers = resp.headers
                etag = headers['ETag']
                if (cached_etag is None) or etag != cached_etag:
                    local_filename = url.split(b'/')[-1]
                    r = requests.get(url, stream=True, allow_redirects=True)
                    if r.status_code == 200:
                        with open(local_filename, 'wb') as f:
                            for chunk in r.iter_content(chunk_size=8192):
                                f.write(chunk)
        return fname.decode('utf-8')
    except Exception as e:
        pass

def noaa_temperatures(fname, yearly=True, start=1971, 
                      span=30, stat='mean', grid_size=5):
    # Reads NetCDF file
    ds = nc.Dataset(fname)
    # Anomaly data (z)
    anom = ds['anom'][:].data
    # Value used to replace missing values
    nanval = ds['anom'].missing_value
    # Original shape is (n_years, 1, 36, 72)
    gridded = anom.squeeze()
    gridded[gridded == nanval] = np.nan
    stat = 'mean'
    funcs = {'mean': np.mean, 'max': np.max, 'min': np.min}
    # Gets the number of full years in the file
    n_years = gridded.shape[0]//12

    end = start + span - 1
    if (start != 1971) or (end != 2000):
        polygons = surface_polygons(grid_size)
        surface_perc = surface_area(polygons, perc=True)
        # Computes weighted average using surface area of grid boxes
        adjustment = np.mean(surface_stat(gridded[(start-1880)*12:(end-1880+1)*12], 
                                          surface_perc))
        # Adjusts baseline
        gridded = gridded - adjustment

    # Computes the average over each year
    griddedy = np.array([funcs[stat](gridded[i*12:(i+1)*12], axis=0) 
                         for i in range(n_years)])
    # NetCDF has latitudes from -87.5 to 87.5 (it uses the centers of the 
    # grid boxes) but we need to flip it upside down to use in Plotly
    griddedy = griddedy[:, ::-1]
    # NetCDF has longitudes from 0 to 360, but we need to rearrange it from 
    # -180 to 180 to use in Plotly, so we concatenate the last 36 columns 
    # (180 to 360 = -180 to 0) to the first 36 columns (0 to 180)
    griddedy = np.concatenate([griddedy[:, :, 36:], griddedy[:, :, :36]], axis=2)
    return griddedy if yearly else gridded