wcps_rasdaman.py

def wcps_rasdaman(query, ip='saocompute.eurac.edu', file_name='', rtype = 'ProcessCoverages'):
    """
    Sends a WCPS query to a Rasdaman server and wraps the response for further use in Python depending on the 
    the response format chosen in the query.
    
    Args: 
        query (str) -- WCPS query you want to send to the Rasdaman server
        ip (str) -- IP of Rasdaman server (default saocompute.eurac.edu)
        rtype -- get Meta data or coverage
    Returns:
        Either one of the following
        - Numpy array for JSON/CSV formatted response
        - Xarray Dataset for a netCDF formatted response
        - Filepath to a TIFF/JPEG/JP2/PNG file saved to disk, in respect to the response image type
        - The response object, when the response could not be processed
    
    Sources:
        http://xarray.pydata.org/en/stable/io.html#netcdf
        http://xarray.pydata.org/en/stable/data-structures.html
        
    Author: Harald Kristen, Alexander Jacob
    Date: 2020-04-15
    """

    import requests
    import json
    import werkzeug
    import numpy as np
    import io
    import xarray as xr
    import uuid
    import os
    import xml.etree.ElementTree as ET
    import base64
    import urllib.request as request
    import xmltodict
    
    #set the work directory
    work_directory = ''#os.getcwd()
    
    #print('WCPS init')
    
    if ip == 'saocompute.eurac.edu/sincohmap' or ip == 'saocompute.eurac.edu':
        url = 'http://' + ip + '/rasdaman/ows?SERVICE=WCS&VERSION=2.0.1&REQUEST=ProcessCoverages'  
    else:
        url = 'http://' + ip + ':8080/rasdaman/ows?SERVICE=WCS&VERSION=2.0.1&REQUEST=ProcessCoverages'
        
    query = werkzeug.urls.url_fix(query)
    complete_url = url + '&query=' + query
    
    if rtype == 'DescribeCoverage':
        complete_url = 'http://' + ip + ':8080/rasdaman/ows?SERVICE=WCS&VERSION=2.0.1&REQUEST=DescribeCoverage&COVERAGEID=' + query
        
    #Fix the special characters used the input query like ' ', $ and so on
    
    print('This is the URL, used for the request:\n' + complete_url)
    
    
    try:
        #Send the request to Rasdaman and save the response in the variable "r"
        r = requests.get(complete_url, stream=True)
    except Exception as ex:
        print(tpye(ex))
        print(ex.args)
        print(ex)

    #If there is an error, plot the error message that comes from Rasdaman & exit script
    if r.status_code != requests.codes.ok:
        print('HTTP Error ' + str(r.status_code))
        root = ET.fromstring(r.text)
        for element in root.iter():
            if element.tag == '{http://www.opengis.net/ows/2.0}ExceptionText':
                print(element.text)

    print('currently receiving content of type: ' + r.headers['Content-Type'])

    #print('This is the URL, used for the request \n' + r.url)

    if r.headers['Content-Type'] == 'text/plain':
        #print('return type is text')
        # Convert CSV or json to NumpyArray
        response_text = r.text()
        output = response_text
        # The JSON version also works for 2D arrays.
        if response_text.startswith("{"):
            loaded = r.json()
            output = np.array(loaded)
        else:
            output = np.fromstring(response_text[1:-1], dtype = float, sep = ',')
    
    elif r.headers['Content-Type'] == 'text/xml': 
        # Create XML object as Ordered Dictionary here....
        file = request.urlopen(complete_url)
        data = file.read()
        file.close()
        output = xmltodict.parse(data)

    elif r.headers['Content-Type'] == 'application/json':
        # Convert JSON to NumpyArray
        loaded = r.json()
        output = np.array(loaded)

    elif r.headers['Content-Type'] == 'application/netcdf':
        print(r.headers)
        # create x array dataset from input stream
        if file_name == '':
            file_name = 'wcps_' + str(uuid.uuid4()) + '.nc'
        print('the following file has been saved locally: ' + file_name)
        with io.open(file_name, 'wb') as outfile:
            outfile.write(r.content)        
        output_open = xr.open_dataset(file_name)
        # Xarray is normally lazy loading netCDF files
        # As we want to perform intense computation, we load the file directly in the main memory with Dataset.load()
        output = xr.Dataset.load(output_open)

    elif r.headers['Content-Type'] in ['image/tiff', 'image/png', 'image/jp2', 'image/jpeg']:
        # Write response in choosen image format to disk and print filepath
        image_type = r.headers['Content-Type']
        if file_name == '':
            file_ending = image_type[6:]
            # write TIFF to disk and print filepath
            tf = 'wcps_' + str(uuid.uuid4())
            file_name = tf + '.' + file_ending
        with io.open(file_name, 'wb') as outfile:
            outfile.write(r.content)
        print('the following file has been saved locally: ' + file_name)
        output = file_name

    else:
        output = r
        output_type = r.headers['Content-Type']
        print('The response could not be processed, as it is a ' + output_type)

    return output