comparison json_data_source.py @ 0:44119c5d8cc6

Initial commit.

0:44119c5d8cc6

#!/usr/bin/env python
import json
import optparse
import urllib
import os.path

CHUNK_SIZE = 2**20 #1mb


def chunk_write( source_stream, target_stream, source_method = "read", target_method="write" ):
    source_method = getattr( source_stream, source_method )
    target_method = getattr( target_stream, target_method )
    while True:
        chunk = source_method( CHUNK_SIZE )
        if chunk:
            target_method( chunk )
        else:
            break


def deconstruct_multi_filename( multi_filename ):
    keys = [ 'primary', 'id', 'name', 'visible', 'file_type' ]
    return ( dict( zip( keys, multi_filename.split('_') ) ) )


def construct_multi_filename( id, name, file_type ):
    """ Implementation of *Number of Output datasets cannot be determined until tool run* from documentation_.
    .. _documentation: http://wiki.galaxyproject.org/Admin/Tools/Multiple%20Output%20Files
    """
    filename = "%s_%s_%s_%s_%s" % ( 'primary', id, name, 'visible', file_type )
    return filename


def download_from_query( query_data, target_output_filename  ):
    """ Download file from the json data and write it to target_output_filename.
    """
    query_url = query_data.get( 'url' )
    query_file_type = query_data.get( 'extension' )
    query_stream = urllib.urlopen( query_url )
    output_stream = open( target_output_filename, 'wb' )
    chunk_write( query_stream, output_stream )
    query_stream.close()
    output_stream.close()


def download_extra_data( query_ext_data, base_path ):
    for ext_data in query_ext_data:
        query_stream = urllib.urlopen( ext_data.get( 'url' ) )
        output_stream = open( os.path.normpath( '/'.join( [ base_path, extra_item.get( 'path' ) ] ) ), 'wb' )
        chunk_write( query_stream, output_stream )
        query_stream.close()
        output_stream.close()


def download_from_json_data( options, args ):
    """ Parse the returned JSON data and download files. Write metadata
    to flat JSON file.
    """
    json_params = json.loads( open( options.json_param_file, 'r' ).read() )
    datasource_params = json_params.get( 'param_dict' )
    dataset_id = base_dataset_id = json_params['output_data'][0]['dataset_id']
    hda_id = json_params['output_data'][0]['hda_id']
    dataset_url = json_params['url']

    url_param = datasource_params.get( file_url_name, None )
    output_filename = datasource_params.get( "output1", None )
    output_base_path = options.path

    # get JSON response from data source
    # TODO: make sure response is not enormous
    query_params = json.loads(urllib.urlopen( dataset_url ).read())
    metadata_to_write = []
    # download and write files
    for query_item in query_params:
        if isinstance( query_item, list ):
            # do something with the nested list as a collection
            for query_subitem in query_item:
                multi_name = construct_multi_filename( hda_id, output_filename, extension )
                target_output_filename = os.path.normpath( '/'.join( [ output_base_path, multi_name ] ) )
                download_from_query( query_subitem, target_output_filename )
                if query_item.get( 'extra_data' ):
                    download_extra_data( query_item.get( 'extra_data' ), output_base_path )
                metadata_to_write.append( query_item )

        elif isinstance( query_item, dict ):
            # what is the difference between hda_id and dataset_id?
            multi_name = construct_multi_filename( hda_id, output_filename, extension )
            target_output_filename = os.path.normpath( [ '/'.join( output_base_path, multi_name ) ] )
            download_from_query( query_item, target_output_filename )
            if query_item.get( 'extra_data' ):
                download_extra_data( query_item.get( 'extra_data' ), output_base_path )
            metadata_to_write.append( query_item )

    with open( json_params['job_config']['TOOL_PROVIDED_JOB_METADATA_FILE'], 'wb' ) as metadata_parameter_file:
        # write JSON metadata from flattened list
        metadata_parameter_file.write( json.dumps( metadata_to_write ) )


def __main__():
    """ Read the JSON return from a data source. Parse each line and request
    the data, download to "newfilepath", and write metadata.

    Schema
    ------

        [ {"url":"http://url_of_file",
           "name":"encode WigData",
           "extension":"wig",
           "metadata":{"db_key":"hg19"},
           "extra_data":[ {"url":"http://url_of_ext_file",
                           "path":"rel/path/to/ext_file"}
                        ]
          }
        ]

    """
    # Parse the command line options
    usage = "Usage: json_data_source.py max_size --json_param_file filename [options]"
    parser = optparse.OptionParser(usage = usage)
    parser.add_option("-j", "--json_param_file", type="string",
                    action="store", dest="json_param_file", help="json schema return data")
    parser.add_option("-p", "--path", type="string",
                    action="store", dest="newfilepath", help="new file path")

    (options, args) = parser.parse_args()
    download_from_json_data( options, args )


if __name__ == "__main__": __main__()