ace-query.py

''' ace-query.py
    This file makes queries through sparql to dbpedia
    It returns types and subjects values that are associated
    with given keywords. It will also compare similarities
    between multiple keywords to see if all keywords given share
    any similar attributes.
'''
import sys
import traceback
import urllib, urllib2
import json

def query(q,epr,f='application/json'):
    try:
        params = {'query': q}
        params = urllib.urlencode(params)
        opener = urllib2.build_opener(urllib2.HTTPHandler)
        request = urllib2.Request(epr+'?'+params)
        request.add_header('Accept', f)
        request.get_method = lambda: 'GET'
        url = opener.open(request)
        return url.read()
    except:
        e = sys.exc_info()[0]
        traceback.print_exc(file=sys.stdout)
        raise e

def main():
    argc = len(sys.argv)

    if argc == 1:
        print( "Error: no keyword given.")
        exit(0)

    bl_types = []
    bl_subjects = []

    for i in range(argc-1):
        keyword = sys.argv[i+1]

        blah = print_types(keyword)
        bl_types.append(blah)

        blah = print_subjects(keyword)
        bl_subjects.append(blah)

    #check_similarities()
    exit(0)



def check_similarities(tl,sl):
    final_list = []
    tl_len = len(tl)
    sl_len = len(sl)

    if tl_len <= 1 or sl_len <= 1 :
        return

    '''for i in range(tl_len-1):
        first = tl[i]
        last = tl[i+1]

        # check all members of 'first' to see if they are in 'last'
        #for j in range(first):
        #    if first[j] in last'''


def print_types(q):
    myquery = """
        PREFIX w3t: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        SELECT ?y
        WHERE {
         dbr:""" + q + """ w3t:type ?y.
        }
        """

    # we want to extract the type of the keyword from the returned json string
    #ex - http://www.w3.org/1999/02/22-rdf-syntax-ns#type
    json_obj =  query(myquery,"http://dbpedia.org/sparql")
    json_data = json.loads(json_obj)
    results = json_data['results']
    results_data = results['bindings']

    if len(results_data) == 0:
        print q + ' has no type attribute'
        return []
    else:
        print q + ' has the following type attributes:'

    type_list = []
    for i in range(len(results_data)):
        current_index = results_data[i]
        current_obj = current_index['y']
        obj_type = current_obj['value']
        print '\t' + obj_type
        type_list.append(obj_type)
    return type_list

def print_subjects(q):
    # --------------------------------------------------------------------------
    # now get the subject lines that the keyword falls under
    # --------------------------------------------------------------------------
    myquery = """
       PREFIX purl: <http://purl.org/dc/terms/>
       SELECT ?y
       WHERE {
        dbr:""" + q + """ purl:subject ?y.
       }
       """

    # we want to extract the type of the keyword from the returned json string
    #ex - http://www.w3.org/1999/02/22-rdf-syntax-ns#type
    json_obj =  query(myquery,"http://dbpedia.org/sparql")
    json_data = json.loads(json_obj)
    results = json_data['results']
    results_data = results['bindings']

    if len(results_data) == 0:
        print q + ' has no subject attribute'
        return []
    else:
        print q + ' has the following subject attributes:'
    type_list = []
    for i in range(len(results_data)):
        current_index = results_data[i]
        current_obj = current_index['y']
        obj_type = current_obj['value']
        print '\t' + obj_type
        type_list.append(obj_type)
    return type_list



if __name__ == "__main__":
    main()