GeoLyzard

From SemanticLab

Connection

• the development version of geoLyzard is hosted on gecko.wu.ac.at
  • for connecting to the service create a SSH tunnel to xmdimrill and forward the ports

ssh user@xmdimrill.wu.ac.at -L 8080:gecko.wu.ac.at:8180

Usage

• the service can be used with any XML-RPC Client
  • example in Python

from xmlrpclib import ServerProxy, Error
 
FILE = 'example.xml'
 
xml_cont = open(FILE,'r').read().strip()
GEO_TAGGER_URL = 'http://localhost:8080/geotagger/geotagger'
 
enc_dict = {}
for i in xrange(1,2):
    enc_dict[str(i)]=base64.b64encode(xml_cont)
 
""" 
geoLocation takes a Gazetteer as parameter and the an encoded xml-file
valid gazetteers: C, C5000, C10000, c50000, C100000, C500000
 
"""
annotations = xmlrpc_server.Tagger.getGeoLocation("C1000000", enc_dict)

• two disambiguation algorithms are available: Default and Amitay
  • setting the disambiguation algorithm with Python:

from xmlrpclib import ServerProxy, Error
 
GEO_TAGGER_URL = 'http://localhost:8080/geotagger/geotagger'
xmlrpc_server = ServerProxy(GEO_TAGGER_URL)
 
xmlrpc_server.Tagger.setDisambiguationAlgorithm('Amitay')
# or
xmlrpc_server.Tagger.setDisambiguationAlgorithm('Default')

• depending on the algorithm several parameters are availabe:
  • Amitay:
    • minFocusPoints - Minimum of focus points, values below, will be removed from the result (default: 0.0)
    • nonAmbigiousConfidence - Confidence for non ambiguous locations (default: 0.5)
    • maxPopulationConfidence - Maximum population confidence (default: 0.5)
    • minPopulationConfidence - Minimum population confidence (default: 0.25)
    • discountFactor -Discount factor for calculating focus points of parents (default: 0.7)
  • Default:
    • narrowLocationBias (default: 3.3)

# example setting a parameter
xmlrpc_server.Tagger.setDisambiguationParameters('minFocusPoints', '100.0')