As i understood Natural Language Understanding doesn't natively support Italian language for relation extraction. My question is : If I create a model for annotation of a specific domain in Italian language in WKS will i be able to use it in NLU? Is it possible to upload the model from WKS to NLU (like it was in alchemy)? Will it then manage Italian text?
NLU document said that
The entities and relations features can be expanded to support even more languages by using custom models in Watson Knowledge Studio.
You need to deploy your model to the your NLU instance at bluemix. then, you can try with Python, or JAVA, the NLU SDK and samples are here: https://github.com/watson-developer-cloud/java-sdk
Here is a simple python example:
import sys import os import json sys.path.append(os.path.join(os.getcwd(),'..')) import watson_developer_cloud import watson_developer_cloud.natural_language_understanding.features.v1 as features nlu = watson_developer_cloud.NaturalLanguageUnderstandingV1( version='2017-02-27', username='your user here', password='your password here') response = nlu.analyze( text='I need a Band 8 or lower Java Developer with Expertise in DB2 and SQL and Agile Methods and the Education industry', features=[features.Entities(model='10:b194c826-79bc-4f06-93c9-020dd05d47a6')]) print(json.dumps(response, indent=2))
your model need to be here -> features=[features.Entities(model='10:b194c826-79bc-4f06-93c9-020dd05d47a6')])