The Python binding requires Python 2.6 or greater and is available through pip:

pip install rosette_api

# 1. Set utf-8 encoding.
# -*- coding: utf-8 -*-

# 2. Imports from rosette.api.
from rosette.api import API, DocumentParameters, MorphologyOutput

# 3. Create API object.
api = API("[your_api-key]")

# 4. Create parameters object
params = DocumentParameters()

# 5. Set parameters.
params["content"] = u"Was ist so böse an der Europäischen Zentralbank?"

# 6. Make a call.
result = api.morphology(params)

# result is a Python dictionary that contains

{u'lemmas': [{u'text': u'Was', u'lemma': u'was'}, {u'text': u'ist', u'lemma': u'sein'}, {u'text': u'so', u'lemma': u'so'}, {u'text': u'böse', u'lemma': u'böse'}, {u'text': u'an', u'lemma': u'an'}, {u'text': u'der', u'lemma': u'der'}, {u'text': u'Europäischen', u'lemma': u'europäisch'}, {u'text': u'Zentralbank', u'lemma': u'Zentralbank'}, {u'text': u'?', u'lemma': u'?'}]}

The samples use the following procedure:

1. If the application reads text in, set encoding to utf-8 in the first line of the script.

2. Import the rosette.api packages that your application needs. The rosette.api packages include

   • API
   • DocumentParameters
   • NameMatchingParameters
   • NameTranslationParameters
   • MorphologyOutput
   • DataFormat
   • InputUnit

3. Create an API object with the user_key parameter.

4. Create a parameters object for your request input:

   Parameter	Endpoint
   NameMatchingParameters	for /matched-name
   NameTranslationParameters	for /translated-name
   DocumentParameters	for all other endpoints

5. Set the parameters required for your operation: "content" or "contentUri" for DocumentParameters; "name" and "targetLanguage" for NameTranslationParameters; "name1.text" and "name2.text" for NameMatchingParameters; Other parameters are optional.

6. Create an API method for the endpoint you are calling. The methods are

   • entities(linked) where linked is False for entity extraction and True for entity linking.
   • categories()
   • sentiment()
   • language()
   • morphology(tag) where tag is a member of MorphologyOutput: LEMMAS, PARTS_OF_SPEECH, COMPOUND_COMPONENTS, HAN_READINGS, or COMPLETE. An empty tag is equivalent to COMPLETE.
   • sentences()
   • tokens()
   • matched_name()
   • translated_name()

7. The API will return a dictionary with the results.