Merge pull request #55 from cp2boston/RCB-464_text-embedding

text-embedding added

Author: hgaudet

examples/text_embedding.py

@@ -0,0 +1,29 @@
+# -*- coding: utf-8 -*-
+
+"""
+Example code to call Rosette API to get text vectors from a piece of text.
+"""
+
+import argparse
+import json
+import os
+
+from rosette.api import API, DocumentParameters
+
+
+def run(key, altUrl='https://api.rosette.com/rest/v1/'):
+    # Create an API instance
+    api = API(user_key=key, service_url=altUrl)
+    embeddings_data = "Cambridge, Massachusetts"
+    params = DocumentParameters()
+    params["content"] = embeddings_data
+    return api.text_embedding(params)
+
+parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter, description='Calls the ' + os.path.splitext(os.path.basename(__file__))[0] + ' endpoint')
+parser.add_argument('-k', '--key', help='Rosette API Key', required=True)
+parser.add_argument('-u', '--url', help="Alternative API URL", default='https://api.rosette.com/rest/v1/')
+
+if __name__ == '__main__':
+    args = parser.parse_args()
+    result = run(args.key, args.url)
+    print(json.dumps(result, indent=2, ensure_ascii=False, sort_keys=True).encode("utf8"))

rosette/api.py

@@ -24,7 +24,6 @@
 import sys
 import time
 import os
-from socket import gaierror
 import requests
 import re
 import warnings
@@ -35,15 +34,6 @@
 _IsPy3 = sys.version_info[0] == 3
 
 
-try:
-    import urlparse
-except ImportError:
-    import urllib.parse as urlparse
-try:
-    import httplib
-except ImportError:
-    import http.client as httplib
-
 if _IsPy3:
     _GZIP_SIGNATURE = _GZIP_BYTEARRAY
 else:
@@ -843,3 +833,11 @@ def matched_name(self, parameters):
         @type parameters: L{NameSimilarityParameters}
         @return: A python dictionary containing the results of name matching."""
         return self.name_similarity(parameters)
+
+    def text_embedding(self, parameters):
+        """
+        Create an L{EndpointCaller}  to identify text vectors found in the texts
+        to which it is applied and call it.
+        @type parameters: L{DocumentParameters} or L{str}
+        @return: A python dictionary containing the results of text embedding."""
+        return EndpointCaller(self, "text-embedding").call(parameters)

tests/test_rosette_api.py

@@ -579,3 +579,14 @@ def test_for_name_translation_required_parameters(api, json_response):
 
     httpretty.disable()
     httpretty.reset()
+
+
+def test_the_text_embedded_endpoint(api, json_response, doc_params):
+    httpretty.enable()
+    httpretty.register_uri(httpretty.POST, "https://api.rosette.com/rest/v1/text-embedding",
+                           body=json_response, status=200, content_type="application/json")
+
+    result = api.text_embedding(doc_params)
+    assert result["name"] == "Rosette API"
+    httpretty.disable()
+    httpretty.reset()