Isolate stream property.

Author: daspecster

speech/google/cloud/speech/_gax.py

@@ -182,7 +182,7 @@ def streaming_recognize(self, sample, language_code=None,
                        .cloud_speech_pb2.StreamingRecognizeResponse`
         :returns: ``StreamingRecognizeResponse`` instances.
         """
-        if sample.stream.closed:
+        if sample.stream is None or sample.stream.closed:
             raise ValueError('Stream is closed.')
 
         requests = _stream_requests(sample, language_code=language_code,

speech/google/cloud/speech/client.py

@@ -66,13 +66,12 @@ def __init__(self, credentials=None, http=None, use_gax=None):
         else:
             self._use_gax = use_gax
 
-    def sample(self, content=None, source_uri=None, encoding=None,
+    def sample(self, content=None, source_uri=None, stream=None, encoding=None,
                sample_rate=None):
         """Factory: construct Sample to use when making recognize requests.
 
-        :type content: bytes or file
-        :param content: (Optional) Bytes object containing audio or file-like
-                        object which yields bytes.
+        :type content: bytes
+        :param content: (Optional) Bytes object containing audio data.
 
         :type source_uri: str
         :param source_uri: (Optional) URI that points to a file that contains
@@ -81,6 +80,9 @@ def sample(self, content=None, source_uri=None, encoding=None,
                            supported, which must be specified in the following
                            format: ``gs://bucket_name/object_name``.
 
+        :type stream: file
+        :param stream: (Optional) File like object to stream.
+
         :type encoding: str
         :param encoding: encoding of audio data sent in all RecognitionAudio
                          messages, can be one of: :attr:`~.Encoding.LINEAR16`,
@@ -98,7 +100,7 @@ def sample(self, content=None, source_uri=None, encoding=None,
         :rtype: :class:`~google.cloud.speech.sample.Sample`
         :returns: Instance of ``Sample``.
         """
-        return Sample(content=content, source_uri=source_uri,
+        return Sample(content=content, source_uri=source_uri, stream=stream,
                       encoding=encoding, sample_rate=sample_rate, client=self)
 
     @property

speech/google/cloud/speech/sample.py

@@ -22,9 +22,8 @@
 class Sample(object):
     """Representation of an audio sample to be used with Google Speech API.
 
-    :type content: bytes or file
-    :param content: (Optional) Bytes object containing audio or file-like
-                    object which yields bytes.
+    :type content: bytes
+    :param content: (Optional) Bytes object containing audio data.
 
     :type source_uri: str
     :param source_uri: (Optional) URI that points to a file that contains
@@ -33,6 +32,9 @@ class Sample(object):
                        supported, which must be specified in the following
                        format: ``gs://bucket_name/object_name``.
 
+    :type stream: file
+    :param stream: (Optional) File like object to stream.
+
     :type encoding: str
     :param encoding: encoding of audio data sent in all RecognitionAudio
                      messages, can be one of: :attr:`~.Encoding.LINEAR16`,
@@ -53,17 +55,21 @@ class Sample(object):
     default_encoding = Encoding.FLAC
     default_sample_rate = 16000
 
-    def __init__(self, content=None, source_uri=None,
+    def __init__(self, content=None, source_uri=None, stream=None,
                  encoding=None, sample_rate=None, client=None):
         self._client = client
 
-        no_source = content is None and source_uri is None
-        both_source = content is not None and source_uri is not None
+        no_source = content is None and source_uri is None and stream is None
+        both_source = (content is not None
+                       and source_uri is not None
+                       and stream is not None)
         if no_source or both_source:
-            raise ValueError('Supply one of \'content\' or \'source_uri\'')
+            raise ValueError('Supply one of '
+                             '\'content\',  \'source_uri\', \'stream\'')
 
         self._content = content
         self._source_uri = source_uri
+        self._stream = stream
 
         if sample_rate is not None and not 8000 <= sample_rate <= 48000:
             raise ValueError('The value of sample_rate must be between 8000'
@@ -97,13 +103,10 @@ def source_uri(self):
     def content(self):
         """Bytes of audio content.
 
-        :rtype: bytes or file
-        :returns: Byte stream of audio content or file like object.
+        :rtype: bytes
+        :returns: Byte stream of audio content.
         """
-        try:
-            return self._content.read()
-        except AttributeError:
-            return self._content
+        return self._content
 
     @property
     def sample_rate(self):
@@ -121,10 +124,9 @@ def stream(self):
         :rtype: file
         :returns: File like object to stream.
         """
-        stream = self._content
-        if getattr(stream, 'read', None) is None:
+        if getattr(self._stream, 'read', None) is None:
             return None
-        return stream
+        return self._stream
 
     @property
     def encoding(self):

speech/unit_tests/test__gax.py

@@ -106,7 +106,7 @@ def test_stream_requests(self):
         from google.cloud.grpc.speech.v1beta1.cloud_speech_pb2 import (
             StreamingRecognizeRequest)
 
-        sample = Sample(content=BytesIO(self.AUDIO_CONTENT),
+        sample = Sample(stream=BytesIO(self.AUDIO_CONTENT),
                         encoding=speech.Encoding.FLAC,
                         sample_rate=self.SAMPLE_RATE)
         language_code = 'US-en'

speech/unit_tests/test_client.py

@@ -520,7 +520,7 @@ def speech_api(channel=None):
                      make_secure_channel=make_channel):
             client._speech_api = _gax.GAPICSpeechAPI(client)
 
-        sample = client.sample(content=stream,
+        sample = client.sample(stream=stream,
                                encoding=Encoding.LINEAR16,
                                sample_rate=self.SAMPLE_RATE)
 
@@ -593,7 +593,7 @@ def speech_api(channel=None):
                      make_secure_channel=make_channel):
             client._speech_api = _gax.GAPICSpeechAPI(client)
 
-        sample = client.sample(content=stream,
+        sample = client.sample(stream=stream,
                                encoding=Encoding.LINEAR16,
                                sample_rate=self.SAMPLE_RATE)
 
@@ -637,7 +637,7 @@ def speech_api(channel=None):
                      make_secure_channel=make_channel):
             client._speech_api = _gax.GAPICSpeechAPI(client)
 
-        sample = client.sample(content=stream,
+        sample = client.sample(stream=stream,
                                encoding=Encoding.LINEAR16,
                                sample_rate=self.SAMPLE_RATE)
 

speech/unit_tests/test_sample.py

@@ -38,28 +38,31 @@ def test_initialize_sample(self):
         self.assertEqual(sample.sample_rate, self.SAMPLE_RATE)
 
     def test_content_and_source_uri(self):
+        from io import BytesIO
+
         with self.assertRaises(ValueError):
             self._make_one(content='awefawagaeragere',
                            source_uri=self.AUDIO_SOURCE_URI)
 
-    def test_content_is_file_like(self):
-        from io import BytesIO
-        from google.cloud.speech.encoding import Encoding
+        with self.assertRaises(ValueError):
+            self._make_one(stream=BytesIO(b'awefawagaeragere'),
+                           source_uri=self.AUDIO_SOURCE_URI)
 
-        test_bytes = b'testing 1 2 3 4'
-        content = BytesIO(test_bytes)
-        sample = self._make_one(content=content,
-                                encoding=Encoding.FLAC,
-                                sample_rate=self.SAMPLE_RATE)
-        self.assertEqual(sample.content, test_bytes)
+        with self.assertRaises(ValueError):
+            self._make_one(content='awefawagaeragere',
+                           stream=BytesIO(b'awefawagaeragere'),
+                           source_uri=self.AUDIO_SOURCE_URI)
 
     def test_stream_property(self):
+        from io import BytesIO
         from google.cloud.speech.encoding import Encoding
 
-        content = b'abc 1 2 3'
-        sample = self._make_one(content=content, encoding=Encoding.FLAC,
+        data = b'abc 1 2 3 4'
+        stream = BytesIO(data)
+        sample = self._make_one(stream=stream, encoding=Encoding.FLAC,
                                 sample_rate=self.SAMPLE_RATE)
-        self.assertIsNone(sample.stream)
+        self.assertEqual(sample.stream, stream)
+        self.assertEqual(sample.stream.read(), data)
 
     def test_bytes_converts_to_file_like_object(self):
         from google.cloud import speech

system_tests/speech.py

@@ -118,7 +118,7 @@ def _make_async_request(self, content=None, source_uri=None,
     def _make_streaming_request(self, file_obj, single_utterance=True,
                                 interim_results=False):
         client = Config.CLIENT
-        sample = client.sample(content=file_obj,
+        sample = client.sample(stream=file_obj,
                                encoding=speech.Encoding.LINEAR16,
                                sample_rate=16000)
         return sample.streaming_recognize(single_utterance=single_utterance,