Add Polly client integration tests (#63)

Author: jonathan343

clients/aws-sdk-polly/tests/integration/__init__.py

@@ -0,0 +1,25 @@
+# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+from smithy_aws_core.identity import EnvironmentCredentialsResolver
+
+from aws_sdk_polly.client import PollyClient
+from aws_sdk_polly.config import Config
+
+REGION = "us-east-1"
+VOICE_ID = "Matthew"
+ENGINE = "generative"
+OUTPUT_FORMAT = "mp3"
+SAMPLE_RATE = "24000"
+TEST_TEXT = "Hello from the AWS SDK for Python Polly integration tests."
+
+
+def create_polly_client(region: str) -> PollyClient:
+    """Helper to create a PollyClient for a given region."""
+    return PollyClient(
+        config=Config(
+            endpoint_uri=f"https://polly.{region}.amazonaws.com",
+            region=region,
+            aws_credentials_identity_resolver=EnvironmentCredentialsResolver(),
+        )
+    )

clients/aws-sdk-polly/tests/integration/test_bidirectional_streaming.py

@@ -0,0 +1,114 @@
+# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+"""Test bidirectional streaming event stream handling."""
+
+import asyncio
+
+from smithy_core.aio.eventstream import DuplexEventStream
+
+from aws_sdk_polly.models import (
+    CloseStreamEvent,
+    StartSpeechSynthesisStreamActionStream,
+    StartSpeechSynthesisStreamActionStreamCloseStreamEvent,
+    StartSpeechSynthesisStreamActionStreamTextEvent,
+    StartSpeechSynthesisStreamEventStream,
+    StartSpeechSynthesisStreamEventStreamAudioEvent,
+    StartSpeechSynthesisStreamEventStreamServiceFailureException,
+    StartSpeechSynthesisStreamEventStreamServiceQuotaExceededException,
+    StartSpeechSynthesisStreamEventStreamStreamClosedEvent,
+    StartSpeechSynthesisStreamEventStreamThrottlingException,
+    StartSpeechSynthesisStreamEventStreamUnknown,
+    StartSpeechSynthesisStreamEventStreamValidationException,
+    StartSpeechSynthesisStreamInput,
+    StartSpeechSynthesisStreamOutput,
+    TextEvent,
+)
+
+from . import (
+    ENGINE,
+    OUTPUT_FORMAT,
+    REGION,
+    SAMPLE_RATE,
+    TEST_TEXT,
+    VOICE_ID,
+    create_polly_client,
+)
+
+ERROR_EVENT_TYPES = (
+    StartSpeechSynthesisStreamEventStreamValidationException,
+    StartSpeechSynthesisStreamEventStreamServiceQuotaExceededException,
+    StartSpeechSynthesisStreamEventStreamServiceFailureException,
+    StartSpeechSynthesisStreamEventStreamThrottlingException,
+)
+
+
+async def _send_text(
+    stream: DuplexEventStream[
+        StartSpeechSynthesisStreamActionStream,
+        StartSpeechSynthesisStreamEventStream,
+        StartSpeechSynthesisStreamOutput,
+    ],
+) -> None:
+    """Send text input and close the input stream."""
+    await stream.input_stream.send(
+        StartSpeechSynthesisStreamActionStreamTextEvent(value=TextEvent(text=TEST_TEXT))
+    )
+    await stream.input_stream.send(
+        StartSpeechSynthesisStreamActionStreamCloseStreamEvent(CloseStreamEvent())
+    )
+    await stream.input_stream.close()
+
+
+async def _receive_audio(
+    stream: DuplexEventStream[
+        StartSpeechSynthesisStreamActionStream,
+        StartSpeechSynthesisStreamEventStream,
+        StartSpeechSynthesisStreamOutput,
+    ],
+) -> tuple[int, int | None]:
+    """Receive synthesized audio and the final stream summary."""
+    audio_bytes = 0
+    request_characters: int | None = None
+
+    _, output_stream = await stream.await_output()
+    if output_stream is None:
+        return audio_bytes, request_characters
+
+    async for event in output_stream:
+        if isinstance(event, StartSpeechSynthesisStreamEventStreamAudioEvent):
+            if event.value.audio_chunk:
+                audio_bytes += len(event.value.audio_chunk)
+        elif isinstance(event, StartSpeechSynthesisStreamEventStreamStreamClosedEvent):
+            request_characters = event.value.request_characters
+            break
+        elif isinstance(event, ERROR_EVENT_TYPES):
+            raise event.value
+        elif isinstance(event, StartSpeechSynthesisStreamEventStreamUnknown):
+            raise RuntimeError(f"Received unknown event in stream: {event.tag}")
+        else:
+            raise RuntimeError(
+                f"Received unexpected event type in stream: {type(event).__name__}"
+            )
+
+    return audio_bytes, request_characters
+
+
+async def test_start_speech_synthesis_stream() -> None:
+    """Test bidirectional streaming with text input and audio output."""
+    client = create_polly_client(REGION)
+
+    stream = await client.start_speech_synthesis_stream(
+        input=StartSpeechSynthesisStreamInput(
+            engine=ENGINE,
+            output_format=OUTPUT_FORMAT,
+            sample_rate=SAMPLE_RATE,
+            voice_id=VOICE_ID,
+        )
+    )
+
+    results = await asyncio.gather(_send_text(stream), _receive_audio(stream))
+    audio_bytes, request_characters = results[1]
+
+    assert audio_bytes > 0, "Expected to receive synthesized audio"
+    assert request_characters == len(TEST_TEXT)

clients/aws-sdk-polly/tests/integration/test_non_streaming.py

@@ -0,0 +1,27 @@
+# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+"""Test non-streaming output type handling."""
+
+from aws_sdk_polly.models import DescribeVoicesInput, DescribeVoicesOutput
+
+from . import ENGINE, REGION, VOICE_ID, create_polly_client
+
+
+async def test_describe_voices() -> None:
+    """Test non-streaming DescribeVoices operation."""
+    client = create_polly_client(REGION)
+
+    response = await client.describe_voices(input=DescribeVoicesInput(engine=ENGINE))
+
+    assert isinstance(response, DescribeVoicesOutput)
+    assert response.voices is not None
+    assert len(response.voices) > 0
+
+    voices_by_id = {voice.id: voice for voice in response.voices if voice.id}
+    assert VOICE_ID in voices_by_id
+
+    voice = voices_by_id[VOICE_ID]
+    assert voice.language_code == "en-US"
+    assert voice.supported_engines is not None
+    assert ENGINE in voice.supported_engines

clients/aws-sdk-polly/tests/integration/test_output_streaming.py

@@ -0,0 +1,40 @@
+# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+"""Test output streaming blob handling."""
+
+from smithy_core.aio.utils import read_streaming_blob_async
+
+from aws_sdk_polly.models import SynthesizeSpeechInput, SynthesizeSpeechOutput
+
+from . import (
+    ENGINE,
+    OUTPUT_FORMAT,
+    REGION,
+    SAMPLE_RATE,
+    TEST_TEXT,
+    VOICE_ID,
+    create_polly_client,
+)
+
+
+async def test_synthesize_speech() -> None:
+    """Test output-streaming SynthesizeSpeech operation."""
+    client = create_polly_client(REGION)
+
+    response = await client.synthesize_speech(
+        input=SynthesizeSpeechInput(
+            engine=ENGINE,
+            output_format=OUTPUT_FORMAT,
+            sample_rate=SAMPLE_RATE,
+            text=TEST_TEXT,
+            voice_id=VOICE_ID,
+        )
+    )
+
+    assert isinstance(response, SynthesizeSpeechOutput)
+    assert response.content_type == "audio/mpeg"
+    assert response.request_characters == len(TEST_TEXT)
+
+    audio = await read_streaming_blob_async(response.audio_stream)
+    assert len(audio) > 0