SmartAudio/package/avs/avs-sdk/files/avs-device-sdk/Integration/test/AlexaCommunicationsLibraryT...

474 lines
20 KiB
C++

/*
* Copyright 2016-2018 Amazon.com, Inc. or its affiliates. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License").
* You may not use this file except in compliance with the License.
* A copy of the License is located at
*
* http://aws.amazon.com/apache2.0/
*
* or in the "license" file accompanying this file. This file is distributed
* on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
* express or implied. See the License for the specific language governing
* permissions and limitations under the License.
*/
/// @file AlexaCommunicationsLibraryTest.cpp
#include <future>
#include <fstream>
#include <chrono>
#include <gtest/gtest.h>
#include <ACL/AVSConnectionManager.h>
#include <AVSCommon/AVS/Attachment/AttachmentManager.h>
#include <AVSCommon/AVS/Attachment/InProcessAttachment.h>
#include <CBLAuthDelegate/SQLiteCBLAuthDelegateStorage.h>
#include <ContextManager/ContextManager.h>
#include "Integration/ACLTestContext.h"
#include "Integration/ClientMessageHandler.h"
#include "Integration/ConnectionStatusObserver.h"
#include "Integration/ObservableMessageRequest.h"
namespace alexaClientSDK {
namespace integration {
namespace test {
using namespace acl;
using namespace avsCommon::avs;
using namespace avsCommon::avs::attachment;
using namespace avsCommon::sdkInterfaces;
using namespace avsCommon::utils::sds;
using namespace registrationManager;
/// This is a basic synchronize JSON message which may be used to initiate a connection with AVS.
// clang-format off
static const std::string SYNCHRONIZE_STATE_JSON =
"{"
"\"context\":[{"
"\"header\":{"
"\"name\":\"SpeechState\","
"\"namespace\":\"SpeechSynthesizer\""
"},"
"\"payload\":{"
"\"playerActivity\":\"FINISHED\","
"\"offsetInMilliseconds\":0,"
"\"token\":\"\""
"}"
"}],"
"\"event\":{"
"\"header\":{"
"\"messageId\":\"00000000-0000-0000-0000-000000000000\","
"\"name\":\"SynchronizeState\","
"\"namespace\":\"System\""
"},"
"\"payload\":{"
"}"
"}"
"}";
// clang-format on
/// This is a partial JSON string that should not be parseable.
static const std::string BAD_SYNCHRONIZE_STATE_JSON = "{";
/**
* This string specifies a Recognize event using the specified profile.
*
* CLOSE_TALK performs end-of-speech detection on the client, so a stop-capture directive will not be received from AVS.
* NEAR_FIELD performs end-of-speech detection in AVS, so a stop-capture directive will be received from AVS.
*/
// clang-format off
#define RECOGNIZE_EVENT_JSON(PROFILE) \
"{" \
"\"event\":{" \
"\"payload\":{" \
"\"format\":\"AUDIO_L16_RATE_16000_CHANNELS_1\"," \
"\"profile\":\"" #PROFILE "\"" \
"}," \
"\"header\":{" \
"\"dialogRequestId\":\"dialogRequestId123\"," \
"\"messageId\":\"messageId123\"," \
"\"name\":\"Recognize\"," \
"\"namespace\":\"SpeechRecognizer\"" \
"}" \
"}," \
"\"context\":[{" \
"\"payload\":{" \
"\"activeAlerts\":[]," \
"\"allAlerts\":[]" \
"}," \
"\"header\":{" \
"\"name\":\"AlertsState\"," \
"\"namespace\":\"Alerts\"" \
"}" \
"}," \
"{" \
"\"payload\":{" \
"\"playerActivity\":\"IDLE\"," \
"\"offsetInMilliseconds\":0," \
"\"token\":\"\"" \
"}," \
"\"header\":{" \
"\"name\":\"PlaybackState\"," \
"\"namespace\":\"AudioPlayer\"" \
"}" \
"}," \
"{" \
"\"payload\":{" \
"\"muted\":false," \
"\"volume\":0" \
"}," \
"\"header\":{" \
"\"name\":\"VolumeState\"," \
"\"namespace\":\"Speaker\"" \
"}" \
"}," \
"{" \
"\"payload\":{" \
"\"playerActivity\":\"FINISHED\"," \
"\"offsetInMilliseconds\":0," \
"\"token\":\"\"" \
"}," \
"\"header\":{" \
"\"name\":\"SpeechState\"," \
"\"namespace\":\"SpeechSynthesizer\"" \
"}" \
"}]" \
"}"
// clang-format on
/// This string specifies a Recognize event using the CLOSE_TALK profile.
static const std::string CT_RECOGNIZE_EVENT_JSON = RECOGNIZE_EVENT_JSON(CLOSE_TALK);
/// This string specifies a Recognize event using the NEAR_FIELD profile.
static const std::string NF_RECOGNIZE_EVENT_JSON = RECOGNIZE_EVENT_JSON(NEAR_FIELD);
/// This string specifies an ExpectSpeechTimedOut event.
// clang-format off
static const std::string EXPECT_SPEECH_TIMED_OUT_EVENT_JSON =
"{"
"\"event\": {"
"\"header\": {"
"\"namespace\": \"SpeechRecognizer\","
"\"name\": \"ExpectSpeechTimedOut\","
"\"messageId\": \"messageId123\","
"},"
"\"payload\": {"
"}"
"}"
"}";
// clang-format on
/// This is a 16 bit 16 kHz little endian linear PCM audio file containing a recognized message for AVS
static const std::string RECOGNIZE_AUDIO_FILE_NAME = "recognize_test.wav";
/// This is a 16 bit 16 kHz little endian linear PCM audio file containing several seconds of silence
static const std::string SILENCE_AUDIO_FILE_NAME = "silence_test.wav";
/**
* The value of MAX_THREADS is determined by the maximum number of streams we can have active at once which is defined
* in HTTP2Connection.cpp as a constant MAX_STREAMS=10. Streams include events, the downchannel and ping.
* Since we establish a downchannel when we connect, we can only have (MAX_STREAMS - 1) events sent at once.
* Therefore, the value of MAX_THREADS = MAX_STREAMS -1.
*/
static const int MAX_CONCURRENT_STREAMS = 9;
/// Path to the AlexaClientSDKConfig.json file (from command line arguments).
static std::string g_configPath;
/// Path to resources (e.g. audio files) for tests (from command line arguments).
static std::string g_inputPath;
class AlexaCommunicationsLibraryTest : public ::testing::Test {
public:
void SetUp() override {
m_context = ACLTestContext::create(g_configPath);
ASSERT_TRUE(m_context);
m_clientMessageHandler = std::make_shared<ClientMessageHandler>(m_context->getAttachmentManager());
m_avsConnectionManager = AVSConnectionManager::create(
m_context->getMessageRouter(), false, {m_context->getConnectionStatusObserver()}, {m_clientMessageHandler});
ASSERT_TRUE(m_avsConnectionManager);
connect();
}
void TearDown() override {
// Note nullptr checks needed to avoid segaults if @c SetUp() failed.
if (m_avsConnectionManager) {
disconnect();
m_avsConnectionManager->shutdown();
}
m_context.reset();
}
void connect() {
m_avsConnectionManager->enable();
m_context->waitForConnected();
}
void disconnect() {
if (m_avsConnectionManager) {
m_avsConnectionManager->disable();
m_context->waitForDisconnected();
}
}
/*
* Function to send an Event to AVS.
*/
void sendEvent(
const std::string& jsonContent,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status expectedStatus,
std::chrono::seconds timeout,
std::shared_ptr<avsCommon::avs::attachment::AttachmentReader> attachmentReader = nullptr) {
auto messageRequest = std::make_shared<ObservableMessageRequest>(jsonContent, attachmentReader);
m_avsConnectionManager->sendMessage(messageRequest);
ASSERT_TRUE(messageRequest->waitFor(expectedStatus, timeout));
ASSERT_TRUE(messageRequest->hasSendCompleted() || messageRequest->wasExceptionReceived());
}
/**
* Function to create an attachmentReader given a filename.
*/
std::shared_ptr<InProcessAttachmentReader> createAttachmentReader(const std::string& fileName) {
// 1MB is large enough for our test audio samples.
const int mbBytes = 1024 * 1024;
// create an SDS with 1MB size, so we can write the entire audio file into it.
auto sdsBufferSize = InProcessSDS::calculateBufferSize(mbBytes);
auto sdsBuffer = std::make_shared<InProcessSDSTraits::Buffer>(sdsBufferSize);
std::shared_ptr<InProcessSDS> sds = InProcessSDS::create(sdsBuffer);
// open the file.
auto is = std::make_shared<std::ifstream>(fileName, std::ios::binary);
EXPECT_TRUE(is->is_open());
// read the data from the file into the SDS, via the local buffer.
std::vector<char> localBuffer(mbBytes);
auto attachmentWriter = InProcessAttachmentWriter::create(sds);
while (*is) {
// data -> local buffer
is->read(localBuffer.data(), mbBytes);
size_t numBytesRead = static_cast<size_t>(is->gcount());
// local buffer -> sds
AttachmentWriter::WriteStatus writeStatus = AttachmentWriter::WriteStatus::OK;
attachmentWriter->write(localBuffer.data(), numBytesRead, &writeStatus);
// confirm the SDS write operation went ok
bool writeStatusOk =
(AttachmentWriter::WriteStatus::OK == writeStatus ||
AttachmentWriter::WriteStatus::CLOSED == writeStatus);
EXPECT_TRUE(writeStatusOk);
}
// closing the writer indicates to readers of the SDS that there is no more data that will be written.
attachmentWriter->close();
// create and return the reader.
std::shared_ptr<InProcessAttachmentReader> attachmentReader =
InProcessAttachmentReader::create(ReaderPolicy::NONBLOCKING, sds);
EXPECT_NE(attachmentReader, nullptr);
return attachmentReader;
}
/**
* This function sends an Event to AVS, deciding to include an audio attachment based upon a random factor.
*/
void sendRandomEvent() {
std::shared_ptr<InProcessAttachmentReader> attachmentReader;
std::string eventJson = SYNCHRONIZE_STATE_JSON;
auto status = avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS_NO_CONTENT;
if (rand() % 2) {
eventJson = CT_RECOGNIZE_EVENT_JSON;
attachmentReader = createAttachmentReader(g_inputPath + "/" + RECOGNIZE_AUDIO_FILE_NAME);
status = avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS;
}
sendEvent(eventJson, status, std::chrono::seconds(40), attachmentReader);
}
/// Context for running ACL based tests.
std::unique_ptr<ACLTestContext> m_context;
std::shared_ptr<ClientMessageHandler> m_clientMessageHandler;
std::shared_ptr<AVSConnectionManager> m_avsConnectionManager;
};
/// Test connecting and disconnecting from AVS.
TEST_F(AlexaCommunicationsLibraryTest, testConnectAndDisconnect) {
// Connect is called in SetUp and disconnect is called in TearDown. Simply check that we are connected.
ASSERT_TRUE(m_avsConnectionManager->isConnected());
}
/**
* Test sending an Event to AVS.
*
* This test sends a SynchronizeState Event, which does not require an attachment, or receive a Directive in response.
*
* @see https://developer.amazon.com/public/solutions/alexa/alexa-voice-service/reference/system#synchronizestate
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendEvent) {
sendEvent(
SYNCHRONIZE_STATE_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS_NO_CONTENT,
std::chrono::seconds(10));
}
/**
* Function that tests the behavior of the ACL when an improperly formatted message is sent, expecting the server
* to return a bad request status.
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendInvalidEvent) {
sendEvent(
BAD_SYNCHRONIZE_STATE_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::BAD_REQUEST,
std::chrono::seconds(10));
}
/**
* Test ability to send an event with attachments to Alexa Voice Service.
*
* This test sends a RecognizeAudio event, which requires an attachment of the audio stream. In this case,
* we send a pre-recorded audio file which asks Alexa to "tell me a joke".
*
* @see https://developer.amazon.com/public/solutions/alexa/alexa-voice-service/reference/speechrecognizer#recognize
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendEventWithAttachment) {
auto attachmentReader = createAttachmentReader(g_inputPath + "/" + RECOGNIZE_AUDIO_FILE_NAME);
sendEvent(
CT_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS,
std::chrono::seconds(10),
attachmentReader);
}
/**
* Test sending an Event and receiving a Directive in response.
*
* This test sends a Recognize event, which includes an audio attachment. In this case, the audio file sent to AVS
* asks Alexa to "tell me a joke". The Speak Directive we expect in response will be the joke.
*
* @see https://developer.amazon.com/public/solutions/alexa/alexa-voice-service/reference/audioplayer#states
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendEventAndReceiveDirective) {
auto attachmentReader = createAttachmentReader(g_inputPath + "/" + RECOGNIZE_AUDIO_FILE_NAME);
sendEvent(
CT_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS,
std::chrono::seconds(10),
attachmentReader);
// We expect to receive Directives in response to the recognize Event. Wait for the first one.
ASSERT_TRUE(m_clientMessageHandler->waitForNext(std::chrono::seconds(20)));
}
/**
* Test sending multiple Events in succession and verify that AVS responds to each of them.
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendEventsSerially) {
const int NUMBER_OF_SUCCESSIVE_SENDS = 10;
for (int i = 0; i < NUMBER_OF_SUCCESSIVE_SENDS; ++i) {
auto attachmentReader = createAttachmentReader(g_inputPath + "/" + RECOGNIZE_AUDIO_FILE_NAME);
sendEvent(
CT_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS,
std::chrono::seconds(10),
attachmentReader);
}
}
/**
* Test sending multiple Events concurrently and verify that AVS responds to each of them.
*/
TEST_F(AlexaCommunicationsLibraryTest, testSendEventsConcurrently) {
std::vector<std::future<void>> futures;
for (int i = 0; i < MAX_CONCURRENT_STREAMS; ++i) {
auto future = std::async(std::launch::async, [this]() { sendRandomEvent(); });
futures.push_back(std::move(future));
}
for (auto& future : futures) {
auto status = future.wait_for(std::chrono::seconds(20));
ASSERT_EQ(status, std::future_status::ready);
}
}
/**
* Test receiving an AVS Directive on the DownChannel.
*
* DownChannel Directives are unsolicited - they may be pushed to the client by AVS at any time.
* This test exercises this functionality by sending a RecognizeAudio Event with silence for the attached audio.
* By using the NEAR_FIELD audio profile, we expect AVS to detect the end of speech and send a StopCapture Directive.
*
* @see https://developer.amazon.com/public/solutions/alexa/alexa-voice-service/reference/speechrecognizer#profiles
*/
TEST_F(AlexaCommunicationsLibraryTest, testReceiveDirectiveOnDownchannel) {
auto attachmentReader = createAttachmentReader(g_inputPath + "/" + SILENCE_AUDIO_FILE_NAME);
sendEvent(
NF_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS_NO_CONTENT,
std::chrono::seconds(10),
attachmentReader);
// Wait for the StopCapture Directive to be received.
ASSERT_TRUE(m_clientMessageHandler->waitForNext(std::chrono::seconds(20)));
}
/**
* Test that a connection to AVS persists between sending Events.
*/
TEST_F(AlexaCommunicationsLibraryTest, testPersistentConnection) {
auto attachmentReader = createAttachmentReader(g_inputPath + "/" + RECOGNIZE_AUDIO_FILE_NAME);
sendEvent(
CT_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS,
std::chrono::seconds(10),
attachmentReader);
ASSERT_FALSE(m_context->getConnectionStatusObserver()->waitFor(
ConnectionStatusObserverInterface::Status::DISCONNECTED, std::chrono::seconds(20)))
<< "Connection changed after a response was received";
sendEvent(
CT_RECOGNIZE_EVENT_JSON,
avsCommon::sdkInterfaces::MessageRequestObserverInterface::Status::SUCCESS_NO_CONTENT,
std::chrono::seconds(10),
attachmentReader);
}
/**
* Test add- and removeConnectionStatuObserver, expecting the observer to be updated only when it is added.
*/
TEST_F(AlexaCommunicationsLibraryTest, testMultipleConnectionStatusObservers) {
auto observer = std::make_shared<ConnectionStatusObserver>();
m_avsConnectionManager->addConnectionStatusObserver(observer);
ASSERT_TRUE(observer->waitFor(ConnectionStatusObserverInterface::Status::CONNECTED));
// Remove the observer and disconnect, expecting the status to not change.
m_avsConnectionManager->removeConnectionStatusObserver(observer);
disconnect();
ASSERT_EQ(observer->getConnectionStatus(), ConnectionStatusObserverInterface::Status::CONNECTED);
ASSERT_TRUE(
m_context->getConnectionStatusObserver()->waitFor(ConnectionStatusObserverInterface::Status::DISCONNECTED));
}
} // namespace test
} // namespace integration
} // namespace alexaClientSDK
int main(int argc, char** argv) {
::testing::InitGoogleTest(&argc, argv);
if (argc < 3) {
std::cerr << "USAGE: " << std::string(argv[0]) << " <path_to_auth_delgate_config> <path_to_inputs_folder>"
<< std::endl;
return 1;
} else {
alexaClientSDK::integration::test::g_configPath = std::string(argv[1]);
alexaClientSDK::integration::test::g_inputPath = std::string(argv[2]);
return RUN_ALL_TESTS();
}
}