在Android Studio中使用OkHttp向AVS发送事件https请求后，response.body().string()没有得到downchannelStream响应

Question

我将遵循本教程：https://developer.amazon.com/en-US/docs/alexa/alexa-voice-service/manage-http2-connection.html

总结一下这个问题，在使用安卓中的OkHttp向AVS发送事件https请求后，我没有收到任何使用response.body().string()返回的downchannelStream响应。

在这里，我通过创建一个指令http请求来建立下行通道流，这意味着要根据教程保持打开状态：

private void establishDownChanDirective(String accessToken, OkHttpClient downChannelClient) throws IOException {
    // OKHttp header creation.
    final Request getRequest = new Request.Builder()
            .url("https://alexa.na.gateway.devices.a2z.com/" + AVS_API_VERSION + "/directives")//endpoint url
            .get()
            .addHeader("authorization", "Bearer " + accessToken)
            .build();

    Log.d("Request_header", getRequest.toString());

    downChannelClient.newCall(getRequest).enqueue(new Callback() {
        @Override
        public void onFailure(@NotNull Call call, @NotNull IOException e) {
            Log.d("downChannelResp", "failure: " + e.getMessage());
            call.cancel();
        }

        @Override
        public void onResponse(@NotNull Call call, @NotNull Response response) throws IOException {
            Log.d("downChannelResp", "Down channel recieved! Test 1");
            processResponse(response, "downChannelResp", true);
            Log.d("downChannelResp", "Down channel recieved! Test 2");

            responseDirective = response;
        }
    });
}

接下来，我尝试通过发送一个事件来与AVS同步：

private void sendSyncEvent(OkHttpClient downChannelClient, String accessToken) throws IOException {
    String msgId = UUID.randomUUID().toString();
    String speakToken = "";
    long offsetMili = 20; // if lags put down to 10.
    String playerActivity = "PLAYING";

    final String JSON_SYNC = "{\"context\":[{\"header\":{\"namespace\":\"SpeechRecognizer\",\"name\":\"RecognizerState\"},\"payload\":{\"wakeword\":\"ALEXA\"}},{\"header\":{\"namespace\":\"SpeechSynthesizer\",\"name\":\"SpeechState\"},\"payload\":{\"token\":\"" + speakToken + "\",\"offsetInMilliseconds\":" + offsetMili + ",\"playerActivity\":\"" + playerActivity + "\"}}],\"event\":{\"header\":{\"namespace\":\"System\",\"name\":\"SynchronizeState\",\"messageId\":\"" + msgId + "\"},\"payload\":{}}}";

    List<MultipartBody.Part> partList = new ArrayList<>();
    MultipartBody.Part syncPart = MultipartBody.Part.create(Headers.of(
            "Content-Disposition", "form-data; name=\"metadata\""),
            RequestBody.create(JSON_SYNC, JSON_TYPE));
    partList.add(syncPart);

    RequestBody body = new MultipartBody(ByteString.encodeUtf8(BOUNDARY_TERM), MultipartBody.FORM, partList);

    Log.d("part", syncPart.headers().toString());
    Log.d("body", body.contentType().toString());

    final Request postRequest = new Request.Builder()
            .url("https://alexa.na.gateway.devices.a2z.com/"+AVS_API_VERSION+"/events")//endpoint url
            .post(body)
            .addHeader("authorization", "Bearer " + accessToken)
            .addHeader("content-type", "multipart/form-data; boundary=" + BOUNDARY_TERM) // Don't know whether or not this is needed.
            .build();

    Log.d("post_request", postRequest.toString());
    Log.d("post_req_body", JSON_SYNC);

    downChannelClient.newCall(postRequest).enqueue(new Callback() {
        @Override
        public void onFailure(@NotNull Call call, @NotNull IOException e) {
            Log.d("syncResp", "failure: " + e.getMessage());
            call.cancel();
        }

        @Override
        public void onResponse(@NotNull Call call, @NotNull Response response) throws IOException {
            processResponse(response, "syncResp", false);
        }
    });
}

然后，我尝试发送一个测试识别事件，该事件(根据本教程)旨在通过初始downChannelStream返回响应：

private void testRecognizeEventAVS(OkHttpClient downChannelClient, String accessToken) throws IOException {
    final MediaType AUDIO_TYPE = MediaType.parse("application/octet-stream");

    String audioMsgId = UUID.randomUUID().toString();
    String dialogId = UUID.randomUUID().toString();
    final String JSON_SPEECH_EVENT = "{\"event\": {\"header\": {\"namespace\": \"SpeechRecognizer\",\"name\": \"Recognize\",\"messageId\": \"" + audioMsgId + "\",\"dialogRequestId\": \"" + dialogId + "\"},\"payload\": {\"profile\": \"CLOSE_TALK\", \"format\": \"AUDIO_L16_RATE_16000_CHANNELS_1\"}},\"context\": [{\"header\": {\"namespace\": \"AudioPlayer\",\"name\": \"PlaybackState\"},\"payload\": {\"token\": \"\",\"offsetInMilliseconds\": 0,\"playerActivity\": \"FINISHED\"}}, {\"header\": {\"namespace\": \"SpeechSynthesizer\",\"name\": \"SpeechState\"},\"payload\": {\"token\": \"\",\"offsetInMilliseconds\": 0,\"playerActivity\": \"FINISHED\"}}, { \"header\" : { \"namespace\" : \"Alerts\", \"name\" : \"AlertsState\" }, \"payload\" : { \"allAlerts\" : [ ], \"activeAlerts\" : [ ] } }, {\"header\": {\"namespace\": \"Speaker\",\"name\": \"VolumeState\"},\"payload\": {\"volume\": 25,\"muted\": false}}]}";

    List<MultipartBody.Part> partList = new ArrayList<>();

    // Metadata Part
    Map<String, String> metaHeaders = new HashMap<String, String>();
    metaHeaders.put("Content-Disposition", "form-data; name=\"metadata\"");
    MultipartBody.Part metaPart = MultipartBody.Part.create(Headers.of(metaHeaders), RequestBody.create(JSON_SPEECH_EVENT, JSON_TYPE));
    partList.add(metaPart);

    // Audio Part
    Map<String, String> audioHeaders = new HashMap<String, String>();
    audioHeaders.put("Content-Disposition", "form-data; name=\"metadata\"");
    MultipartBody.Part audioPart = MultipartBody.Part.create(Headers.of(audioHeaders), RequestBody.create(createTestFile(), AUDIO_TYPE));
    partList.add(audioPart);

    RequestBody reqBody = new MultipartBody(ByteString.encodeUtf8(BOUNDARY_TERM), MultipartBody.FORM, partList);

    Log.d("metaPart", metaPart.headers().toString());
    Log.d("audioPart", audioPart.headers().toString());
    Log.d("body", reqBody.contentType().toString());

    // https://developer.amazon.com/en-US/docs/alexa/alexa-voice-service/structure-http2-request.html

    Request speechRequest = new Request.Builder()
            .url("https://alexa.na.gateway.devices.a2z.com/"+AVS_API_VERSION+"/events")
            .addHeader("authorization", "Bearer " + accessToken)
            .addHeader("content-type", "multipart/form-data; boundary=" + BOUNDARY_TERM) // Don't know whether or not this is needed.
            .post(reqBody)
            .build();

    Log.d("speech_request", speechRequest.toString());

    downChannelClient.newCall(speechRequest).enqueue(new Callback() {
        @Override
        public void onFailure(@NotNull Call call, @NotNull IOException e) {
            Log.d("speechResp", "failure: " + e.getMessage());
            call.cancel();
        }

        @Override
        public void onResponse(@NotNull Call call, @NotNull Response response) throws IOException {
            processResponse(response, "speechResp", false);
        }
    });
}

这是上述每个方法中使用的processResponse方法，用于获取响应并将响应信息输出到Android日志中：

private void processResponse(Response response, final String TAG, boolean readBodySource) throws IOException {
    //Log.d(TAG, "response-string: " + response.body().string()); // This never shows up and always stops the rest of this method running for the response from establishDownChanDirective().
    Log.d(TAG, "response-success: " + response.isSuccessful());
    Log.d(TAG, "response" + response.toString());

    // Tried this from stack over flow posts, but right now we aren't even receiving a response-string from the downChannelDirective, so we need to figure that out first.
    if (readBodySource) {
        BufferedSource bufferedSource = response.body().source();

        Buffer buffer = new Buffer();

        while (!bufferedSource.exhausted()) {
            Log.w("bufferedSource", "downchannel recieved!");
            long bs = bufferedSource.read(buffer, 8192);
            Log.d("bufferedSource_read", String.valueOf(bs));
            Log.d("buffersize", String.valueOf(buffer.size()));
        }

        Log.d("buffer_response", buffer.toString());
    }
}

该方法注释掉了string-response，但是当它没有被注释掉时，它只给出D/syncResp: response-string:作为输出，而response-string对于syncResp和speechResp来说只是一个空字符串。但是，对于downChannelResp，它不提供任何输出，并且完全停止运行Log.d(TAG, "response-string: " + response.body().string());下面的其余代码。

现在当我运行这个的时候。

try {
    establishDownChanDirective(accessToken, downChannelClient); // Establish a down channel directive that will remain open.
    sendSyncEvent(downChannelClient, accessToken); // Send a Syncronize event through the same connection as the down channel directive.
    testRecognizeEventAVS(downChannelClient, accessToken); // Send a Speech directive through the same connection as the down channel directive.
    Log.d("OkHttp", "Test: Http stuff finished.");
    if (responseDirective != null) {
        Log.d("OkHttp", "Response: " + responseDirective.body().string());
    } else {
        Log.d("OkHttp", "No response!");
    }
} catch (IOException e) {
    Log.d("OkHttpError", "error: START{" + e.toString() + "}END");
    e.printStackTrace();
}

...it将此作为输出：

D/Request_header: Request{method=GET, url=https://alexa.na.gateway.devices.a2z.com/v20160207/directives, headers=[authorization:Bearer <the access token - censored for this post>]}
D/part: Content-Disposition: form-data; name="metadata"
D/body: multipart/form-data; boundary=------------------------qM9tn4VZyj
D/post_request: Request{method=POST, url=https://alexa.na.gateway.devices.a2z.com/v20160207/events, headers=[authorization:Bearer <the access token - censored for this post>, content-type:multipart/form-data; boundary=------------------------qM9tn4VZyj]}
D/post_req_body: {"context":[{"header":{"namespace":"SpeechRecognizer","name":"RecognizerState"},"payload":{"wakeword":"ALEXA"}},{"header":{"namespace":"SpeechSynthesizer","name":"SpeechState"},"payload":{"token":"","offsetInMilliseconds":20,"playerActivity":"PLAYING"}}],"event":{"header":{"namespace":"System","name":"SynchronizeState","messageId":"2c46b1a9-8b41-47be-bd09-61166b78492e"},"payload":{}}}
D/parent: /storage/emulated/0/Android/data/aut.rnd.alexa/files
D/fileexists: true
D/media_file: successfully created: true
D/metaPart: Content-Disposition: form-data; name="metadata"
D/audioPart: Content-Disposition: form-data; name="metadata"
D/body: multipart/form-data; boundary=------------------------qM9tn4VZyj
D/speech_request: Request{method=POST, url=https://alexa.na.gateway.devices.a2z.com/v20160207/events, headers=[authorization:Bearer <the access token - censored for this post>, content-type:multipart/form-data; boundary=------------------------qM9tn4VZyj]}
D/OkHttp: Test: Http stuff finished.
    No response!
D/downChannelResp: Down channel recieved! Test 1
    response-success: true
    responseResponse{protocol=h2, code=200, message=, url=https://alexa.na.gateway.devices.a2z.com/v20160207/directives}
W/bufferedSource: downchannel recieved!
D/bufferedSource_read: 18
D/buffersize: 18
D/syncResp: response-success: true
    responseResponse{protocol=h2, code=204, message=, url=https://alexa.na.gateway.devices.a2z.com/v20160207/events}
D/speechResp: response-success: true
    responseResponse{protocol=h2, code=204, message=, url=https://alexa.na.gateway.devices.a2z.com/v20160207/events}

这是意想不到的，因为响应应该返回可以转换为JSON的数据，但它似乎根本没有返回任何东西。

Answer 1

这可能是你的问题。https://square.github.io/okhttp/4.x/okhttp/okhttp3/-response-body/#the-response-body-can-be-consumed-only-once

响应body只能消费一次。

这个类可以用来流式传输非常大的响应。例如，可以使用此类读取大于分配给当前进程的整个内存的响应。它甚至可以流式传输比当前设备上总存储量更大的响应，这是视频流应用程序的常见要求。