Network traffic annotation added to speech_recognition_engine.

content/browser/speech/speech_recognition_engine.cc

 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "content/browser/speech/speech_recognition_engine.h"
 
 #include <algorithm>
 #include <vector>
 
 #include "base/big_endian.h"
 #include "base/bind.h"
 #include "base/rand_util.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_util.h"
 #include "base/strings/utf_string_conversions.h"
 #include "base/time/time.h"
 #include "content/browser/speech/audio_buffer.h"
 #include "content/browser/speech/proto/google_streaming_api.pb.h"
 #include "content/public/common/speech_recognition_error.h"
 #include "content/public/common/speech_recognition_result.h"
 #include "google_apis/google_api_keys.h"
 #include "net/base/escape.h"
 #include "net/base/load_flags.h"
+#include "net/traffic_annotation/network_traffic_annotation.h"
 #include "net/url_request/http_user_agent_settings.h"
 #include "net/url_request/url_fetcher.h"
 #include "net/url_request/url_request_context.h"
 #include "net/url_request/url_request_context_getter.h"
 #include "net/url_request/url_request_status.h"
 
 using net::URLFetcher;
 
 namespace content {
 namespace {
@@ skipping 296 matching lines @@
   // Setup downstream fetcher.
   std::vector<std::string> downstream_args;
   downstream_args.push_back(
       "key=" + net::EscapeQueryParamValue(google_apis::GetAPIKey(), true));
   downstream_args.push_back("pair=" + request_key);
   downstream_args.push_back("output=pb");
   GURL downstream_url(std::string(kWebServiceBaseUrl) +
                       std::string(kDownstreamUrl) +
                       base::JoinString(downstream_args, "&"));
 
-  downstream_fetcher_ = URLFetcher::Create(
-      kDownstreamUrlFetcherIdForTesting, downstream_url, URLFetcher::GET, this);
+  net::NetworkTrafficAnnotationTag downstream_traffic_annotation =
+      net::DefineNetworkTrafficAnnotation("speech_recognition_downstream", R"(
+        semantics {
+          sender: "Speech Recognition"
+          description:
+            "Chrome provides translation from speech audio recorded with a "
+            "microphone to text, by using the Google speech recognition web "
+            "service. Audio is sent to Google's servers and text is returned."
+          trigger:
+            "The user chooses to start the recognition by clicking the "
+            "microphone icon in the Google search field."
+          data: "Audio recorded with the microphone."

[Henrik Grunell, 2017/05/09 09:03:30]

This only gets text data, afaik it only sends a request for the data providing a
unique random id for this particular speech recognition request. I'm not sure
what should be written here.

[Ramin Halavati, 2017/05/09 09:24:18]

You mean this request just sends the id to initialize the speech recognition
request, and the next one sends the actual data?
Could you please briefly explain the steps?

[Henrik Grunell, 2017/05/11 07:14:08]

A unique id is generated for a speech recognition request. This network request
(downstream) sends an id for getting the text response. Then the next request
(upstream) sends the audio data along with the id. When the server has finished
processing the audio data and produced a text response it replies to first
request.

[Ramin Halavati, 2017/05/11 07:21:09]

Done.

+          destination: GOOGLE_OWNED_SERVICE
+        }
+        policy {
+          cookies_allowed: false
+          setting:
+            "The user must allow the browser to access the microphone in a "
+            "popup notification. This is set per site (hostname pattern). In "
+            "the content settings, microphone access can be turned off for all "
+            "sites and site specific settings can be changed."
+          policy_exception_justification: "Not implemented."
+        })");
+  downstream_fetcher_ =
+      URLFetcher::Create(kDownstreamUrlFetcherIdForTesting, downstream_url,
+                         URLFetcher::GET, this, downstream_traffic_annotation);
   downstream_fetcher_->SetRequestContext(url_context_.get());
   downstream_fetcher_->SetLoadFlags(net::LOAD_DO_NOT_SAVE_COOKIES |
                                     net::LOAD_DO_NOT_SEND_COOKIES |
                                     net::LOAD_DO_NOT_SEND_AUTH_DATA);
   downstream_fetcher_->Start();
 
   // Setup upstream fetcher.
   // TODO(hans): Support for user-selected grammars.
   std::vector<std::string> upstream_args;
   upstream_args.push_back("key=" +
@@ skipping 34 matching lines @@
     if (preamble_encoder_)
       audio_format = preamble_encoder_->GetMimeType() + ",";
     audio_format += encoder_->GetMimeType();
     upstream_args.push_back(
         "audioFormat=" + net::EscapeQueryParamValue(audio_format, true));
   }
   GURL upstream_url(std::string(kWebServiceBaseUrl) +
                     std::string(kUpstreamUrl) +
                     base::JoinString(upstream_args, "&"));
 
-  upstream_fetcher_ = URLFetcher::Create(kUpstreamUrlFetcherIdForTesting,
-                                         upstream_url, URLFetcher::POST, this);
+  net::NetworkTrafficAnnotationTag upstream_traffic_annotation =
+      net::DefineNetworkTrafficAnnotation("speech_recognition_upstream", R"(
+        semantics {
+          sender: "Speech Recognition"
+          description:
+            "Chrome provides translation from speech audio recorded with a "
+            "microphone to text, by using the Google speech recognition web "
+            "service. Audio is sent to Google's servers and text is returned."

[Henrik Grunell, 2017/05/09 09:03:30]

Change to "Audio is sent to Google's servers (upstream) and text is returned
(downstream)."

Here and above in downstream annotation.

I think it could make sense to have a general description, or should the it be
strictly for this request?

[Ramin Halavati, 2017/05/09 09:24:18]

We are working on a next version of annotation functions, which will allow
defining a general annotation and adding details to it for different cases, but
it is not finalized yet.
We can merge these two after that has landed.

[Henrik Grunell, 2017/05/11 07:14:08]

Acknowledged.

+          trigger:
+            "The user chooses to start the recognition by clicking the "
+            "microphone icon in the Google search field."
+          data: "Audio recorded with the microphone."
+          destination: GOOGLE_OWNED_SERVICE
+        }
+        policy {
+          cookies_allowed: false
+          setting:
+            "The user must allow the browser to access the microphone in a "
+            "popup notification. This is set per site (hostname pattern). In "
+            "the content settings, microphone access can be turned off for all "
+            "sites and site specific settings can be changed."
+          policy_exception_justification: "Not implemented."
+        })");
+  upstream_fetcher_ =
+      URLFetcher::Create(kUpstreamUrlFetcherIdForTesting, upstream_url,
+                         URLFetcher::POST, this, upstream_traffic_annotation);
   if (use_framed_post_data_)
     upstream_fetcher_->SetChunkedUpload("application/octet-stream");
   else
     upstream_fetcher_->SetChunkedUpload(encoder_->GetMimeType());
   upstream_fetcher_->SetRequestContext(url_context_.get());
   upstream_fetcher_->SetReferrer(config_.origin_url);
   upstream_fetcher_->SetLoadFlags(net::LOAD_DO_NOT_SAVE_COOKIES |
                                   net::LOAD_DO_NOT_SEND_COOKIES |
                                   net::LOAD_DO_NOT_SEND_AUTH_DATA);
   upstream_fetcher_->Start();
@@ skipping 235 matching lines @@
 }
 
 SpeechRecognitionEngine::FSMEventArgs::FSMEventArgs(FSMEvent event_value)
     : event(event_value) {
 }
 
 SpeechRecognitionEngine::FSMEventArgs::~FSMEventArgs() {
 }
 
 }  // namespace content