Conversational speech tool: timing model with data access.

webrtc/modules/audio_processing/test/conversational_speech/timing.h

Index: webrtc/modules/audio_processing/test/conversational_speech/timing.h
diff --git a/webrtc/modules/audio_processing/test/conversational_speech/timing.h b/webrtc/modules/audio_processing/test/conversational_speech/timing.h
new file mode 100644
index 0000000000000000000000000000000000000000..ff8d5e14cd0a21eac27bc91ba8f5f056802659f1
--- /dev/null
+++ b/webrtc/modules/audio_processing/test/conversational_speech/timing.h
@@ -0,0 +1,62 @@
+/*
+ *  Copyright (c) 2017 The WebRTC project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+#ifndef WEBRTC_MODULES_AUDIO_PROCESSING_TEST_CONVERSATIONAL_SPEECH_TIMING_H_
+#define WEBRTC_MODULES_AUDIO_PROCESSING_TEST_CONVERSATIONAL_SPEECH_TIMING_H_
+
+#include <memory>
+#include <string>
+#include <vector>
+
+namespace webrtc {
+namespace test {
+namespace conversational_speech {
+
+class Timing {
+ public:
+  struct Turn{
+    Turn(std::string speaker_name_, std::string audiotrack_file_name_,
+         int offset_)

[kwiberg-webrtc, 2017/03/16 14:10:09]

No trailing underscore on non-member variables, please. (IIRC you can give them
the same names as the members without running into trouble. But it's probably
better to just give them other names.)

[AleBzk, 2017/03/20 14:25:28]

Done.

+        : speaker_name(speaker_name_),
+          audiotrack_file_name(audiotrack_file_name_),
+          offset(offset_) {}
+    bool operator ==(const Turn &b) const;
+    std::string speaker_name;
+    std::string audiotrack_file_name;
+    int offset;
+  };
+
+  Timing();
+  ~Timing();
+
+  // Removes all the existing turn entries.
+  void Clear();
+
+  // Appends the next turn.
+  void AppendTurn(std::unique_ptr<Timing::Turn> turn);
+
+  // Loads the turns from a file.
+  void Load(const std::string& timing_filepath);
+
+  // Writes the turns into a file.
+  void Save(const std::string& timing_filepath) const;
+
+  const std::vector<std::unique_ptr<Timing::Turn>>& turns() const;
+  std::size_t Size() const;
+
+ private:
+  std::vector<std::unique_ptr<Turn>> turns_;

[kwiberg-webrtc, 2017/03/16 14:10:09]

Turn is a simple value type. Why not just have a std::vector<Turn>? Do you
expect that to lead to excessive copying of Turn objects? It doesn't look like
that would happen.

[AleBzk, 2017/03/20 14:25:28]

True. At this moment copying is not an issue. I expect <20 Turn objects and I
don't expect that they will ever become more complex than a struct with 3
members (2 strings and one integer).

[kwiberg-webrtc, 2017/03/21 22:35:15]

OK, definitely just use a std::vector<Turn> then.

+};
+
+}  // namespace conversational_speech
+}  // namespace test
+}  // namespace webrtc
+
+#endif  // WEBRTC_MODULES_AUDIO_PROCESSING_TEST_CONVERSATIONAL_SPEECH_TIMING_H_