speech_summary.h

/* Copyright 2018 ID R&D Inc. All Rights Reserved. */
 
#pragma once
 
#include <voicesdk/core/config.h>
#include <voicesdk/core/common/intervals.h>
 
#include <iostream>
#include <vector>
#include <memory>
 
namespace voicesdk {
 
    struct SpeechEvent {
        bool isVoice = false;
 
        AudioInterval audioInterval;
 
        SpeechEvent() = default;
 
        SpeechEvent(bool isVoice, const AudioInterval& audioInterval) :
            isVoice(isVoice), audioInterval(audioInterval) {}
 
        friend std::ostream& operator<<(std::ostream& os, const SpeechEvent& obj) {
            os << "SpeechEvent["
               << "isVoice:" << obj.isVoice << ", "
               << "audioInterval:" << obj.audioInterval << "]";
            return os;
        }
 
        bool operator==(const SpeechEvent& other) const {
            return other.isVoice == isVoice
                && other.audioInterval == audioInterval;
        }
    };
 
    struct SpeechInfo {
        float speechLengthMs = 0.0f;
 
        float backgroundLengthMs = 0.0f;
 
        float totalLengthMs = 0.0f;
 
        SpeechInfo() = default;
 
        SpeechInfo(float speechLengthMs, float backgroundLengthMs, float totalLengthMs) :
            speechLengthMs(speechLengthMs),
            backgroundLengthMs(backgroundLengthMs),
            totalLengthMs(totalLengthMs)
        {}
 
        friend std::ostream& operator<<(std::ostream& os, const SpeechInfo& obj) {
            os << "SpeechInfo["
               << "speechLengthMs:" << obj.speechLengthMs << ", "
               << "backgroundLengthMs:" << obj.backgroundLengthMs << ", "
               << "totalLengthMs:" << obj.totalLengthMs << "]";
            return os;
        }
 
        bool operator==(const SpeechInfo& other) const {
            return other.speechLengthMs == speechLengthMs
                && other.backgroundLengthMs == backgroundLengthMs
                && other.totalLengthMs == totalLengthMs;
        }
    };
 
    struct SpeechSummary {
        std::vector<SpeechEvent> speechEvents;
 
        SpeechInfo speechInfo;
 
        SpeechSummary() = default;
 
        SpeechSummary(const std::vector<SpeechEvent>& speechEvents, const SpeechInfo& speechInfo) :
            speechEvents(speechEvents),
            speechInfo(speechInfo)
        {}
 
        friend std::ostream& operator<<(std::ostream& os, const SpeechSummary& obj) {
            os << "SpeechSummary["
               << "speechEvents.size():" << obj.speechEvents.size() << ", "
               << "speechInfo:" << obj.speechInfo << "]";
            return os;
        }
 
        bool operator==(const SpeechSummary& other) const {
            return other.speechInfo == speechInfo
                && other.speechEvents == speechEvents;
        }
    };
 
    class VOICE_SDK_API SpeechSummaryStream {
    public:
 
        using Ptr = std::shared_ptr<SpeechSummaryStream>;
 
        virtual void reset() = 0;
 
        virtual void finalize() = 0;
 
        virtual bool hasSpeechEvents() const = 0;
 
        virtual SpeechEvent getSpeechEvent() = 0;
 
        virtual SpeechSummary getTotalSpeechSummary() const = 0;
 
        virtual SpeechInfo getTotalSpeechInfo() const = 0;
 
        virtual float getCurrentBackgroundLength() const = 0;
 
        virtual void addSamples(
            const uint8_t *bytes,
            size_t bytesNum) = 0;
 
        virtual void addSamples(
            const int16_t *pcm16Samples,
            size_t samplesNum) = 0;
 
        virtual void addSamples(
            const float *floatSamples,
            size_t samplesNum) = 0;
 
        virtual bool compare(const SpeechSummaryStream::Ptr& other) const = 0;
 
        virtual ~SpeechSummaryStream() = default;
    };
 
    class VOICE_SDK_API SpeechSummaryStreamOpus {
    public:
 
        using Ptr = std::shared_ptr<SpeechSummaryStreamOpus>;
 
        static SpeechSummaryStreamOpus::Ptr create(const std::string& initPath, size_t sampleRate);
 
        virtual void reset() = 0;
 
        virtual void finalize() = 0;
 
        virtual bool hasSpeechEvents() const = 0;
 
        virtual SpeechEvent getSpeechEvent() = 0;
 
        virtual SpeechSummary getTotalSpeechSummary() const = 0;
 
        virtual SpeechInfo getTotalSpeechInfo() const = 0;
 
        virtual float getCurrentBackgroundLength() const = 0;
 
        virtual void addPacket(
                const uint8_t* bytes,
                size_t bytesNum) = 0;
 
        virtual bool compare(const SpeechSummaryStreamOpus::Ptr& other) const = 0;
 
        virtual ~SpeechSummaryStreamOpus() = default;
    };
 
    class VOICE_SDK_API SpeechSummaryEngine {
    public:
 
        using Ptr = std::shared_ptr<SpeechSummaryEngine>;
 
        static SpeechSummaryEngine::Ptr create(const std::string& initPath = "");
 
        virtual ~SpeechSummaryEngine() = default;
 
        virtual SpeechSummary getSpeechSummary(
            const uint8_t *bytes,
            size_t bytesNum,
            size_t sampleRate) = 0;
 
        virtual SpeechSummary getSpeechSummary(
            const int16_t *pcm16Samples,
            size_t samplesNum,
            size_t sampleRate) = 0;
 
        virtual SpeechSummary getSpeechSummary(
            const float *floatSamples,
            size_t samplesNum,
            size_t sampleRate) = 0;
 
        virtual SpeechSummary getSpeechSummary(
            const std::string& audioFile) = 0;
 
        /* Speech summary stream methods */
 
        virtual SpeechSummaryStream::Ptr createStream(int sampleRate) = 0;
    };
}