#include "ffcpp/ffcpp.h"
#include "ffcpp/MediaFile.h"
#include "ffcpp/FifoQueue.h"
#include "ffcpp/Scaler.h"
#include "ffcpp/Resampler.h"
//#include <iostream>

constexpr int VIDEO_STREAM_INDEX = 0;
constexpr int AUDIO_STREAM_INDEX = 1;

constexpr int VIDEO_WIDTH = 854;
constexpr int VIDEO_HEIGHT = 480;

namespace ff = ffcpp;

void flushEncoder(ff::MediaFile& file, ff::Codec& encoder, const ff::Stream& inStream, const ff::Stream& outStream,
                  int streamIndex, AVRational srcTimeBase, AVRational dstTimeBase) {
	if(encoder.capabilities() & AV_CODEC_CAP_DELAY) {
		while (1) {
			auto packet = encoder.encode(nullptr);
			if(!packet) break;

			packet.setStreamIndex(streamIndex);
			packet.rescaleTimestamps(srcTimeBase, dstTimeBase);
			file.writePacket(packet);
		}
	}
}

int main(int argc, char** argv) {
	ff::init();
	ff::MediaFile input(argv[1], ff::Mode::Read);
	ff::MediaFile output(argv[2], ff::Mode::Write);

	ff::Stream& vStream = input.videoStream();
	ff::Stream& aStream = input.audioStream();

	ff::Codec& vDecoder = vStream.codec();
	ff::Codec& aDecoder = aStream.codec();

	double aspect = 1.0*vDecoder.width()/vDecoder.height();
	int outHeight = (int)(VIDEO_WIDTH/aspect) & ~1;
	ff::Stream& outVStream = output.addVideoStream(AV_CODEC_ID_H264, VIDEO_WIDTH, outHeight, vDecoder.timeBase(), AV_PIX_FMT_YUV420P);
	ff::Codec& vEncoder = outVStream.codec();

	ff::Stream& outAStream = output.addAudioStream(AV_CODEC_ID_VORBIS, 2, 44100, AV_SAMPLE_FMT_FLTP);
	ff::Codec& aEncoder = outAStream.codec();

	output.writeHeader();

	int64_t aPts = 0;
	ff::FifoQueue fifo(aEncoder.sampleFormat(), aEncoder.channels(), aEncoder.frameSize());
	ff::Scaler scaler(vDecoder, vEncoder);
	ff::Resampler resampler(aDecoder, aEncoder);
	bool needScaling = ff::Scaler::needScaling(vDecoder, vEncoder);
	bool needResampling = ff::Resampler::needResampling(aDecoder, aEncoder);
	while(auto packet = input.readPacket()) {
		AVMediaType packetType = input.packetType(packet);
		if(packetType == AVMEDIA_TYPE_AUDIO) {
			auto frame = aDecoder.decode(packet);
			if(needResampling)
				frame = resampler.resample(frame);
			fifo.addSamples(frame);

			// FIXME: we're losing last samples in case when fifo queue isn't full enough for encoder
			while(fifo.enoughSamples()) {
				auto frame = aEncoder.createAudioFrame();
				fifo.readFrame(frame);
				frame.setPts(aPts);
				aPts += frame.samplesCount();
				auto encPacket = aEncoder.encode(frame);
				if(!encPacket) continue;
				encPacket.setStreamIndex(AUDIO_STREAM_INDEX);
				encPacket.rescaleTimestamps(aEncoder.timeBase(), outAStream.timeBase());
				output.writePacket(encPacket);
			}
		} else if(packetType == AVMEDIA_TYPE_VIDEO) {
			auto frame = vDecoder.decode(packet);
			if(needScaling)
				frame = scaler.scale(frame);
			frame.setPictureType(AV_PICTURE_TYPE_NONE);
			auto encPacket = vEncoder.encode(frame);
			if(!encPacket) continue;
			encPacket.setStreamIndex(VIDEO_STREAM_INDEX);
			encPacket.rescaleTimestamps(vStream.timeBase(), outVStream.timeBase());
			output.writePacket(encPacket);
		}
	}

	flushEncoder(output, vEncoder, vStream, outVStream, VIDEO_STREAM_INDEX, vStream.timeBase(), outVStream.timeBase());
	flushEncoder(output, aEncoder, aStream, outAStream, AUDIO_STREAM_INDEX, aEncoder.timeBase(), outAStream.timeBase());
	output.writeTrailer();

	return 0;
}