#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
#include <sysexits.h>
#include <fcntl.h>
#include <math.h>

#include <xas/spatial.h>
#include <xas/synth.h>
#include <xas/vox.h>
#include <xas/audio.h>
#include <xas/riff.h>

static void usage(int argc, char **argv, const char *message, ...) {
    va_list args;

    va_start(args, message);

    if (message) {
        vfprintf(stderr, message, args);
        fputc('\n', stderr);
    }

    va_end(args);

    fprintf(stderr, "usage: %s output.wav\n", argv[0]);

    exit(EX_USAGE);
}

int main(int argc, char **argv) {
    xas_spatial_scene *scene;

    xas_synth *sine;

    xas_audio_stream *synth,
                       *voice,
                       *output,
                       *wave;

    xas_vox *vox;

    xas_audio_format format = {
        .channels    = XAS_AUDIO_STEREO,
        .sample_size = XAS_AUDIO_PCM_16_BIT,
        .sample_rate = 44100
    };

    size_t buffer_size = 735,
           duration_s  = 300,
           i;

    xas_spatial_coord speakers[2] = {
        { -0.09, 0.0, 0.0 },
        {  0.09, 0.0, 0.0 }
    };

    if (argc != 2) {
        usage(argc, argv, "No output file provided");
    }

    if ((wave = xas_riff_new_file(argv[1],
                                    format,
                                    O_WRONLY | O_CREAT | O_TRUNC)) == NULL) {
        goto error_riff_new_file;
    }

    if ((sine = xas_synth_new(format,
                                buffer_size,
                                XAS_SYNTH_SINE)) == NULL) {
        goto error_synth_new;
    }

    if ((synth = xas_synth_stream_new(sine)) == NULL) {
        goto error_synth_stream_new;
    }

    if ((vox = xas_vox_new(format,
                             buffer_size,
                             "/usr/bin/text2wave")) == NULL) {
        goto error_vox_new;
    }

    xas_vox_set_parameter_float(vox, "Duration_Stretch", 1.3);

    if ((voice = xas_vox_stream_new(vox)) == NULL) {
        goto error_vox_stream_new;
    }

    if ((scene = xas_spatial_scene_new(format,
                                         speakers[0],
                                         speakers[1])) == NULL) {
        goto error_spatial_scene_new;
    }

    if ((output = xas_spatial_scene_stream_new(scene,
                                                 buffer_size)) == NULL) {
        goto error_spatial_scene_stream_new;
    }

    if (xas_spatial_scene_add_object(scene,
                                       (xas_spatial_coord){ 5.2, 0.0, 0.0 },
                                       synth,
                                       sine) == NULL) {
        goto error_spatial_scene_add_object;
    }

    if (xas_spatial_scene_add_object(scene,
                                       (xas_spatial_coord){ -5.2, 0.0, 0.0 },
                                       voice,
                                       vox) == NULL) {
        goto error_spatial_scene_add_object;
    }

    xas_synth_set_frequency(sine, 2600);

    xas_vox_set_voice(vox, "voice_cmu_us_slt_cg");
    xas_synth_start(sine);

    xas_vox_sayf(vox, "I want to eat your soul.\n");
    xas_vox_sayf(vox, "You don't understand.\n");
    xas_vox_sayf(vox, "I really want to eat your soul.\n");

    xas_vox_generate(vox);

    for (i=0; i<duration_s; i++) {
        void *buf;
        ssize_t readlen;

        if ((readlen = xas_audio_stream_read(output,
                                               &buf,
                                               buffer_size)) < 0) {
            goto error_audio_stream_read;
        }

        if (xas_audio_stream_write(wave, buf, readlen) < 0) {
            goto error_audio_stream_write;
        }
    }

    xas_audio_stream_flush(wave);

    xas_audio_stream_destroy(output);
    xas_spatial_scene_destroy(scene);
    xas_audio_stream_destroy(voice);
    xas_vox_destroy(vox);
    xas_audio_stream_destroy(synth);
    xas_audio_stream_destroy(wave);

    return EX_OK;

error_audio_stream_read:
error_audio_stream_write:
error_spatial_scene_add_object:
    xas_audio_stream_destroy(output);

error_spatial_scene_stream_new:
    xas_spatial_scene_destroy(scene);

error_spatial_scene_new:
    xas_audio_stream_destroy(voice);

error_vox_stream_new:
    xas_vox_destroy(vox);

error_vox_new:
    xas_audio_stream_destroy(synth);

error_synth_stream_new:
    xas_synth_destroy(sine);

error_synth_new:
    xas_audio_stream_destroy(wave);

error_riff_new_file:
    return EX_OSERR;
}