{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "from suno_utils.utils.text import read_jsonl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load codec for decoding\n",
    "from suno_utils.tasks.dac_vae_100hz_peaq import (  # NOTE: works for 25hz as well\n",
    "    preload_models as preload_codec_models,\n",
    "    decode as codec_decode,\n",
    "    encode as codec_encode,\n",
    "    get_embedding_rate,\n",
    "    load_model as load_codec_model,\n",
    ")\n",
    "CODEC_FILEPATH = \"s3://suno-data/christian/25hz_vae_peaq_kl_0.005.pth\"\n",
    "preload_codec_models(CODEC_FILEPATH)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the memmaps and check the shapes\n",
    "VAE_DIM = 128\n",
    "VAE_RATE_HZ = 25\n",
    "VAE_MEMMAP_SIZE = 750\n",
    "SEMANTIC_VOCAB_SIZE = 4000\n",
    "SEMANTIC_MEMMAP_SIZE = 750\n",
    "\n",
    "out_dir = \"/app/suno/data/diffusion_ft/interesting_clips_up_v1_20241118_full\"\n",
    "metas_tr = read_jsonl(os.path.join(out_dir, f\"metas_val.jsonl\"))\n",
    "print(len(metas_tr)/ 2)\n",
    "mm_semantic_tr = np.memmap(os.path.join(out_dir, f\"data_semantic_val.bin\"), dtype=np.uint16, mode=\"r\")\n",
    "mm_vae_tr = np.memmap(os.path.join(out_dir, f\"data_vae_val.bin\"), dtype=np.float16, mode=\"r\")\n",
    "\n",
    "\n",
    "mm_vae_tr = mm_vae_tr.reshape(-1, VAE_MEMMAP_SIZE, VAE_DIM)\n",
    "print(mm_vae_tr.shape)\n",
    "\n",
    "mm_semantic_tr = mm_semantic_tr.reshape(-1, SEMANTIC_MEMMAP_SIZE)\n",
    "print(mm_semantic_tr.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# decode some audio\n",
    "idx = 40\n",
    "# ensure even index\n",
    "assert idx % 2 == 0\n",
    "print(metas_tr[idx])\n",
    "print(\"positive\")\n",
    "audio = codec_decode(mm_vae_tr[idx])\n",
    "audio.normalize_volume().play()\n",
    "\n",
    "print(metas_tr[idx + 1])\n",
    "print(\"negative\")\n",
    "audio = codec_decode(mm_vae_tr[idx + 1])\n",
    "audio.normalize_volume().play()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "suno_env2",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}