DATA03_training_sample_gen3_VAE_refine.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3a414f2d-8b54-4659-ab2b-b9ff5b5a4259",
   "metadata": {},
   "source": [
    "# Generate training samples for VQ-VAE\n",
    "\n",
    "This notebook geneerate training samples (batch size = 1) for VQ-VAE.\n",
    "* All samples, including testing set samples were generated.\n",
    "* Samples were saved as .npy files; their file names contain the datetime information, so the testing samplles will not be touched."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "df2ef1ec-9b90-4574-8ed9-806573e0ba76",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import time\n",
    "import h5py\n",
    "import numpy as np\n",
    "from glob import glob\n",
    "\n",
    "# ------------------------------------------------------- #\n",
    "# Turn-off warnings\n",
    "import logging\n",
    "import warnings\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' \n",
    "logging.getLogger(\"tensorflow\").setLevel(logging.ERROR)\n",
    "\n",
    "# ------------------------------------------------------- #\n",
    "# Turn-off tensoflow-specific warnings\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "\n",
    "tf.autograph.set_verbosity(0)\n",
    "tf.get_logger().setLevel('ERROR')\n",
    "\n",
    "# ------------------------------------------------------- #\n",
    "# Import customized modules and settings\n",
    "sys.path.insert(0, '/glade/u/home/ksha/GAN_proj/')\n",
    "sys.path.insert(0, '/glade/u/home/ksha/GAN_proj/libs/')\n",
    "\n",
    "from namelist import *\n",
    "import data_utils as du\n",
    "import model_utils as mu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ae6cd660-668e-4659-9aae-914824eb7a76",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "85218783-5506-4219-94ed-af61b0a36162",
   "metadata": {},
   "source": [
    "# VQ-VAE backbone"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f0b6209a-773e-4ff7-9f2f-39e23077308b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters\n",
    "filter_nums = [64, 128] # number of convolution kernels per down-/upsampling layer \n",
    "latent_dim = 4 # number of latent feature channels\n",
    "activation = 'gelu' # activation function\n",
    "num_embeddings = 128 #128 # number of the VQ codes\n",
    "\n",
    "input_size = (224, 464, 1) # size of MRMS input\n",
    "latent_size = (14, 29, latent_dim) # size of compressed latent features\n",
    "\n",
    "drop_encode = False\n",
    "drop_decode = False\n",
    "\n",
    "model_name_encoder_load = model_dir+'models/VQ_VAE_encoder_stack1_tune0'\n",
    "model_name_decoder_load = model_dir+'models/VQ_VAE_decoder_stack1_tune0'\n",
    "\n",
    "encoder = mu.VQ_VAE_encoder(input_size, filter_nums, latent_dim, num_embeddings, activation, drop_encode)\n",
    "\n",
    "W_old = mu.dummy_loader(model_name_encoder_load)\n",
    "encoder.set_weights(W_old)\n",
    "\n",
    "decoder = mu.VQ_VAE_decoder(latent_size, filter_nums, activation, drop_decode)\n",
    "\n",
    "W_old = mu.dummy_loader(model_name_decoder_load)\n",
    "decoder.set_weights(W_old)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "82bc09b3-157d-4512-ae51-499af903a542",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Connect the encoder and decoder\n",
    "X = keras.Input(shape=input_size)\n",
    "X_encode = encoder(X)\n",
    "X_decode = decoder(X_encode)\n",
    "VQ_VAE = keras.Model(X, X_decode)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9293af4e-95b2-4337-bce6-511c005b072e",
   "metadata": {},
   "source": [
    "## Generate training samples\n",
    "\n",
    "* Locations: `$camp_dir+BATCH_dscale/`\n",
    "* Normalization:\n",
    "```python\n",
    "apcp = np.log(0.1*apcp+1)\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3b6a3739-edc7-4610-a935-7e8de5dd04c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "def norm_precip(x):\n",
    "    return np.log(0.1*x+1)\n",
    "\n",
    "def norm_elev(x):\n",
    "    return x / np.nanmax(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6beb6d4e-c78d-4032-a926-4cb438cd1e82",
   "metadata": {},
   "outputs": [],
   "source": [
    "with h5py.File(save_dir+'CCPA_domain.hdf', 'r') as h5io:\n",
    "    land_mask_CCPA = h5io['land_mask_CCPA'][...]\n",
    "    elev_CCPA = h5io['elev_CCPA'][...]\n",
    "    \n",
    "land_mask = land_mask_CCPA == 1.0\n",
    "ocean_mask = land_mask_CCPA == 0.0\n",
    "grid_shape = land_mask.shape\n",
    "\n",
    "elev_CCPA[ocean_mask] = 0\n",
    "elev_CCPA[elev_CCPA<0] = 0\n",
    "elev_CCPA = norm_elev(elev_CCPA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ebaaa7de-5c5c-4177-972b-d874c6bbd5fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "with h5py.File(camp_dir+'CCPA/CCPA_CDFs_2002_2019.hdf', 'r') as h5io:\n",
    "    CCPA_CDFs = h5io['CCPA_CDFs'][...]\n",
    "CCPA_CDFs_99 = norm_precip(CCPA_CDFs[99, ...])\n",
    "CCPA_CDFs_99[ocean_mask, :] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "19db9787-024a-4f22-a379-23fc04b38838",
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_dir = camp_dir+'BATCH_dscale/'\n",
    "batch_name = 'CCPA_y{}_d{:03d}_h{}.npy' #.format(datetime, hour-of-day)\n",
    "\n",
    "filename_ccpa = camp_dir+'CCPA/CCPA_y{}.hdf'\n",
    "batch_save = np.empty(grid_shape+(4,))\n",
    "\n",
    "for year in range(2002, 2022):\n",
    "    print('Batch gen year: {}'.format(year))\n",
    "    \n",
    "    with h5py.File(filename_ccpa.format(year), 'r') as h5io:\n",
    "        CCPA = h5io['CCPA'][...]\n",
    "        \n",
    "    N_days = len(CCPA)\n",
    "    \n",
    "    for i6hour in range(4):\n",
    "        CCPA_input = norm_precip(CCPA[:, i6hour, ...])\n",
    "        CCPA_input[:, ocean_mask] = 0.0\n",
    "        \n",
    "        CCPA_VAE = VQ_VAE.predict(CCPA_input[..., None], verbose=0)[..., 0]\n",
    "        CCPA_VAE[CCPA_VAE<0] = 0\n",
    "        CCPA_VAE[:, ocean_mask] = 0\n",
    "        \n",
    "        CCPA_CDFs_99_ = CCPA_CDFs_99[..., i6hour]\n",
    "        for iday in range(N_days):\n",
    "            batch_save[...] = np.nan\n",
    "            batch_save[..., 0] = CCPA_VAE[iday, ...]\n",
    "            batch_save[..., 1] = elev_CCPA\n",
    "            batch_save[..., 2] = CCPA_CDFs_99_\n",
    "            batch_save[..., 3] = CCPA_input[iday, ...]\n",
    "\n",
    "            save_name = BATCH_dir+batch_name.format(year, iday, i6hour)\n",
    "            if np.sum(np.isnan(batch_save)) > 0:\n",
    "                print('NaNs --> {}'.format(save_name))\n",
    "            else:\n",
    "                # print(save_name)\n",
    "                np.save(save_name, batch_save)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e886453f-c301-4811-8c45-541edd74d8a9",
   "metadata": {},
   "source": [
    "### Check batch quality"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "d987dd32-34b0-46ac-aa91-9741e985c605",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.QuadMesh at 0x14ee0dee74f0>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.pcolormesh(batch_save[..., 3], cmap=plt.cm.nipy_spectral_r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf49046f-ad8e-4abd-ab64-fdf25209b2c3",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}