make_datasets.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 356,
     "status": "ok",
     "timestamp": 1671436795770,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "AmcJTBKIQoIl",
    "outputId": "1ff1522b-4685-4f57-a899-6e26a4bd0a9e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python 3.7.15\r\n"
     ]
    }
   ],
   "source": [
    "!python3 --version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "executionInfo": {
     "elapsed": 5982,
     "status": "ok",
     "timestamp": 1671448127185,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "OPOF_CovQLHF"
   },
   "outputs": [],
   "source": [
    "# %load_ext lab_black\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "executionInfo": {
     "elapsed": 8,
     "status": "ok",
     "timestamp": 1671448127721,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "sC7xm9abQLJE"
   },
   "outputs": [],
   "source": [
    "import cv2\n",
    "from datetime import timedelta\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "# import odc.stac\n",
    "import pandas as pd\n",
    "from pathlib import Path\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from tqdm import tqdm\n",
    "import gc\n",
    "\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "executionInfo": {
     "elapsed": 9,
     "status": "ok",
     "timestamp": 1671448127724,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "_bCrWJ5xQmKS"
   },
   "outputs": [],
   "source": [
    "DATA_DIR = ''#Path.cwd().parent.resolve() / \"data/final/public\"\n",
    "# DATA_DIR = '/gdrive/My Drive/Colab Notebooks/HarmfulAlgalBloomDetection/'#Path.cwd().parent.resolve() / \"data/final/public\"\n",
    "# assert DATA_DIR.exists()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "executionInfo": {
     "elapsed": 77,
     "status": "ok",
     "timestamp": 1671448128303,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "8uHtQC0_Q3zM",
    "outputId": "8e081799-4518-46d6-9093-68266fd60820"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>date</th>\n",
       "      <th>split</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>aabm</td>\n",
       "      <td>39.080319</td>\n",
       "      <td>-86.430867</td>\n",
       "      <td>2018-05-14</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aabn</td>\n",
       "      <td>36.559700</td>\n",
       "      <td>-121.510000</td>\n",
       "      <td>2016-08-31</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>aacd</td>\n",
       "      <td>35.875083</td>\n",
       "      <td>-78.878434</td>\n",
       "      <td>2020-11-19</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>aaee</td>\n",
       "      <td>35.487000</td>\n",
       "      <td>-79.062133</td>\n",
       "      <td>2016-08-24</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>aaff</td>\n",
       "      <td>38.049471</td>\n",
       "      <td>-99.827001</td>\n",
       "      <td>2019-07-23</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    uid   latitude   longitude        date  split\n",
       "0  aabm  39.080319  -86.430867  2018-05-14  train\n",
       "1  aabn  36.559700 -121.510000  2016-08-31   test\n",
       "2  aacd  35.875083  -78.878434  2020-11-19  train\n",
       "3  aaee  35.487000  -79.062133  2016-08-24  train\n",
       "4  aaff  38.049471  -99.827001  2019-07-23  train"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata = pd.read_csv(DATA_DIR + \"metadata.csv\")\n",
    "metadata.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 73,
     "status": "ok",
     "timestamp": 1671448128305,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "V4XhdDcX1mDK",
    "outputId": "e5cf9bec-34d5-4470-b3c8-e961c7305ae9"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23570"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(metadata.uid.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 63,
     "status": "ok",
     "timestamp": 1671448128306,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "mD__UUQh1x2m",
    "outputId": "0ed231fd-023e-4545-c904-39bf22f2c00d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "12681"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(metadata.loc[:,['latitude','longitude' \t]].drop_duplicates())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 53,
     "status": "ok",
     "timestamp": 1671448128307,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "dwl0GLm02djH",
    "outputId": "3212a638-1e0b-4689-82f8-5eabe3a1c7e6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([   79,   106,   107, ..., 23565, 23566, 23569])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "duplicated_rows=np.where(metadata.loc[:,['latitude','longitude']].duplicated()==True)[0]\n",
    "duplicated_rows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 45,
     "status": "ok",
     "timestamp": 1671448128308,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "Mm-ufEvC2mi_",
    "outputId": "fc5c2a2d-ca3c-4b7e-d0fa-986fda8f3d22"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uid                acdj\n",
       "latitude        37.3906\n",
       "longitude      -120.958\n",
       "date         2015-04-14\n",
       "split             train\n",
       "Name: 79, dtype: object"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.iloc[duplicated_rows[0],:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "executionInfo": {
     "elapsed": 35,
     "status": "ok",
     "timestamp": 1671448128308,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "yEl6IpDf28jr"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "executionInfo": {
     "elapsed": 35,
     "status": "ok",
     "timestamp": 1671448128309,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "RH95IbIn2vtT",
    "outputId": "c76d671b-f663-488a-8436-9b3d304ed018"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>date</th>\n",
       "      <th>split</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>aaqc</td>\n",
       "      <td>37.3906</td>\n",
       "      <td>-120.958</td>\n",
       "      <td>2013-07-09</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>acdj</td>\n",
       "      <td>37.3906</td>\n",
       "      <td>-120.958</td>\n",
       "      <td>2015-04-14</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1701</th>\n",
       "      <td>bxeq</td>\n",
       "      <td>37.3906</td>\n",
       "      <td>-120.958</td>\n",
       "      <td>2019-09-10</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2302</th>\n",
       "      <td>conk</td>\n",
       "      <td>37.3906</td>\n",
       "      <td>-120.958</td>\n",
       "      <td>2019-12-04</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2409</th>\n",
       "      <td>crja</td>\n",
       "      <td>37.3906</td>\n",
       "      <td>-120.958</td>\n",
       "      <td>2021-07-13</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       uid  latitude  longitude        date  split\n",
       "22    aaqc   37.3906   -120.958  2013-07-09  train\n",
       "79    acdj   37.3906   -120.958  2015-04-14  train\n",
       "1701  bxeq   37.3906   -120.958  2019-09-10  train\n",
       "2302  conk   37.3906   -120.958  2019-12-04  train\n",
       "2409  crja   37.3906   -120.958  2021-07-13  train"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.loc[metadata['latitude']==metadata.latitude.iloc[duplicated_rows[0]],:].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 469,
     "status": "ok",
     "timestamp": 1671448128746,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "-z622nn3RLdV",
    "outputId": "24112676-8eb2-459f-e895-0cb98a892ed0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "train    17060\n",
       "test      6510\n",
       "Name: split, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.split.value_counts(dropna=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 143
    },
    "executionInfo": {
     "elapsed": 41,
     "status": "ok",
     "timestamp": 1671448130388,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "BjhVjWbkRa5N",
    "outputId": "f7d8bf97-9524-441d-e9e8-5a8f88c34c41"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>min_date</th>\n",
       "      <th>max_date</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>split</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <td>2013-01-08</td>\n",
       "      <td>2021-12-29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train</th>\n",
       "      <td>2013-01-04</td>\n",
       "      <td>2021-12-14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        min_date   max_date\n",
       "split                      \n",
       "test  2013-01-08 2021-12-29\n",
       "train 2013-01-04 2021-12-14"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# convert date to pd.datetime\n",
    "metadata.date = pd.to_datetime(metadata.date)\n",
    "\n",
    "# what is the date range?\n",
    "metadata.groupby(\"split\").agg(min_date=(\"date\", min), max_date=(\"date\", max))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 311
    },
    "executionInfo": {
     "elapsed": 39,
     "status": "ok",
     "timestamp": 1671448130390,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "IdKW5sW5Rgmn",
    "outputId": "d86e26de-ed95-46bb-d72a-04cb4190c3fd"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# what years are in the data?\n",
    "pd.crosstab(metadata.date.dt.year, metadata.split).plot(kind=\"bar\")\n",
    "plt.ylabel(\"Number of samples\")\n",
    "plt.xlabel(\"Year\")\n",
    "plt.title(\"Distribution of years in the data\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 35,
     "status": "ok",
     "timestamp": 1671448130390,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "hi93IqmGRoMH",
    "outputId": "f1756d72-a185-4b23-d123-4ebfa92ac1dc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "summer    10813\n",
       "spring     5045\n",
       "fall       4758\n",
       "winter     2954\n",
       "Name: season, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# what seasons are the data points from?\n",
    "\n",
    "metadata[\"season\"] = (\n",
    "    metadata.date.dt.month.replace([12, 1, 2], \"winter\")\n",
    "    .replace([3, 4, 5], \"spring\")\n",
    "    .replace([6, 7, 8], \"summer\")\n",
    "    .replace([9, 10, 11], \"fall\")\n",
    ")\n",
    "metadata.season.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "executionInfo": {
     "elapsed": 94,
     "status": "ok",
     "timestamp": 1671448134135,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "d3LNKOfBRuoA",
    "outputId": "59875884-1447-46ea-b1b3-df8e74d997b3"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>region</th>\n",
       "      <th>severity</th>\n",
       "      <th>density</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>aabm</td>\n",
       "      <td>midwest</td>\n",
       "      <td>1</td>\n",
       "      <td>585.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aacd</td>\n",
       "      <td>south</td>\n",
       "      <td>1</td>\n",
       "      <td>290.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>aaee</td>\n",
       "      <td>south</td>\n",
       "      <td>1</td>\n",
       "      <td>1614.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>aaff</td>\n",
       "      <td>midwest</td>\n",
       "      <td>3</td>\n",
       "      <td>111825.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>aafl</td>\n",
       "      <td>midwest</td>\n",
       "      <td>4</td>\n",
       "      <td>2017313.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    uid   region  severity    density\n",
       "0  aabm  midwest         1      585.0\n",
       "1  aacd    south         1      290.0\n",
       "2  aaee    south         1     1614.0\n",
       "3  aaff  midwest         3   111825.0\n",
       "4  aafl  midwest         4  2017313.0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels = pd.read_csv(DATA_DIR + \"train_labels.csv\")\n",
    "train_labels.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17060, 4)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "executionInfo": {
     "elapsed": 88,
     "status": "ok",
     "timestamp": 1671448134137,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "t3RXLgLeR5H3"
   },
   "outputs": [],
   "source": [
    "train_labels_and_metadata = train_labels.merge(\n",
    "    metadata, how=\"left\", left_on=\"uid\", right_on=\"uid\", validate=\"1:1\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "executionInfo": {
     "elapsed": 86,
     "status": "ok",
     "timestamp": 1671448134139,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "7h_9Ei0sR-W5",
    "outputId": "e745f945-3cd7-4921-87ac-23b5588f11bf"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Train labels severity level counts')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "severity_counts = (\n",
    "    train_labels.replace(\n",
    "        {\n",
    "            \"severity\": {\n",
    "                1: \"1 (<20,000)\",\n",
    "                2: \"2 (20,000-100,000)\",\n",
    "                3: \"3 (100,000 - 1,000,000)\",\n",
    "                4: \"4 (1,00,000 - 10,000,000)\",\n",
    "                5: \"5 (>10,000,00)\",\n",
    "            }\n",
    "        }\n",
    "    )\n",
    "    .severity.value_counts()\n",
    "    .sort_index(ascending=False)\n",
    ")\n",
    "plt.barh(severity_counts.index, severity_counts.values)\n",
    "plt.xlabel(\"Number of samples\")\n",
    "plt.ylabel(\"Severity (range in cells/mL)\")\n",
    "plt.title(\"Train labels severity level counts\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 74,
     "status": "ok",
     "timestamp": 1671448134141,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "h_-HXWylSAmg",
    "outputId": "6d3d5f91-72c0-4072-9084-1dab6b94cbd6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1.706000e+04\n",
       "mean     1.074537e+06\n",
       "std      6.836693e+06\n",
       "min      0.000000e+00\n",
       "25%      4.066000e+03\n",
       "50%      3.270975e+04\n",
       "75%      4.849192e+05\n",
       "max      8.046675e+08\n",
       "Name: density, dtype: float64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels.density.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 66,
     "status": "ok",
     "timestamp": 1671448134142,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "LDSjxLrSSC6X",
    "outputId": "1196413b-71ea-4d9c-ac48-8bb1ffdb3d74"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "91"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(train_labels.density == 0).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((23570, 6), (17060, 4))"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.shape,train_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ADD train labels to metadata\n",
    "\n",
    "metadata=metadata.merge(train_labels,on='uid',how='left')\n",
    "metadata.loc[:,['severity','density']]=metadata.loc[:,['severity','density']].fillna(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>date</th>\n",
       "      <th>split</th>\n",
       "      <th>season</th>\n",
       "      <th>region</th>\n",
       "      <th>severity</th>\n",
       "      <th>density</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>aabm</td>\n",
       "      <td>39.080319</td>\n",
       "      <td>-86.430867</td>\n",
       "      <td>2018-05-14</td>\n",
       "      <td>train</td>\n",
       "      <td>spring</td>\n",
       "      <td>midwest</td>\n",
       "      <td>1.0</td>\n",
       "      <td>585.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aabn</td>\n",
       "      <td>36.559700</td>\n",
       "      <td>-121.510000</td>\n",
       "      <td>2016-08-31</td>\n",
       "      <td>test</td>\n",
       "      <td>summer</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>aacd</td>\n",
       "      <td>35.875083</td>\n",
       "      <td>-78.878434</td>\n",
       "      <td>2020-11-19</td>\n",
       "      <td>train</td>\n",
       "      <td>fall</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>290.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>aaee</td>\n",
       "      <td>35.487000</td>\n",
       "      <td>-79.062133</td>\n",
       "      <td>2016-08-24</td>\n",
       "      <td>train</td>\n",
       "      <td>summer</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1614.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>aaff</td>\n",
       "      <td>38.049471</td>\n",
       "      <td>-99.827001</td>\n",
       "      <td>2019-07-23</td>\n",
       "      <td>train</td>\n",
       "      <td>summer</td>\n",
       "      <td>midwest</td>\n",
       "      <td>3.0</td>\n",
       "      <td>111825.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23565</th>\n",
       "      <td>zzvv</td>\n",
       "      <td>36.708500</td>\n",
       "      <td>-121.749000</td>\n",
       "      <td>2014-12-02</td>\n",
       "      <td>test</td>\n",
       "      <td>winter</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23566</th>\n",
       "      <td>zzwo</td>\n",
       "      <td>39.792190</td>\n",
       "      <td>-99.971050</td>\n",
       "      <td>2017-06-19</td>\n",
       "      <td>train</td>\n",
       "      <td>summer</td>\n",
       "      <td>midwest</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48510.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23567</th>\n",
       "      <td>zzwq</td>\n",
       "      <td>35.794000</td>\n",
       "      <td>-79.012551</td>\n",
       "      <td>2015-03-24</td>\n",
       "      <td>train</td>\n",
       "      <td>spring</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1271.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23568</th>\n",
       "      <td>zzyb</td>\n",
       "      <td>35.742000</td>\n",
       "      <td>-79.238600</td>\n",
       "      <td>2016-11-21</td>\n",
       "      <td>train</td>\n",
       "      <td>fall</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9682.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23569</th>\n",
       "      <td>zzzi</td>\n",
       "      <td>39.767323</td>\n",
       "      <td>-96.028617</td>\n",
       "      <td>2015-08-31</td>\n",
       "      <td>test</td>\n",
       "      <td>summer</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23570 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        uid   latitude   longitude       date  split  season   region  \\\n",
       "0      aabm  39.080319  -86.430867 2018-05-14  train  spring  midwest   \n",
       "1      aabn  36.559700 -121.510000 2016-08-31   test  summer      NaN   \n",
       "2      aacd  35.875083  -78.878434 2020-11-19  train    fall    south   \n",
       "3      aaee  35.487000  -79.062133 2016-08-24  train  summer    south   \n",
       "4      aaff  38.049471  -99.827001 2019-07-23  train  summer  midwest   \n",
       "...     ...        ...         ...        ...    ...     ...      ...   \n",
       "23565  zzvv  36.708500 -121.749000 2014-12-02   test  winter      NaN   \n",
       "23566  zzwo  39.792190  -99.971050 2017-06-19  train  summer  midwest   \n",
       "23567  zzwq  35.794000  -79.012551 2015-03-24  train  spring    south   \n",
       "23568  zzyb  35.742000  -79.238600 2016-11-21  train    fall    south   \n",
       "23569  zzzi  39.767323  -96.028617 2015-08-31   test  summer      NaN   \n",
       "\n",
       "       severity   density  \n",
       "0           1.0     585.0  \n",
       "1          -1.0      -1.0  \n",
       "2           1.0     290.0  \n",
       "3           1.0    1614.0  \n",
       "4           3.0  111825.0  \n",
       "...         ...       ...  \n",
       "23565      -1.0      -1.0  \n",
       "23566       2.0   48510.0  \n",
       "23567       1.0    1271.0  \n",
       "23568       1.0    9682.0  \n",
       "23569      -1.0      -1.0  \n",
       "\n",
       "[23570 rows x 9 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 238
    },
    "executionInfo": {
     "elapsed": 56,
     "status": "ok",
     "timestamp": 1671448134143,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "dQEVMT1CSEno",
    "outputId": "5313ed0e-6c44-4107-d534-6bfd10273fa4"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>region</th>\n",
       "      <th>severity</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>uid</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>aabn</th>\n",
       "      <td>west</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aair</th>\n",
       "      <td>west</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aajw</th>\n",
       "      <td>northeast</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aalr</th>\n",
       "      <td>midwest</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aalw</th>\n",
       "      <td>west</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         region  severity\n",
       "uid                      \n",
       "aabn       west         1\n",
       "aair       west         1\n",
       "aajw  northeast         1\n",
       "aalr    midwest         1\n",
       "aalw       west         1"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_format = pd.read_csv(DATA_DIR + \"submission_format.csv\", index_col=0)\n",
    "submission_format.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 54,
     "status": "ok",
     "timestamp": 1671448134145,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "BbzZohUOSGf3",
    "outputId": "ff1a2174-b0e3-4f82-8340-5d923af9489a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6510, 2)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_format.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(23570, 9)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(submission_format)+len(train_labels)-len(metadata)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 527,
     "status": "ok",
     "timestamp": 1671448134628,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "wmRXPRsiSLHB",
    "outputId": "6022fe39-47c7-4d6a-d562-d1fc14e3167d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uid                         garm\n",
       "latitude                41.98006\n",
       "longitude             -110.65734\n",
       "date         2021-09-27 00:00:00\n",
       "split                      train\n",
       "season                      fall\n",
       "region                      west\n",
       "severity                     5.0\n",
       "density              12267752.74\n",
       "Name: 5461, dtype: object"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example_row = metadata[metadata.uid == \"garm\"].iloc[0]\n",
    "example_row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def image_to_features(image_array):\n",
    "    \"\"\"\n",
    "    Convert an image array of the form (color band, height, width) to a\n",
    "    1-dimensional list of features. Returns a list where the first three\n",
    "    values are the averages of each color band, and the second three\n",
    "    values are the medians of each color band.\n",
    "    \"\"\"\n",
    "    averages = image_array.mean(axis=(1, 2)).tolist()\n",
    "    medians = np.median(image_array, axis=(1, 2)).tolist()\n",
    "    stds = np.std(image_array, axis=(1, 2)).tolist()\n",
    "\n",
    "    # c0 = int(image_array.shape[1]/2)\n",
    "    # c1 = int(image_array.shape[1]/2)\n",
    "    # averages_zoomin = image_array[:,c0-1:c0+1,c1-1:c1+1].mean(axis=(1, 2)).tolist()\n",
    "    # medians_zoomin = np.median(image_array[:,c0-1:c0+1,c1-1:c1+1], axis=(1, 2)).tolist()\n",
    "\n",
    "    return averages + medians + stds #+averages_zoomin + medians_zoomin"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 13,
     "status": "ok",
     "timestamp": 1671448166552,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "77GIYd9DT_5b",
    "outputId": "c35a45e7-7216-4ec4-e082-dc397c853b24"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "train    17060\n",
       "test      6510\n",
       "Name: split, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# take a random subset of the training data for the benchmark\n",
    "# train_subset = metadata[metadata[\"split\"] == \"train\"].sample(n=2500, random_state=2)\n",
    "train_subset = metadata[metadata[\"split\"] == \"train\"]\n",
    "\n",
    "# combine train subset with all test data\n",
    "metadata_subset = pd.concat([train_subset, metadata[metadata[\"split\"] == \"test\"]])\n",
    "metadata_subset.split.value_counts(dropna=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "executionInfo": {
     "elapsed": 12,
     "status": "ok",
     "timestamp": 1671448166554,
     "user": {
      "displayName": "Ioannis Nasios",
      "userId": "11761353373054103725"
     },
     "user_tz": -120
    },
    "id": "9YwedsccBwEi"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.abs(altDF.altitude-altDF.altitude.astype(int)).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to be used in group kfold split\n",
    "\n",
    "metadata['latlong']=metadata['latitude'].round(2).astype(str)+'_'+metadata['longitude'].round(2).astype(str)\n",
    "metadata['latlong1']=metadata['latitude'].round(1).astype(str)+'_'+metadata['longitude'].round(1).astype(str)\n",
    "metadata['latlong0']=metadata['latitude'].round(0).astype(str)+'_'+metadata['longitude'].round(0).astype(str)\n",
    "metadata['latlong10']=metadata['latitude'].round(1).astype(str)+'_'+metadata['longitude'].round(0).astype(str)\n",
    "unique_points=np.unique(metadata.latlong)\n",
    "unique_points1=np.unique(metadata.latlong1)\n",
    "unique_points0=np.unique(metadata.latlong0)\n",
    "unique_points10=np.unique(metadata.latlong10)\n",
    "# metadata.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# metadata.shape,altDF.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(23556, 'oeuw_alt123.jpg')"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#GET geomorhological features\n",
    "DEM_DATA_DIR = \"benchmark/dem/\"#DATA_DIR.parents[1] / \"benchmark\"\n",
    "DEM_DATA_DIR = \"benchmark/dem2K/\"#DATA_DIR.parents[1] / \"benchmark\"\n",
    "geomorthology_files=os.listdir(DEM_DATA_DIR)\n",
    "len(geomorthology_files),geomorthology_files[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 23556/23556 [02:47<00:00, 140.54it/s]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "# RESIZE_SIZE=128\n",
    "DEM_RESIZE_SIZE=32\n",
    "dem_featuresL=[]\n",
    "rgbDEM=[]\n",
    "for f in tqdm(geomorthology_files):\n",
    "# for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    im = cv2.imread(DEM_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "\n",
    "    dem_features = image_to_features(np.expand_dims(im,0))\n",
    "    dem_featuresL.append(dem_features)\n",
    "    if im.max()!=im.min():\n",
    "        im=(im-im.min())/(im.max()-im.min())\n",
    "    else:\n",
    "        im*=0\n",
    "    im=cv2.resize(im,(DEM_RESIZE_SIZE, DEM_RESIZE_SIZE))\n",
    "    rgbDEM.append((im*255).astype('uint8'))\n",
    "        \n",
    "rgbDEM=np.stack(rgbDEM)    \n",
    "dem_featuresL=np.stack(dem_featuresL)    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed('dem2K', a=rgbDEM, b=dem_featuresL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(['oeuw', 'tjhp', 'crel', 'zqhi', 'estg'], [123, 19, 8, 85, -4])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dem_featuresL.shape\n",
    "geo_uids=[x.split('_')[0] for x in geomorthology_files]\n",
    "geo_alts=[int(x.split('_')[1].split('.')[0][3:]) for x in geomorthology_files]\n",
    "geo_uids[:5],geo_alts[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(rgbDEM[101])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(23556, 32, 32)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rgbDEM.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>altitude</th>\n",
       "      <th>DEMmean</th>\n",
       "      <th>DEMmedian</th>\n",
       "      <th>DEMstd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>oeuw</td>\n",
       "      <td>123</td>\n",
       "      <td>108.174279</td>\n",
       "      <td>104.5</td>\n",
       "      <td>47.364784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>tjhp</td>\n",
       "      <td>19</td>\n",
       "      <td>107.601538</td>\n",
       "      <td>103.0</td>\n",
       "      <td>33.233621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>crel</td>\n",
       "      <td>8</td>\n",
       "      <td>62.676179</td>\n",
       "      <td>40.0</td>\n",
       "      <td>64.743936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>zqhi</td>\n",
       "      <td>85</td>\n",
       "      <td>109.822581</td>\n",
       "      <td>105.0</td>\n",
       "      <td>54.010357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>estg</td>\n",
       "      <td>-4</td>\n",
       "      <td>62.576923</td>\n",
       "      <td>46.0</td>\n",
       "      <td>50.925415</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    uid  altitude     DEMmean  DEMmedian     DEMstd\n",
       "0  oeuw       123  108.174279      104.5  47.364784\n",
       "1  tjhp        19  107.601538      103.0  33.233621\n",
       "2  crel         8   62.676179       40.0  64.743936\n",
       "3  zqhi        85  109.822581      105.0  54.010357\n",
       "4  estg        -4   62.576923       46.0  50.925415"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# altDF=pd.DataFrame(np.stack([[('_').join(x.split('_')[:2]),float(x.split('_')[-1][3:-4])] for x in geomorthology_files]),columns=['latlong','altitude'])\n",
    "# altDF['altitude']=altDF['altitude'].astype(float)\n",
    "altDF=pd.DataFrame(geo_uids,columns=['uid'])\n",
    "altDF['altitude']=geo_alts\n",
    "altDF['altitude']=altDF['altitude'].fillna(altDF['altitude'].median)\n",
    "\n",
    "altDF['DEMmean']=dem_featuresL[:,0]\n",
    "altDF['DEMmedian']=dem_featuresL[:,1]\n",
    "altDF['DEMstd']=dem_featuresL[:,2]\n",
    "altDF.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26, 32)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv2.imread(DEM_DATA_DIR+f, cv2.IMREAD_GRAYSCALE).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# CLIMATE_DATA_DIR = \"climate/\"\n",
    "# CLIMATE_DATA_DIR = \"climate_again/\"\n",
    "# CLIMATE_DATA_DIR = \"climate_1/\"\n",
    "# CLIMATE_DATA_DIR = \"climate_14/\"\n",
    "# CLIMATE_DATA_DIR = \"climate_14b/\"\n",
    "# CLIMATE_DATA_DIR = \"climate_d14_b5000/\"\n",
    "# climate_files=os.listdir(CLIMATE_DATA_DIR)\n",
    "# CLIM_RESIZE_SIZE=3\n",
    "# clim_featuresL=[]\n",
    "# rgbCLIM=[]\n",
    "# climIds=[]\n",
    "# for f in tqdm(climate_files):\n",
    "#     try:\n",
    "        \n",
    "#     # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "#     #     im = cv2.imread(CLIMATE_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "#         im = np.load(CLIMATE_DATA_DIR+f)['a']\n",
    "#         notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "#         if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "\n",
    "#     #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "#             clim_features = image_to_features(im2)\n",
    "#     #         clim_features = image_to_features(im)\n",
    "#             clim_featuresL.append(clim_features)\n",
    "#         #     if im.max()!=im.min():\n",
    "#         #         im=(im-im.min())/(im.max()-im.min())\n",
    "#         #     else:\n",
    "#         #         im*=0\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "#             climIds.append(f.split('.')[0])\n",
    "#     except:\n",
    "#         pass\n",
    "        \n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "# clim_featuresL=np.stack(clim_featuresL)    \n",
    "# climIds=np.stack(climIds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clim_featuresL.shape,rgbCLIM.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clim_featuresLw1=[]\n",
    "# climIdsw1=[]\n",
    "# for f in tqdm(climate_files):\n",
    "#     try:\n",
    "        \n",
    "#     # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "#     #     im = cv2.imread(CLIMATE_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "#         im = np.load(CLIMATE_DATA_DIR+f)['a']\n",
    "#         im=im[:7,...] #only first weak\n",
    "#         notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "#         if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "\n",
    "#     #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "#             clim_features = image_to_features(im2)\n",
    "#     #         clim_features = image_to_features(im)\n",
    "#             clim_featuresLw1.append(clim_features)\n",
    "\n",
    "\n",
    "#             climIdsw1.append(f.split('.')[0])\n",
    "#     except:\n",
    "#         pass\n",
    "    \n",
    "# clim_featuresLw1=np.stack(clim_featuresLw1)    \n",
    "# climIdsw1=np.stack(climIdsw1)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clim_featuresLw1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLIMATE_DATA_DIR = \"climate_d14_b5000/\"\n",
    "CLIMATE_DATA_DIR = \"climate_d28_b5000/\"\n",
    "climate_files=os.listdir(CLIMATE_DATA_DIR)\n",
    "CLIM_RESIZE_SIZE=3\n",
    "clim_featuresL=[]\n",
    "# rgbCLIM=[]\n",
    "climSeries=[]\n",
    "climIds=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_DATA_DIR+f)['a']\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "    #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "\n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresL.append(clim_features)\n",
    "        #     if im.max()!=im.min():\n",
    "        #         im=(im-im.min())/(im.max()-im.min())\n",
    "        #     else:\n",
    "        #         im*=0\n",
    "            climSeries.append(im2[0,0,:])\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "            climIds.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "        \n",
    "climSeries=np.stack(climSeries)\n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "clim_featuresL=np.stack(clim_featuresL)    \n",
    "climIds=np.stack(climIds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# climSeries.shape,clim_featuresL.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.plot(climSeries[10],'.-')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "clim_featuresLw1=[]\n",
    "climIdsw1=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_DATA_DIR+f)['a']\n",
    "        im=im[:7,...] #only first weak\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "            \n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresLw1.append(clim_features)\n",
    "\n",
    "\n",
    "            climIdsw1.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "    \n",
    "clim_featuresLw1=np.stack(clim_featuresLw1)    \n",
    "climIdsw1=np.stack(climIdsw1)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7383"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLIMATE_R_DATA_DIR = \"climate_rain_d28_b5000/\"\n",
    "climate_files=os.listdir(CLIMATE_R_DATA_DIR)\n",
    "CLIM_RESIZE_SIZE=3\n",
    "clim_featuresR=[]\n",
    "# rgbCLIM=[]\n",
    "climSeriesR=[]\n",
    "climIdsR=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_R_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_R_DATA_DIR+f)['a']\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "    #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "\n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresR.append(clim_features)\n",
    "        #     if im.max()!=im.min():\n",
    "        #         im=(im-im.min())/(im.max()-im.min())\n",
    "        #     else:\n",
    "        #         im*=0\n",
    "            climSeriesR.append(im2[0,0,:])\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "            climIdsR.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "        \n",
    "climSeriesR=np.stack(climSeriesR)\n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "clim_featuresR=np.stack(clim_featuresR)    \n",
    "climIdsR=np.stack(climIdsR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.plot(climSeriesR[100],'.-')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLIMATE_G_DATA_DIR = \"climate_gust_d28_b5000/\"\n",
    "climate_files=os.listdir(CLIMATE_G_DATA_DIR)\n",
    "CLIM_RESIZE_SIZE=3\n",
    "clim_featuresG=[]\n",
    "# rgbCLIM=[]\n",
    "climSeriesG=[]\n",
    "climIdsG=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_G_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_G_DATA_DIR+f)['a']\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "    #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "\n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresG.append(clim_features)\n",
    "        #     if im.max()!=im.min():\n",
    "        #         im=(im-im.min())/(im.max()-im.min())\n",
    "        #     else:\n",
    "        #         im*=0\n",
    "            climSeriesG.append(im2[0,0,:])\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "            climIdsG.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "        \n",
    "climSeriesG=np.stack(climSeriesG)\n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "clim_featuresG=np.stack(clim_featuresG)    \n",
    "climIdsG=np.stack(climIdsG)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.plot(climSeriesG[100],'.-')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLIMATE_S_DATA_DIR = \"climate_snowc_d28_b15000/\"\n",
    "climate_files=os.listdir(CLIMATE_S_DATA_DIR)\n",
    "CLIM_RESIZE_SIZE=9\n",
    "clim_featuresS=[]\n",
    "# rgbCLIM=[]\n",
    "climSeriesS=[]\n",
    "climIdsS=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_G_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_S_DATA_DIR+f)['a']\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "    #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "\n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresS.append(clim_features)\n",
    "        #     if im.max()!=im.min():\n",
    "        #         im=(im-im.min())/(im.max()-im.min())\n",
    "        #     else:\n",
    "        #         im*=0\n",
    "            climSeriesS.append(im2[0,0,:])\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "            climIdsS.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "        \n",
    "climSeriesS=np.stack(climSeriesS)\n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "clim_featuresS=np.stack(clim_featuresS)    \n",
    "climIdsS=np.stack(climIdsS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLIMATE_H_DATA_DIR = \"climate_hgt_d28_b15000/\"\n",
    "climate_files=os.listdir(CLIMATE_H_DATA_DIR)\n",
    "CLIM_RESIZE_SIZE=9\n",
    "clim_featuresH=[]\n",
    "# rgbCLIM=[]\n",
    "climSeriesH=[]\n",
    "climIdsH=[]\n",
    "for f in tqdm(climate_files):\n",
    "    try:\n",
    "        \n",
    "    # for row in tqdm(metadata_subset.iloc[:2,:].itertuples(), total=2):\n",
    "    #     im = cv2.imread(CLIMATE_G_DATA_DIR+f, cv2.IMREAD_GRAYSCALE)\n",
    "        im = np.load(CLIMATE_H_DATA_DIR+f)['a']\n",
    "        notnans = np.where(im[:,0,0]!=-999)[0]\n",
    "        if len(notnans)!=0:\n",
    "#             im=np.transpose(im,(1,2,0))\n",
    "#             im=np.expand_dims(im,0)\n",
    "    #         clim_features = image_to_features(np.expand_dims(im,0))\n",
    "#             im2=im.copy()\n",
    "#             im2[im2==-999]= im2[im2!=-999].mean()\n",
    "#             im2=np.mean(im2,-1)\n",
    "            \n",
    "            im2=np.mean(im,(1,2))\n",
    "            im2[im2==-999]=np.nan\n",
    "            # d14[d14==-999]=np.mean(d14[d14!=-999])\n",
    "            im2=(pd.Series(im2).ffill()+pd.Series(im2).bfill())/2\n",
    "            im2=pd.Series(im2).ffill()\n",
    "            im2= pd.Series(im2).bfill().values\n",
    "            im2 = np.reshape(im2,(1,1,len(im2)))\n",
    "\n",
    "            clim_features = image_to_features(im2)\n",
    "    #         clim_features = image_to_features(im)\n",
    "            clim_featuresH.append(clim_features)\n",
    "        #     if im.max()!=im.min():\n",
    "        #         im=(im-im.min())/(im.max()-im.min())\n",
    "        #     else:\n",
    "        #         im*=0\n",
    "            climSeriesH.append(im2[0,0,:])\n",
    "#             im=np.expand_dims(im[0,...,notnans[0]], -1)\n",
    "#             im=cv2.resize(im.astype(np.float32),(CLIM_RESIZE_SIZE, CLIM_RESIZE_SIZE))\n",
    "#         #     rgbCLIM.append((im*255).astype('uint8'))\n",
    "#             rgbCLIM.append(im.astype(int))\n",
    "\n",
    "            climIdsH.append(f.split('.')[0])\n",
    "    except:\n",
    "        pass\n",
    "        \n",
    "climSeriesH=np.stack(climSeriesH)\n",
    "# rgbCLIM=np.stack(rgbCLIM)    \n",
    "clim_featuresH=np.stack(clim_featuresH)    \n",
    "climIdsH=np.stack(climIdsH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed(CLIMATE_DATA_DIR[:-1], a=climSeries, b=clim_featuresL, c=climIds)\n",
    "np.savez_compressed(CLIMATE_DATA_DIR[:-1]+'w1',  b=clim_featuresLw1, c=climIdsw1)\n",
    "np.savez_compressed(CLIMATE_R_DATA_DIR[:-1], a=climSeriesR, b=clim_featuresR, c=climIdsR)\n",
    "np.savez_compressed(CLIMATE_G_DATA_DIR[:-1], a=climSeriesG, b=clim_featuresG, c=climIdsG)\n",
    "np.savez_compressed(CLIMATE_S_DATA_DIR[:-1], a=climSeriesS, b=clim_featuresS, c=climIdsS)\n",
    "np.savez_compressed(CLIMATE_H_DATA_DIR[:-1], a=climSeriesH, b=clim_featuresH, c=climIdsH)\n",
    "\n",
    "temp=np.load('climate_d28_b5000.npz')\n",
    "climSeries=temp['a']\n",
    "clim_featuresL=temp['b']\n",
    "climIds=temp['c']\n",
    "\n",
    "temp=np.load('climate_d28_b5000w1.npz')\n",
    "clim_featuresLw1=temp['b']\n",
    "climIdsw1=temp['c']\n",
    "\n",
    "temp=np.load('climate_rain_d28_b5000.npz')\n",
    "climSeriesR=temp['a']\n",
    "clim_featuresR=temp['b']\n",
    "climIdsR=temp['c']\n",
    "\n",
    "temp=np.load('climate_gust_d28_b5000.npz')\n",
    "climSeriesG=temp['a']\n",
    "clim_featuresG=temp['b']\n",
    "climIdsG=temp['c']\n",
    "\n",
    "temp=np.load('climate_snowc_d28_b15000.npz')\n",
    "climSeriesS=temp['a']\n",
    "clim_featuresS=temp['b']\n",
    "climIdsS=temp['c']\n",
    "\n",
    "temp=np.load('climate_hgt_d28_b15000.npz')\n",
    "climSeriesH=temp['a']\n",
    "clim_featuresH=temp['b']\n",
    "climIdsH=temp['c']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((20502, 28), (20502, 28), (20502, 28), (20547, 28), (20547, 28))"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "climSeries.shape, climSeriesR.shape, climSeriesG.shape, climSeriesS.shape, climSeriesH.shape,"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# climIdsSeries=np.concatenate((np.expand_dims(climIds,1),climSeries),1)\n",
    "# climIdsSeriesR=np.concatenate((np.expand_dims(climIdsR,1),climSeriesR),1)\n",
    "# climIdsSeriesG=np.concatenate((np.expand_dims(climIdsG,1),climSeriesG),1)\n",
    "# climIdsSeriesS=np.concatenate((np.expand_dims(climIdsS,1),climSeriesS),1)\n",
    "# climIdsSeriesH=np.concatenate((np.expand_dims(climIdsH,1),climSeriesH),1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "climIdsSeries=pd.DataFrame(np.concatenate((np.expand_dims(climIds,1),climSeries),1), columns=['uid']+['Temp'+str(i) for i in range(28)])\n",
    "\n",
    "climIdsSeriesR=pd.DataFrame(np.concatenate((np.expand_dims(climIdsR,1),climSeriesR),1), columns=['uid']+['Rain'+str(i) for i in range(28)])\n",
    "\n",
    "climIdsSeriesG=pd.DataFrame(np.concatenate((np.expand_dims(climIdsG,1),climSeriesG),1), columns=['uid']+['Gust'+str(i) for i in range(28)])\n",
    "\n",
    "climIdsSeriesS=pd.DataFrame(np.concatenate((np.expand_dims(climIdsS,1),climSeriesS),1), columns=['uid']+['Snowc'+str(i) for i in range(28)])\n",
    "\n",
    "climIdsSeriesH=pd.DataFrame(np.concatenate((np.expand_dims(climIdsH,1),climSeriesH),1), columns=['uid']+['HGT'+str(i) for i in range(28)])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rma,rmi = rgbCLIM.max(),rgbCLIM.min()\n",
    "# rma,rmi "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rgbCLIM=(((rgbCLIM.astype(float)-rmi)/(rma-rmi))*255).astype('uint8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# # rgbCLIM[0][:,:,[0,15,29]].astype('uint8').max()\n",
    "# rgbCLIM.max(),rgbCLIM.min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # plt.imshow(rgbCLIM[20][:,:,0].astype('uint8'))\n",
    "# plt.imshow(rgbCLIM[21].astype('uint8'))\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rgbCLIM.shape,rgbCLIM[1].shape,np.transpose(np.load(CLIMATE_DATA_DIR+f)['a'],(1,2,0)).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((20502,), (20502, 3))"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "climIds.shape,clim_featuresL.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clim_featuresL=np.squeeze(clim_featuresL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>climate_mean</th>\n",
       "      <th>climate_median</th>\n",
       "      <th>climate_std</th>\n",
       "      <th>uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15.510000</td>\n",
       "      <td>14.72</td>\n",
       "      <td>2.891339</td>\n",
       "      <td>waln</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>19.679286</td>\n",
       "      <td>19.53</td>\n",
       "      <td>3.257052</td>\n",
       "      <td>wlwd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>23.263571</td>\n",
       "      <td>23.54</td>\n",
       "      <td>2.815118</td>\n",
       "      <td>zxdf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.067143</td>\n",
       "      <td>11.58</td>\n",
       "      <td>2.001658</td>\n",
       "      <td>rzcf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>27.039286</td>\n",
       "      <td>27.98</td>\n",
       "      <td>3.080568</td>\n",
       "      <td>djsa</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   climate_mean  climate_median  climate_std   uid\n",
       "0     15.510000           14.72     2.891339  waln\n",
       "1     19.679286           19.53     3.257052  wlwd\n",
       "2     23.263571           23.54     2.815118  zxdf\n",
       "3     11.067143           11.58     2.001658  rzcf\n",
       "4     27.039286           27.98     3.080568  djsa"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresDF=pd.concat([pd.DataFrame(clim_featuresL,columns=['climate_mean','climate_median','climate_std']),\n",
    "                           pd.DataFrame(climIds,columns=['uid'])],1)\n",
    "clim_featuresDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>climate_mean_fw</th>\n",
       "      <th>climate_median_fw</th>\n",
       "      <th>climate_std_fw</th>\n",
       "      <th>uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15.577143</td>\n",
       "      <td>14.40</td>\n",
       "      <td>3.785377</td>\n",
       "      <td>waln</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18.914286</td>\n",
       "      <td>18.52</td>\n",
       "      <td>3.163031</td>\n",
       "      <td>wlwd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20.622857</td>\n",
       "      <td>20.52</td>\n",
       "      <td>2.451622</td>\n",
       "      <td>zxdf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.360000</td>\n",
       "      <td>11.72</td>\n",
       "      <td>2.610134</td>\n",
       "      <td>rzcf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>26.665714</td>\n",
       "      <td>26.48</td>\n",
       "      <td>1.777662</td>\n",
       "      <td>djsa</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   climate_mean_fw  climate_median_fw  climate_std_fw   uid\n",
       "0        15.577143              14.40        3.785377  waln\n",
       "1        18.914286              18.52        3.163031  wlwd\n",
       "2        20.622857              20.52        2.451622  zxdf\n",
       "3        11.360000              11.72        2.610134  rzcf\n",
       "4        26.665714              26.48        1.777662  djsa"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresW1DF=pd.concat([pd.DataFrame(clim_featuresLw1,columns=['climate_mean_fw','climate_median_fw','climate_std_fw']),\n",
    "                           pd.DataFrame(climIdsw1,columns=['uid'])],1)\n",
    "clim_featuresW1DF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rain_mean</th>\n",
       "      <th>rain_median</th>\n",
       "      <th>rain_std</th>\n",
       "      <th>uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.181429</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.381947</td>\n",
       "      <td>waln</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.212143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.378166</td>\n",
       "      <td>wlwd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.092857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.267059</td>\n",
       "      <td>zxdf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.060000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.180317</td>\n",
       "      <td>rzcf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.100714</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.268088</td>\n",
       "      <td>djsa</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rain_mean  rain_median  rain_std   uid\n",
       "0   0.181429          0.0  0.381947  waln\n",
       "1   0.212143          0.0  0.378166  wlwd\n",
       "2   0.092857          0.0  0.267059  zxdf\n",
       "3   0.060000          0.0  0.180317  rzcf\n",
       "4   0.100714          0.0  0.268088  djsa"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresRDF=pd.concat([pd.DataFrame(clim_featuresR,columns=['rain_mean','rain_median','rain_std']),\n",
    "                           pd.DataFrame(climIdsR,columns=['uid'])],1)\n",
    "clim_featuresRDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rain_median\n",
       "0.000000       19695\n",
       "1.000000         134\n",
       "0.040000          91\n",
       "0.020000          56\n",
       "0.060000          40\n",
       "               ...  \n",
       "0.266667           1\n",
       "0.490000           1\n",
       "0.220000           1\n",
       "0.140000           1\n",
       "0.441667           1\n",
       "Length: 107, dtype: int64"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresRDF[['rain_median']].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gust_mean</th>\n",
       "      <th>gust_median</th>\n",
       "      <th>gust_std</th>\n",
       "      <th>uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6.577143</td>\n",
       "      <td>6.62</td>\n",
       "      <td>2.962208</td>\n",
       "      <td>waln</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.080714</td>\n",
       "      <td>5.42</td>\n",
       "      <td>2.294039</td>\n",
       "      <td>wlwd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.024286</td>\n",
       "      <td>2.26</td>\n",
       "      <td>2.655266</td>\n",
       "      <td>zxdf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.992857</td>\n",
       "      <td>1.90</td>\n",
       "      <td>1.157734</td>\n",
       "      <td>rzcf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.527857</td>\n",
       "      <td>4.50</td>\n",
       "      <td>2.498653</td>\n",
       "      <td>djsa</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gust_mean  gust_median  gust_std   uid\n",
       "0   6.577143         6.62  2.962208  waln\n",
       "1   5.080714         5.42  2.294039  wlwd\n",
       "2   3.024286         2.26  2.655266  zxdf\n",
       "3   1.992857         1.90  1.157734  rzcf\n",
       "4   4.527857         4.50  2.498653  djsa"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresGDF=pd.concat([pd.DataFrame(clim_featuresG,columns=['gust_mean','gust_median','gust_std']),\n",
    "                           pd.DataFrame(climIdsG,columns=['uid'])],1)\n",
    "clim_featuresGDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \n",
      "/var/venv/cuda112py37/lib/python3.7/site-packages/ipykernel_launcher.py:5: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  \"\"\"\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hgt_mean</th>\n",
       "      <th>hgt_median</th>\n",
       "      <th>hgt_std</th>\n",
       "      <th>uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120.442308</td>\n",
       "      <td>120.442308</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>waln</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>109.023669</td>\n",
       "      <td>109.023669</td>\n",
       "      <td>2.842171e-14</td>\n",
       "      <td>wlwd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>309.899408</td>\n",
       "      <td>309.899408</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>zxdf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>484.071429</td>\n",
       "      <td>484.071429</td>\n",
       "      <td>1.705303e-13</td>\n",
       "      <td>rzcf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>161.089744</td>\n",
       "      <td>161.089744</td>\n",
       "      <td>2.842171e-14</td>\n",
       "      <td>djsa</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     hgt_mean  hgt_median       hgt_std   uid\n",
       "0  120.442308  120.442308  0.000000e+00  waln\n",
       "1  109.023669  109.023669  2.842171e-14  wlwd\n",
       "2  309.899408  309.899408  0.000000e+00  zxdf\n",
       "3  484.071429  484.071429  1.705303e-13  rzcf\n",
       "4  161.089744  161.089744  2.842171e-14  djsa"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clim_featuresSDF=pd.concat([pd.DataFrame(clim_featuresS,columns=['snowc_mean','snowc_median','snowc_std']),\n",
    "                           pd.DataFrame(climIdsS,columns=['uid'])],1)\n",
    "clim_featuresSDF.head()\n",
    "clim_featuresHDF=pd.concat([pd.DataFrame(clim_featuresH,columns=['hgt_mean','hgt_median','hgt_std']),\n",
    "                           pd.DataFrame(climIdsH,columns=['uid'])],1)\n",
    "clim_featuresHDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjAAAAGdCAYAAAAMm0nCAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAnSUlEQVR4nO3df3AU533H8c8hcccPSycElk5XC4FJDBYggUVQFBsCgUoIDUnG1LEBG9ko4DDCP5DtYjUuCNxBCrQE7FIorYF0AgNxx2AHGoyEMThG/BJz5VesGgIWLpxoDOiQPBZIuv7RYekFMBLWafWI92tmZ273efb2exvG+uTZZ3cdwWAwKAAAAIN0srsAAACAliLAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACME2l3AeHS1NSks2fPKioqSg6Hw+5yAABAMwSDQV2+fFler1edOt16nKXDBpizZ88qMTHR7jIAAMAdOHPmjO67775btnfYABMVFSXp/05AdHS0zdUAAIDmCAQCSkxMtP6O30qHDTDXLhtFR0cTYAAAMMztpn8wiRcAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYJwWB5jdu3drwoQJ8nq9cjgc2rx5c0i7w+G46bJ48WKrT58+fW5oLykpCfmew4cPa8SIEerSpYsSExO1aNGiO/uFAACgw2lxgKmrq1NqaqqWL19+0/Zz586FLKtXr5bD4dDEiRND+i1YsCCk33PPPWe1BQIBZWZmKikpSRUVFVq8eLGKioq0atWqlpYLAAA6oBY/iTc7O1vZ2dm3bPd4PCHr7777rkaPHq37778/ZHtUVNQNfa9Zt26drly5otWrV8vpdGrgwIHy+XxasmSJZsyY0dKSAQBABxPWOTDV1dXaunWr8vLybmgrKSlRz549NXToUC1evFgNDQ1WW3l5uUaOHCmn02lty8rKUmVlpS5evHjTY9XX1ysQCIQsAACgYwrru5B+9atfKSoqSo8++mjI9ueff14PPfSQYmNjtWfPHhUWFurcuXNasmSJJMnv96tv374h+8THx1ttPXr0uOFYxcXFmj9/fph+CQAAaE/CGmBWr16tKVOmqEuXLiHbCwoKrM8pKSlyOp169tlnVVxcLJfLdUfHKiwsDPnea2+zBAAAHU/YAsxHH32kyspKbdy48bZ909PT1dDQoNOnT6t///7yeDyqrq4O6XNt/VbzZlwu1x2HHwAAYJawBZi33npLaWlpSk1NvW1fn8+nTp06KS4uTpKUkZGhn//857p69ao6d+4sSSotLVX//v1vevkIME2fV7fets/pkpw2qAQAzNTiSby1tbXy+Xzy+XySpFOnTsnn86mqqsrqEwgE9Pbbb+unP/3pDfuXl5dr6dKl+s///E/98Y9/1Lp16zR79mw9+eSTVjiZPHmynE6n8vLydOzYMW3cuFHLli0LuUQEAADuXi0egTl48KBGjx5trV8LFbm5uVq7dq0kacOGDQoGg5o0adIN+7tcLm3YsEFFRUWqr69X3759NXv27JBw4na7tX37duXn5ystLU29evXS3LlzuYUaAABIkhzBYDBodxHhEAgE5Ha7VVNTo+joaLvLAUJwCQkAbq65f795FxIAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgnBYHmN27d2vChAnyer1yOBzavHlzSPvTTz8th8MRsowbNy6kz4ULFzRlyhRFR0crJiZGeXl5qq2tDelz+PBhjRgxQl26dFFiYqIWLVrU8l8HAAA6pMiW7lBXV6fU1FRNmzZNjz766E37jBs3TmvWrLHWXS5XSPuUKVN07tw5lZaW6urVq3rmmWc0Y8YMrV+/XpIUCASUmZmpsWPHauXKlTpy5IimTZummJgYzZgxo6UlA22qz6tb7S4BADq8FgeY7OxsZWdnf20fl8slj8dz07Y//OEP2rZtmw4cOKBhw4ZJkt58802NHz9ef//3fy+v16t169bpypUrWr16tZxOpwYOHCifz6clS5YQYHDXaE4QOl2S0waVAED7E5Y5MB9++KHi4uLUv39/zZw5U1988YXVVl5erpiYGCu8SNLYsWPVqVMn7du3z+ozcuRIOZ1Oq09WVpYqKyt18eLFmx6zvr5egUAgZAEAAB1TqweYcePG6d/+7d+0Y8cO/eIXv9CuXbuUnZ2txsZGSZLf71dcXFzIPpGRkYqNjZXf77f6xMfHh/S5tn6tz58rLi6W2+22lsTExNb+aQAAoJ1o8SWk23niiSesz4MHD1ZKSor69eunDz/8UGPGjGntw1kKCwtVUFBgrQcCAUIMAAAdVNhvo77//vvVq1cvnThxQpLk8Xh0/vz5kD4NDQ26cOGCNW/G4/Gouro6pM+19VvNrXG5XIqOjg5ZAABAxxT2APP555/riy++UEJCgiQpIyNDly5dUkVFhdXngw8+UFNTk9LT060+u3fv1tWrV60+paWl6t+/v3r06BHukgEAQDvX4gBTW1srn88nn88nSTp16pR8Pp+qqqpUW1urV155RXv37tXp06e1Y8cO/ehHP9K3vvUtZWVlSZIefPBBjRs3TtOnT9f+/fv18ccfa9asWXriiSfk9XolSZMnT5bT6VReXp6OHTumjRs3atmyZSGXiAAAwN2rxQHm4MGDGjp0qIYOHSpJKigo0NChQzV37lxFRETo8OHD+uEPf6gHHnhAeXl5SktL00cffRTyLJh169ZpwIABGjNmjMaPH69HHnlEq1atstrdbre2b9+uU6dOKS0tTS+99JLmzp3LLdQAAECS5AgGg0G7iwiHQCAgt9utmpoa5sOgTbXlg+x4DgyAjqa5f795FxIAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4kXYXACC8+ry69bZ9TpfktEElANB6GIEBAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHF4FxIA3pcEwDiMwAAAAOMQYAAAgHEIMAAAwDgtDjC7d+/WhAkT5PV65XA4tHnzZqvt6tWrmjNnjgYPHqzu3bvL6/Vq6tSpOnv2bMh39OnTRw6HI2QpKSkJ6XP48GGNGDFCXbp0UWJiohYtWnRnvxAAAHQ4LQ4wdXV1Sk1N1fLly29o+/LLL3Xo0CH97d/+rQ4dOqR33nlHlZWV+uEPf3hD3wULFujcuXPW8txzz1ltgUBAmZmZSkpKUkVFhRYvXqyioiKtWrWqpeUCAIAOqMV3IWVnZys7O/umbW63W6WlpSHb/vEf/1HDhw9XVVWVevfubW2PioqSx+O56fesW7dOV65c0erVq+V0OjVw4ED5fD4tWbJEM2bMaGnJAACggwn7HJiamho5HA7FxMSEbC8pKVHPnj01dOhQLV68WA0NDVZbeXm5Ro4cKafTaW3LyspSZWWlLl68GO6SAQBAOxfW58B89dVXmjNnjiZNmqTo6Ghr+/PPP6+HHnpIsbGx2rNnjwoLC3Xu3DktWbJEkuT3+9W3b9+Q74qPj7faevToccOx6uvrVV9fb60HAoFw/CQAANAOhC3AXL16VT/5yU8UDAa1YsWKkLaCggLrc0pKipxOp5599lkVFxfL5XLd0fGKi4s1f/78b1QzAAAwQ1guIV0LL5999plKS0tDRl9uJj09XQ0NDTp9+rQkyePxqLq6OqTPtfVbzZspLCxUTU2NtZw5c+ab/xAAANAutXqAuRZePv30U5WVlalnz5633cfn86lTp06Ki4uTJGVkZGj37t26evWq1ae0tFT9+/e/6eUjSXK5XIqOjg5ZAABAx9TiS0i1tbU6ceKEtX7q1Cn5fD7FxsYqISFBf/VXf6VDhw5py5YtamxslN/vlyTFxsbK6XSqvLxc+/bt0+jRoxUVFaXy8nLNnj1bTz75pBVOJk+erPnz5ysvL09z5szR0aNHtWzZMv3yl79spZ8NAABM1uIAc/DgQY0ePdpavzafJTc3V0VFRXrvvfckSUOGDAnZb+fOnRo1apRcLpc2bNigoqIi1dfXq2/fvpo9e3bIvBi3263t27crPz9faWlp6tWrl+bOncst1AAAQNIdBJhRo0YpGAzesv3r2iTpoYce0t69e297nJSUFH300UctLQ8AANwFeBcSAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIzT4nchAbg79Xl16237nC7JaYNKAIARGAAAYCACDAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMw8scAYM15wWLANARMQIDAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYJxIuwsA0HH0eXXrbfucLslpg0oAdHQtHoHZvXu3JkyYIK/XK4fDoc2bN4e0B4NBzZ07VwkJCeratavGjh2rTz/9NKTPhQsXNGXKFEVHRysmJkZ5eXmqra0N6XP48GGNGDFCXbp0UWJiohYtWtTyXwcAADqkFgeYuro6paamavny5TdtX7Rokd544w2tXLlS+/btU/fu3ZWVlaWvvvrK6jNlyhQdO3ZMpaWl2rJli3bv3q0ZM2ZY7YFAQJmZmUpKSlJFRYUWL16soqIirVq16g5+IgAA6GgcwWAweMc7OxzatGmTfvzjH0v6v9EXr9erl156SS+//LIkqaamRvHx8Vq7dq2eeOIJ/eEPf1BycrIOHDigYcOGSZK2bdum8ePH6/PPP5fX69WKFSv085//XH6/X06nU5L06quvavPmzfrkk0+aVVsgEJDb7VZNTY2io6Pv9CcCLdacyyh3My4hAfg6zf373aqTeE+dOiW/36+xY8da29xut9LT01VeXi5JKi8vV0xMjBVeJGns2LHq1KmT9u3bZ/UZOXKkFV4kKSsrS5WVlbp48eJNj11fX69AIBCyAACAjqlVA4zf75ckxcfHh2yPj4+32vx+v+Li4kLaIyMjFRsbG9LnZt/x/4/x54qLi+V2u60lMTHxm/8gAADQLnWY26gLCwtVU1NjLWfOnLG7JAAAECatGmA8Ho8kqbq6OmR7dXW11ebxeHT+/PmQ9oaGBl24cCGkz82+4/8f48+5XC5FR0eHLAAAoGNq1QDTt29feTwe7dixw9oWCAS0b98+ZWRkSJIyMjJ06dIlVVRUWH0++OADNTU1KT093eqze/duXb161epTWlqq/v37q0ePHq1ZMgAAMFCLA0xtba18Pp98Pp+k/5u46/P5VFVVJYfDoRdffFF/93d/p/fee09HjhzR1KlT5fV6rTuVHnzwQY0bN07Tp0/X/v379fHHH2vWrFl64okn5PV6JUmTJ0+W0+lUXl6ejh07po0bN2rZsmUqKChotR8OAADM1eIn8R48eFCjR4+21q+FitzcXK1du1Z//dd/rbq6Os2YMUOXLl3SI488om3btqlLly7WPuvWrdOsWbM0ZswYderUSRMnTtQbb7xhtbvdbm3fvl35+flKS0tTr169NHfu3JBnxQAAgLvXN3oOTHvGc2AQDjzj5ZvjOTAAvo4tz4EBAABoCwQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxWvwkXgD4JprzMEAedgfgdhiBAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxom0uwAA+HN9Xt162z6nS3LaoBIA7RUjMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjNPqAaZPnz5yOBw3LPn5+ZKkUaNG3dD2s5/9LOQ7qqqqlJOTo27duikuLk6vvPKKGhoaWrtUAABgqFa/jfrAgQNqbGy01o8ePaq//Mu/1GOPPWZtmz59uhYsWGCtd+vWzfrc2NionJwceTwe7dmzR+fOndPUqVPVuXNnLVy4sLXLBQAABmr1AHPvvfeGrJeUlKhfv376/ve/b23r1q2bPB7PTfffvn27jh8/rrKyMsXHx2vIkCF6/fXXNWfOHBUVFcnpdLZ2yQAAwDBhnQNz5coV/frXv9a0adPkcDis7evWrVOvXr00aNAgFRYW6ssvv7TaysvLNXjwYMXHx1vbsrKyFAgEdOzYsVseq76+XoFAIGQBAAAdU1ifxLt582ZdunRJTz/9tLVt8uTJSkpKktfr1eHDhzVnzhxVVlbqnXfekST5/f6Q8CLJWvf7/bc8VnFxsebPn9/6PwIAALQ7YQ0wb731lrKzs+X1eq1tM2bMsD4PHjxYCQkJGjNmjE6ePKl+/frd8bEKCwtVUFBgrQcCASUmJt7x9wEAgPYrbAHms88+U1lZmTWycivp6emSpBMnTqhfv37yeDzav39/SJ/q6mpJuuW8GUlyuVxyuVzfsGoAAGCCsM2BWbNmjeLi4pST8/UvXPP5fJKkhIQESVJGRoaOHDmi8+fPW31KS0sVHR2t5OTkcJULAAAMEpYRmKamJq1Zs0a5ubmKjLx+iJMnT2r9+vUaP368evbsqcOHD2v27NkaOXKkUlJSJEmZmZlKTk7WU089pUWLFsnv9+u1115Tfn4+IywAAEBSmAJMWVmZqqqqNG3atJDtTqdTZWVlWrp0qerq6pSYmKiJEyfqtddes/pERERoy5YtmjlzpjIyMtS9e3fl5uaGPDcGAADc3cISYDIzMxUMBm/YnpiYqF27dt12/6SkJP3Hf/xHOEoDAAAdAO9CAgAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGIcAAAADjhOVVAoCJ+ry61e4S0Mqa87/p6ZKcNqgEQGtjBAYAABiHAAMAAIxDgAEAAMYhwAAAAOMQYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxuFdSDAe77sBgLsPIzAAAMA4BBgAAGAcAgwAADAOAQYAABiHSbwAjNScydsAOi5GYAAAgHEIMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcXgSL+4KPLUVADoWRmAAAIBxCDAAAMA4rR5gioqK5HA4QpYBAwZY7V999ZXy8/PVs2dP3XPPPZo4caKqq6tDvqOqqko5OTnq1q2b4uLi9Morr6ihoaG1SwUAAIYKyxyYgQMHqqys7PpBIq8fZvbs2dq6davefvttud1uzZo1S48++qg+/vhjSVJjY6NycnLk8Xi0Z88enTt3TlOnTlXnzp21cOHCcJQLAAAME5YAExkZKY/Hc8P2mpoavfXWW1q/fr1+8IMfSJLWrFmjBx98UHv37tV3v/tdbd++XcePH1dZWZni4+M1ZMgQvf7665ozZ46KiorkdDrDUTIAADBIWObAfPrpp/J6vbr//vs1ZcoUVVVVSZIqKip09epVjR071uo7YMAA9e7dW+Xl5ZKk8vJyDR48WPHx8VafrKwsBQIBHTt27JbHrK+vVyAQCFkAAEDH1OoBJj09XWvXrtW2bdu0YsUKnTp1SiNGjNDly5fl9/vldDoVExMTsk98fLz8fr8kye/3h4SXa+3X2m6luLhYbrfbWhITE1v3hwEAgHaj1S8hZWdnW59TUlKUnp6upKQk/eY3v1HXrl1b+3CWwsJCFRQUWOuBQIAQAwBABxX226hjYmL0wAMP6MSJE/J4PLpy5YouXboU0qe6utqaM+PxeG64K+na+s3m1VzjcrkUHR0dsgAAgI4p7AGmtrZWJ0+eVEJCgtLS0tS5c2ft2LHDaq+srFRVVZUyMjIkSRkZGTpy5IjOnz9v9SktLVV0dLSSk5PDXS4AADBAq19CevnllzVhwgQlJSXp7NmzmjdvniIiIjRp0iS53W7l5eWpoKBAsbGxio6O1nPPPaeMjAx997vflSRlZmYqOTlZTz31lBYtWiS/36/XXntN+fn5crlcrV0uAAAwUKsHmM8//1yTJk3SF198oXvvvVePPPKI9u7dq3vvvVeS9Mtf/lKdOnXSxIkTVV9fr6ysLP3TP/2TtX9ERIS2bNmimTNnKiMjQ927d1dubq4WLFjQ2qUCAABDOYLBYNDuIsIhEAjI7XarpqaG+TAdHC9qxDdxuiTH7hIA/D/N/fvNu5AAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIzT6s+BAVoTt0gDAG6GERgAAGAcAgwAADAOl5AA3NWac5mSp/UC7Q8jMAAAwDgEGAAAYBwCDAAAMA4BBgAAGIcAAwAAjMNdSLAND6kDANwpRmAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIzDbdQIC26RBgCEEyMwAADAOIzAAMBtNGdE8XRJThtUAuAaRmAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMYhwAAAAOPwIDu0GK8JAADYjREYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADGafW7kIqLi/XOO+/ok08+UdeuXfW9731Pv/jFL9S/f3+rz6hRo7Rr166Q/Z599lmtXLnSWq+qqtLMmTO1c+dO3XPPPcrNzVVxcbEiI7lxCkD705y7806X5LRBJcDdodXTwK5du5Sfn6/vfOc7amho0N/8zd8oMzNTx48fV/fu3a1+06dP14IFC6z1bt26WZ8bGxuVk5Mjj8ejPXv26Ny5c5o6dao6d+6shQsXtnbJAADAMK0eYLZt2xayvnbtWsXFxamiokIjR460tnfr1k0ej+em37F9+3YdP35cZWVlio+P15AhQ/T6669rzpw5KioqktPpbO2yAQCAQcI+B6ampkaSFBsbG7J93bp16tWrlwYNGqTCwkJ9+eWXVlt5ebkGDx6s+Ph4a1tWVpYCgYCOHTsW7pIBAEA7F9YJJU1NTXrxxRf18MMPa9CgQdb2yZMnKykpSV6vV4cPH9acOXNUWVmpd955R5Lk9/tDwoska93v99/0WPX19aqvr7fWA4FAa/8cAADQToQ1wOTn5+vo0aP6/e9/H7J9xowZ1ufBgwcrISFBY8aM0cmTJ9WvX787OlZxcbHmz5//jeoFAABmCNslpFmzZmnLli3auXOn7rvvvq/tm56eLkk6ceKEJMnj8ai6ujqkz7X1W82bKSwsVE1NjbWcOXPmm/4EAADQTrV6gAkGg5o1a5Y2bdqkDz74QH379r3tPj6fT5KUkJAgScrIyNCRI0d0/vx5q09paamio6OVnJx80+9wuVyKjo4OWQAAQMfU6peQ8vPztX79er377ruKioqy5qy43W517dpVJ0+e1Pr16zV+/Hj17NlThw8f1uzZszVy5EilpKRIkjIzM5WcnKynnnpKixYtkt/v12uvvab8/Hy5XK7WLhkAABim1UdgVqxYoZqaGo0aNUoJCQnWsnHjRkmS0+lUWVmZMjMzNWDAAL300kuaOHGifvvb31rfERERoS1btigiIkIZGRl68sknNXXq1JDnxgAAgLuXIxgMBu0uIhwCgYDcbrdqamq4nNQCzXmaKIA7w5N4gdtr7t9v3oUEAACMQ4ABAADG4c2IANBGeOEj0HoIMADQjhBygObhEhIAADAOAQYAABiHS0h3CW6PBgB0JIzAAAAA4xBgAACAcQgwAADAOAQYAABgHCbxAoBheFYMwAgMAAAwEAEGAAAYhwADAACMQ4ABAADGIcAAAADjEGAAAIBxCDAAAMA4BBgAAGAcAgwAADAOAQYAABiHAAMAAIxDgAEAAMbhZY4A0AHxwkd0dIzAAAAA4xBgAACAcQgwAADAOMyB6QCac60bAICOhAADAHcpJvrCZFxCAgAAxiHAAAAA4xBgAACAcZgDAwC4pda6SYC5NGhtjMAAAADjMAIDAAg77nhCayPAtHM84wUAgBtxCQkAABiHAAMAAIxDgAEAAMZp13Ngli9frsWLF8vv9ys1NVVvvvmmhg8fbndZAIB2jAnDd4d2G2A2btyogoICrVy5Uunp6Vq6dKmysrJUWVmpuLg4u8trFUzQBYDrCB5oiXZ7CWnJkiWaPn26nnnmGSUnJ2vlypXq1q2bVq9ebXdpAADAZu1yBObKlSuqqKhQYWGhta1Tp04aO3asysvLb7pPfX296uvrrfWamhpJUiAQaPX6Bs17v9W/EwBwe71nv92uvqc5js7ParNjdQTX/m4Hg8Gv7dcuA8yf/vQnNTY2Kj4+PmR7fHy8Pvnkk5vuU1xcrPnz59+wPTExMSw1AgDQHO6ldldgpsuXL8vtdt+yvV0GmDtRWFiogoICa72pqUkXLlxQz5495XA4Wu04gUBAiYmJOnPmjKKjo1vte03EubiOc3Ed5+I6zsV1nIvrOBfX3excBINBXb58WV6v92v3bZcBplevXoqIiFB1dXXI9urqank8npvu43K55HK5QrbFxMSEq0RFR0ff9f/wruFcXMe5uI5zcR3n4jrOxXWci+v+/Fx83cjLNe1yEq/T6VRaWpp27NhhbWtqatKOHTuUkZFhY2UAAKA9aJcjMJJUUFCg3NxcDRs2TMOHD9fSpUtVV1enZ555xu7SAACAzdptgHn88cf1P//zP5o7d678fr+GDBmibdu23TCxt625XC7NmzfvhstVdyPOxXWci+s4F9dxLq7jXFzHubjum5wLR/B29ykBAAC0M+1yDgwAAMDXIcAAAADjEGAAAIBxCDAAAMA4BJgWWr58ufr06aMuXbooPT1d+/fvt7ukNrd7925NmDBBXq9XDodDmzdvtrsk2xQXF+s73/mOoqKiFBcXpx//+MeqrKy0uyxbrFixQikpKdYDqTIyMvS73/3O7rJsV1JSIofDoRdffNHuUmxRVFQkh8MRsgwYMMDusmzz3//933ryySfVs2dPde3aVYMHD9bBgwftLqvN9enT54Z/Fw6HQ/n5+c3+DgJMC2zcuFEFBQWaN2+eDh06pNTUVGVlZen8+fN2l9am6urqlJqaquXLl9tdiu127dql/Px87d27V6Wlpbp69aoyMzNVV1dnd2lt7r777lNJSYkqKip08OBB/eAHP9CPfvQjHTt2zO7SbHPgwAH98z//s1JSUuwuxVYDBw7UuXPnrOX3v/+93SXZ4uLFi3r44YfVuXNn/e53v9Px48f1D//wD+rRo4fdpbW5AwcOhPybKC0tlSQ99thjzf+SIJpt+PDhwfz8fGu9sbEx6PV6g8XFxTZWZS9JwU2bNtldRrtx/vz5oKTgrl277C6lXejRo0fwX//1X+0uwxaXL18Ofvvb3w6WlpYGv//97wdfeOEFu0uyxbx584Kpqal2l9EuzJkzJ/jII4/YXUa79MILLwT79esXbGpqavY+jMA005UrV1RRUaGxY8da2zp16qSxY8eqvLzcxsrQntTU1EiSYmNjba7EXo2NjdqwYYPq6uru2td/5OfnKycnJ+S/GXerTz/9VF6vV/fff7+mTJmiqqoqu0uyxXvvvadhw4bpscceU1xcnIYOHap/+Zd/sbss2125ckW//vWvNW3atBa9fJkA00x/+tOf1NjYeMOTgOPj4+X3+22qCu1JU1OTXnzxRT388MMaNGiQ3eXY4siRI7rnnnvkcrn0s5/9TJs2bVJycrLdZbW5DRs26NChQyouLra7FNulp6dr7dq12rZtm1asWKFTp05pxIgRunz5st2ltbk//vGPWrFihb797W/r/fff18yZM/X888/rV7/6ld2l2Wrz5s26dOmSnn766Rbt125fJQCYJj8/X0ePHr1rr+9LUv/+/eXz+VRTU6N///d/V25urnbt2nVXhZgzZ87ohRdeUGlpqbp06WJ3ObbLzs62PqekpCg9PV1JSUn6zW9+o7y8PBsra3tNTU0aNmyYFi5cKEkaOnSojh49qpUrVyo3N9fm6uzz1ltvKTs7W16vt0X7MQLTTL169VJERISqq6tDtldXV8vj8dhUFdqLWbNmacuWLdq5c6fuu+8+u8uxjdPp1Le+9S2lpaWpuLhYqampWrZsmd1ltamKigqdP39eDz30kCIjIxUZGaldu3bpjTfeUGRkpBobG+0u0VYxMTF64IEHdOLECbtLaXMJCQk3hPkHH3zwrr2kJkmfffaZysrK9NOf/rTF+xJgmsnpdCotLU07duywtjU1NWnHjh137TV+SMFgULNmzdKmTZv0wQcfqG/fvnaX1K40NTWpvr7e7jLa1JgxY3TkyBH5fD5rGTZsmKZMmSKfz6eIiAi7S7RVbW2tTp48qYSEBLtLaXMPP/zwDY9Z+K//+i8lJSXZVJH91qxZo7i4OOXk5LR4Xy4htUBBQYFyc3M1bNgwDR8+XEuXLlVdXZ2eeeYZu0trU7W1tSH/7+nUqVPy+XyKjY1V7969bays7eXn52v9+vV69913FRUVZc2Hcrvd6tq1q83Vta3CwkJlZ2erd+/eunz5stavX68PP/xQ77//vt2ltamoqKgb5kB1795dPXv2vCvnRr388suaMGGCkpKSdPbsWc2bN08RERGaNGmS3aW1udmzZ+t73/ueFi5cqJ/85Cfav3+/Vq1apVWrVtldmi2ampq0Zs0a5ebmKjLyDuJI+G6K6pjefPPNYO/evYNOpzM4fPjw4N69e+0uqc3t3LkzKOmGJTc31+7S2tzNzoOk4Jo1a+wurc1NmzYtmJSUFHQ6ncF77703OGbMmOD27dvtLqtduJtvo3788ceDCQkJQafTGfyLv/iL4OOPPx48ceKE3WXZ5re//W1w0KBBQZfLFRwwYEBw1apVdpdkm/fffz8oKVhZWXlH+zuCwWCwdbIUAABA22AODAAAMA4BBgAAGIcAAwAAjEOAAQAAxiHAAAAA4xBgAACAcQgwAADAOAQYAABgHAIMAAAwDgEGAAAYhwADAACMQ4ABAADG+V9cxGHtQdPCVQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(clim_featuresGDF.gust_mean,50)\n",
    "plt.hist(clim_featuresGDF.gust_median,50)\n",
    "plt.show()\n",
    "plt.hist(clim_featuresGDF.gust_std,50)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# plt.imshow(rgbCLIM[0])\n",
    "# plt.show()\n",
    "# # plt.imshow( np.load(CLIMATE_DATA_DIR+f)['a'])\n",
    "# # plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clim_featuresL.shape,rgbCLIM.shape,im.shape, np.load(CLIMATE_DATA_DIR+f)['a'].shape,f,f.split('.')[0]\n",
    "# clim_featuresL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uid</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>date</th>\n",
       "      <th>split</th>\n",
       "      <th>season</th>\n",
       "      <th>region</th>\n",
       "      <th>severity</th>\n",
       "      <th>density</th>\n",
       "      <th>latlong</th>\n",
       "      <th>latlong1</th>\n",
       "      <th>latlong0</th>\n",
       "      <th>latlong10</th>\n",
       "      <th>altitude</th>\n",
       "      <th>DEMmean</th>\n",
       "      <th>DEMmedian</th>\n",
       "      <th>DEMstd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>aabm</td>\n",
       "      <td>39.080319</td>\n",
       "      <td>-86.430867</td>\n",
       "      <td>2018-05-14</td>\n",
       "      <td>train</td>\n",
       "      <td>spring</td>\n",
       "      <td>midwest</td>\n",
       "      <td>1.0</td>\n",
       "      <td>585.0</td>\n",
       "      <td>39.08_-86.43</td>\n",
       "      <td>39.1_-86.4</td>\n",
       "      <td>39.0_-86.0</td>\n",
       "      <td>39.1_-86.0</td>\n",
       "      <td>169.0</td>\n",
       "      <td>61.069930</td>\n",
       "      <td>22.5</td>\n",
       "      <td>68.402346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aabn</td>\n",
       "      <td>36.559700</td>\n",
       "      <td>-121.510000</td>\n",
       "      <td>2016-08-31</td>\n",
       "      <td>test</td>\n",
       "      <td>summer</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>36.56_-121.51</td>\n",
       "      <td>36.6_-121.5</td>\n",
       "      <td>37.0_-122.0</td>\n",
       "      <td>36.6_-122.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>62.063702</td>\n",
       "      <td>40.0</td>\n",
       "      <td>53.451200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>aacd</td>\n",
       "      <td>35.875083</td>\n",
       "      <td>-78.878434</td>\n",
       "      <td>2020-11-19</td>\n",
       "      <td>train</td>\n",
       "      <td>fall</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>290.0</td>\n",
       "      <td>35.88_-78.88</td>\n",
       "      <td>35.9_-78.9</td>\n",
       "      <td>36.0_-79.0</td>\n",
       "      <td>35.9_-79.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>130.127404</td>\n",
       "      <td>133.0</td>\n",
       "      <td>44.633880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>aaee</td>\n",
       "      <td>35.487000</td>\n",
       "      <td>-79.062133</td>\n",
       "      <td>2016-08-24</td>\n",
       "      <td>train</td>\n",
       "      <td>summer</td>\n",
       "      <td>south</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1614.0</td>\n",
       "      <td>35.49_-79.06</td>\n",
       "      <td>35.5_-79.1</td>\n",
       "      <td>35.0_-79.0</td>\n",
       "      <td>35.5_-79.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>147.151442</td>\n",
       "      <td>152.0</td>\n",
       "      <td>50.721441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>aaff</td>\n",
       "      <td>38.049471</td>\n",
       "      <td>-99.827001</td>\n",
       "      <td>2019-07-23</td>\n",
       "      <td>train</td>\n",
       "      <td>summer</td>\n",
       "      <td>midwest</td>\n",
       "      <td>3.0</td>\n",
       "      <td>111825.0</td>\n",
       "      <td>38.05_-99.83</td>\n",
       "      <td>38.0_-99.8</td>\n",
       "      <td>38.0_-100.0</td>\n",
       "      <td>38.0_-100.0</td>\n",
       "      <td>702.0</td>\n",
       "      <td>132.240093</td>\n",
       "      <td>131.5</td>\n",
       "      <td>52.735658</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    uid   latitude   longitude       date  split  season   region  severity  \\\n",
       "0  aabm  39.080319  -86.430867 2018-05-14  train  spring  midwest       1.0   \n",
       "1  aabn  36.559700 -121.510000 2016-08-31   test  summer       -1      -1.0   \n",
       "2  aacd  35.875083  -78.878434 2020-11-19  train    fall    south       1.0   \n",
       "3  aaee  35.487000  -79.062133 2016-08-24  train  summer    south       1.0   \n",
       "4  aaff  38.049471  -99.827001 2019-07-23  train  summer  midwest       3.0   \n",
       "\n",
       "    density        latlong     latlong1     latlong0    latlong10  altitude  \\\n",
       "0     585.0   39.08_-86.43   39.1_-86.4   39.0_-86.0   39.1_-86.0     169.0   \n",
       "1      -1.0  36.56_-121.51  36.6_-121.5  37.0_-122.0  36.6_-122.0      30.0   \n",
       "2     290.0   35.88_-78.88   35.9_-78.9   36.0_-79.0   35.9_-79.0      94.0   \n",
       "3    1614.0   35.49_-79.06   35.5_-79.1   35.0_-79.0   35.5_-79.0     129.0   \n",
       "4  111825.0   38.05_-99.83   38.0_-99.8  38.0_-100.0  38.0_-100.0     702.0   \n",
       "\n",
       "      DEMmean  DEMmedian     DEMstd  \n",
       "0   61.069930       22.5  68.402346  \n",
       "1   62.063702       40.0  53.451200  \n",
       "2  130.127404      133.0  44.633880  \n",
       "3  147.151442      152.0  50.721441  \n",
       "4  132.240093      131.5  52.735658  "
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata=metadata.merge(altDF,how='left').fillna(-1)\n",
    "metadata.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[108.17427885, 104.5       ,  47.36478388],\n",
       "       [107.60153846, 103.        ,  33.23362057],\n",
       "       [ 62.67617866,  40.        ,  64.74393619],\n",
       "       ...,\n",
       "       [ 82.21153846,  78.        ,  33.90528325],\n",
       "       [110.68629808, 116.        ,  56.38459457],\n",
       "       [107.12379808,  96.        ,  55.43781436]])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dem_featuresL.shape,im.shape\n",
    "# (im[...,0]-im[...,1]).sum()\n",
    "# cv2.imread(DEM_DATA_DIR+f, cv2.IMREAD_GRAYSCALE).shape\n",
    "dem_featuresL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14905"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip3 install rasterio"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "    https://epic.awi.de/id/eprint/49883/1/remotesensing-11-01730.pdf\n",
    "\n",
    "Band Wavelength (μm) | Resolution (m) | Band Wavelength (μm) | Resolution (m)  \n",
    "\n",
    "                    |        LANDSAT       |       SENTINEL-2  \n",
    "\n",
    "    Coastal Aerosol | 1 | 0.45–0.451  | 30 | 1  | 0.430–0.450 | 60\n",
    "\n",
    "    Blue            | 2 | 0.452–0.51  | 30 | 2  | 0.448–0.546 | 10\n",
    "\n",
    "    Green           | 3 | 0.533–0.59  | 30 | 3  | 0.538–0.583 | 10\n",
    "\n",
    "    Red             | 4 | 0.636–0.673 | 30 | 4  | 0.646–0.684 | 10\n",
    "\n",
    "    NIR             | 5 | 0.851–0.879 | 30 | 8A | 0.848–0.881 | 20\n",
    "\n",
    "    NIR2                                   | 8 | 0.763–0.908 | 10\n",
    "\n",
    "    SWIR1           | 6 | 1.566–1.651 | 30 | 11 | 1.542–1.685 | 20\n",
    "\n",
    "    SWIR2           | 7 | 2.107–2.294 | 30 | 12 | 2.081–2.323 | 20\n",
    "\n",
    "    Cirrus          | 9 | 1.363–1.384 | 30 | 10 | 1.336–1.441 | 60\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(103, 103, 3)"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import rasterio\n",
    "def read_band(path,band='B4'):\n",
    "    im0=os.listdir(path)[0].split('.')[0]\n",
    "    b=rasterio.open(path+im0+'.'+band+'.tif' ).read()[0,...]\n",
    "#     percB = np.percentile(b,[0.1,99.99])\n",
    "#     b = np.clip(b, percB[0], percB[1])\n",
    "    B=(((b-b.min())/(b.max()-b.min()))*255).astype('uint8')\n",
    "    return(B)\n",
    "\n",
    "TIME_BUFFER_DAYS=20\n",
    "AREA_BUFFER = 500\n",
    "CLOUDS_LESS_THAN=30\n",
    "\n",
    "BASE_DIRs = 'EEtemp/S_t'+str(TIME_BUFFER_DAYS)+'_a'+str(AREA_BUFFER)+'_c'+str(CLOUDS_LESS_THAN)+'/'\n",
    "BASE_DIRl = 'EEtemp/L_t'+str(TIME_BUFFER_DAYS)+'_a'+str(AREA_BUFFER)+'_c'+str(CLOUDS_LESS_THAN)+'/'\n",
    "\n",
    "num=1\n",
    "file = BASE_DIRs+os.listdir(BASE_DIRs)[num]+'/'\n",
    "# file = BASE_DIRl+os.listdir(BASE_DIRl)[num]+'/'\n",
    "B4=read_band(file,band='B4')\n",
    "B3=read_band(file,band='B3')\n",
    "B2=read_band(file,band='B2')\n",
    "RGB=np.stack((B4,B3,B2),-1)\n",
    "RGB.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(36, 35)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num=1\n",
    "file = BASE_DIRl+os.listdir(BASE_DIRl)[num]+'/'\n",
    "im0=os.listdir(file)[0].split('.')[0]+'.B2.tif'\n",
    "b=rasterio.open(file+im0 ).read()[0,...]\n",
    "b.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(b)\n",
    "plt.show()\n",
    "# np.unique(b),b.mean(),b[8:-8,8:-8].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(B4)\n",
    "plt.show()\n",
    "plt.imshow(RGB)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "# RGBr=cv2.resize(RGB, (300,300))\n",
    "# plt.imshow(RGBr)\n",
    "# plt.show()\n",
    "# RGBr=cv2.resize(RGB, (300,300), cv2.INTER_CUBIC)\n",
    "# plt.imshow(RGBr)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "RESIZE_SIZE=64\n",
    "BASE_DIR=BASE_DIRs\n",
    "\n",
    "rgbsS=[]\n",
    "idsS=[]\n",
    "\n",
    "# for uid in tqdm(os.listdir(BASE_DIR)[:200], total=200):\n",
    "for uid in tqdm(os.listdir(BASE_DIR)):\n",
    "\n",
    "    file = BASE_DIR+uid+'/'\n",
    "    \n",
    "    try:\n",
    "        B1=read_band(file,band='B1')\n",
    "        B1=cv2.resize(B1,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B2=read_band(file,band='B2')\n",
    "        B2=cv2.resize(B2,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B3=read_band(file,band='B3')\n",
    "        B3=cv2.resize(B3,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B4=read_band(file,band='B4')\n",
    "        B4=cv2.resize(B4,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "#         B5=read_band(file,band='B8A')############################################\n",
    "        B5=read_band(file,band='B8')\n",
    "        B5=cv2.resize(B5,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B6=read_band(file,band='B11')\n",
    "        B6=cv2.resize(B6,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B7=read_band(file,band='B12')\n",
    "        B7=cv2.resize(B7,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "#         B8=read_band(file,band='B10')\n",
    "#         B8=cv2.resize(B8,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "#         allBands=np.stack((B1,B2,B3,B4,B5,B6,B7,B8),-1)\n",
    "        allBands=np.stack((B1,B2,B3,B4,B5,B6,B7),-1)\n",
    "        \n",
    "#         allBands=cv2.resize(allBands,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        \n",
    "        rgbsS.append(allBands)\n",
    "        idsS.append(uid)\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "#        \n",
    "            \n",
    "rgbsS=np.stack(rgbsS)    \n",
    "idsS=np.stack(idsS) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rgbsS.shape,rgbsS[0][:,:,[3,2,1]].shape,allBands.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.imshow(rgbsS[1112][:,:,[3,2,1]])\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "RESIZE_SIZE=64\n",
    "BASE_DIR=BASE_DIRl\n",
    "\n",
    "rgbsL=[]\n",
    "idsL=[]\n",
    "\n",
    "# for uid in tqdm(os.listdir(BASE_DIR)[:200], total=200):\n",
    "for uid in tqdm(os.listdir(BASE_DIR)):\n",
    "\n",
    "    file = BASE_DIR+uid+'/'\n",
    "    \n",
    "    try:\n",
    "        B1=read_band(file,band='B1')\n",
    "        B1=cv2.resize(B1,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B2=read_band(file,band='B2')\n",
    "        B2=cv2.resize(B2,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B3=read_band(file,band='B3')\n",
    "        B3=cv2.resize(B3,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B4=read_band(file,band='B4')\n",
    "        B4=cv2.resize(B4,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B5=read_band(file,band='B5')\n",
    "        B5=cv2.resize(B5,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B6=read_band(file,band='B6')\n",
    "        B6=cv2.resize(B6,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        B7=read_band(file,band='B7')\n",
    "        B7=cv2.resize(B7,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "#         B8=read_band(file,band='B9')\n",
    "#         B8=cv2.resize(B8,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        \n",
    "        allBands=np.stack((B1,B2,B3,B4,B5,B6,B7),-1)\n",
    "        \n",
    "#         allBands=cv2.resize(allBands,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        \n",
    "        rgbsL.append(allBands)\n",
    "        idsL.append(uid)\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "#        \n",
    "            \n",
    "rgbsL=np.stack(rgbsL)    \n",
    "idsL=np.stack(idsL) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "# RESIZE_SIZE=20\n",
    "# RESIZE_SIZEq=int(RESIZE_SIZE/4)\n",
    "# BASE_DIR=BASE_DIRl\n",
    "\n",
    "# L1011fs=[]\n",
    "# idsLf=[]\n",
    "\n",
    "# # for uid in tqdm(os.listdir(BASE_DIR)[:200], total=200):\n",
    "# for uid in tqdm(os.listdir(BASE_DIR)):\n",
    "\n",
    "#     file = BASE_DIR+uid+'/'\n",
    "    \n",
    "#     try:\n",
    "#         B10=read_band(file,band='B10')\n",
    "#         B10=cv2.resize(B10,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "#         B11=read_band(file,band='B11')\n",
    "#         B11=cv2.resize(B11,(RESIZE_SIZE, RESIZE_SIZE))\n",
    "        \n",
    "#         B10mean=np.mean(B10)\n",
    "#         B10Cmean=np.mean(B10[RESIZE_SIZEq:-RESIZE_SIZEq,RESIZE_SIZEq:-RESIZE_SIZEq])\n",
    "#         B11mean=np.mean(B11)\n",
    "#         B11Cmean=np.mean(B11[RESIZE_SIZEq:-RESIZE_SIZEq,RESIZE_SIZEq:-RESIZE_SIZEq])\n",
    "        \n",
    "        \n",
    "#         L1011fs.append([B10mean,B10Cmean,B11mean,B11Cmean])\n",
    "#         idsLf.append(uid)\n",
    "#     except:\n",
    "#         pass\n",
    "\n",
    "# #        \n",
    "            \n",
    "# L1011fs=np.stack(L1011fs)   \n",
    "# idsLf=np.stack(idsLf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5952"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gc.collect()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed('L_t20_a500_c30', a=idsL, b=rgbsL)\n",
    "np.savez_compressed('S_t20_a500_c30', a=idsS, b=rgbsS)\n",
    "# np.savez_compressed('S_t20_a500_c30_8', a=idsS, b=rgbsS) #band8 instead 8A\n",
    "# np.savez_compressed('L_t20_a500_c30_B1011f', a=idsLf, b=L1011fs) #\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "\n",
    "# S_file='S_t20_a200_c30.npz'\n",
    "# L_file='L_t20_a200_c30.npz'\n",
    "S_file='S_t20_a500_c30.npz'\n",
    "# S_file='S_t20_a500_c30_8.npz'\n",
    "# L_file='L_t20_a500_c30.npz'\n",
    "# temp=np.load(L_file)\n",
    "# idsL=temp['a']\n",
    "# rgbsL=temp['b']\n",
    "temp=np.load(S_file)\n",
    "idsS=temp['a']\n",
    "rgbsS=temp['b']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Lf_file='L_t20_a500_c30_B1011f.npz'\n",
    "# temp=np.load(Lf_file)\n",
    "# idsLf=temp['a']\n",
    "# L1011fs=temp['b']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "# idsLfDF=pd.DataFrame(np.concatenate((np.expand_dims(idsLf,-1), L1011fs),-1),columns=['uid', 'LB10mean','LB10Cmean','LB11mean','LB11Cmean'])\n",
    "# idsLfDF[['LB10mean','LB10Cmean','LB11mean','LB11Cmean']]=idsLfDF[['LB10mean','LB10Cmean','LB11mean','LB11Cmean']].astype(float)\n",
    "# idsLfDF.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyOSc2OvTGXSi0VpLOwRmlqC",
   "provenance": [
    {
     "file_id": "1g4Zk_MZ921O9jUl_p7vXYQlJ8K0Dca1A",
     "timestamp": 1671436706092
    },
    {
     "file_id": "1td1A96pr-GFAohrlPKNamiMoFIHDYT1B",
     "timestamp": 1671186589886
    }
   ]
  },
  "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": "3.7.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}