DA_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:  Logger initialized\n"
     ]
    }
   ],
   "source": [
    "import xarray as xr\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "import DA_core as DA\n",
    "from glob import glob\n",
    "import matplotlib.animation as animation\n",
    "from pathlib import Path\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "\n",
    "data_dir='/scratch/cimes/feiyul/PyQG/data'\n",
    "# data_dir='/work/Feiyu.Lu/PyQG/data'\n",
    "work_dir='/home/fnl/PyQG/pyqg_experiments/examples/DA'\n",
    "B_kws={'cmap':'bwr','levels':np.linspace(-0.75E-10,0.75E-10,26),'extend':'both'}\n",
    "B_kws0={'cmap':'bwr','levels':np.linspace(-2.5E-12,2.5E-12,26),'extend':'both'}\n",
    "\n",
    "B_kws1={'cmap':'bwr','levels':np.linspace(-0.75E-13,0.75E-13,26),'extend':'both'}\n",
    "B_ens_kws={'cmap':'bwr','levels':np.linspace(-2.5E-11,2.5E-11,26),'extend':'both'}\n",
    "B_ens_kws0={'cmap':'bwr','levels':np.linspace(-2.5E-13,2.5E-13,26),'extend':'both'}\n",
    "B_ens_kws1={'cmap':'bwr','levels':np.linspace(-0.25E-14,0.25E-14,26),'extend':'both'}\n",
    "q_kws={'cmap':'bwr','levels':np.linspace(-3.4E-5,3.4E-5,18),'extend':'both'}\n",
    "q_kws1={'cmap':'bwr','levels':np.linspace(-6.8E-7,6.8E-7,18),'extend':'both'}\n",
    "inc_kws={'cmap':'bwr','levels':np.linspace(-7.5E-6,7.5E-6,16),'extend':'both'}\n",
    "spread_kws={'cmap':'hot_r','levels':np.linspace(0,20E-6,11),'extend':'both'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "obs_kwargs={'Nx_truth':128,'obs_freq':10,'obs_err':[1,-5,5,-7],'DA_freq':10,'save_B':False,'nobs':[50,50]}\n",
    "\n",
    "DA_training_paras1={'nens':1280,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.45],**obs_kwargs}\n",
    "DA_training1=DA.DA_exp(**DA_training_paras1)\n",
    "training_ens1=DA_training1.nens\n",
    "\n",
    "DA_training_paras2={'nens':80,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.5],**obs_kwargs}\n",
    "DA_training2=DA.DA_exp(**DA_training_paras2)\n",
    "training_ens2=DA_training2.nens\n",
    "\n",
    "DA_training_paras3={'nens':80,'DA_method':'EnKF','Nx_DA':64,'R_W':100,'inflate':[1,0.5],**obs_kwargs}\n",
    "DA_training3=DA.DA_exp(**DA_training_paras3)\n",
    "training_ens3=DA_training3.nens\n",
    "\n",
    "DA_training_paras4={'nens':1280,'DA_method':'EnKF','Nx_DA':64,'R_W':100,'inflate':[1,0.45],**obs_kwargs}\n",
    "DA_training4=DA.DA_exp(**DA_training_paras4)\n",
    "training_ens4=DA_training4.nens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NoDA_Nx64_from_Nx128_nobs50_50_err1E-5_5E-7\n",
      "3DVar_Nx64_from_Nx128_freq10_R75_nobs50_50_err1E-5_5E-7\n",
      "EnKF_Nx64_from_Nx128_ens5_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7\n",
      "EnKF_Nx64_from_Nx128_ens10_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7\n",
      "EnKF_Nx64_from_Nx128_ens20_freq10_relax0.6_R100_nobs50_50_err1E-5_5E-7\n",
      "EnKF_Nx64_from_Nx128_ens40_freq10_relax0.55_R100_nobs50_50_err1E-5_5E-7\n",
      "EnKF_Nx64_from_Nx128_ens80_freq10_relax0.5_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.35_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n"
     ]
    }
   ],
   "source": [
    "kwargs={'Nx_DA':64,'Nx_truth':128,'training_exp':DA_training3,'training_var':'_2L32f','obs_freq':10,'DA_freq':10,\n",
    "        'nobs':[50,50],'obs_err':[1,-5,5,-7],'B_alpha':0.0,'R_training':8,'R_DA':8}\n",
    "kwargs1={'Nx_DA':64,'Nx_truth':128,'training_exp':DA_training2,'training_var':'_2L32f_interp','obs_freq':10,'DA_freq':10,\n",
    "        'nobs':[50,50],'obs_err':[1,-5,5,-7],'B_alpha':0.0,'R_training':8,'R_DA':8}\n",
    "\n",
    "exp_set='DA64'\n",
    "\n",
    "if exp_set=='DA32':\n",
    "  DA_paras=[{'nens':1,'DA_method':'NoDA','Nx_DA':32,'R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','Nx_DA':32,'R_W':50,**kwargs},\n",
    "            {'nens':5,'DA_method':'EnKF','Nx_DA':32,'R_W':50,'inflate':[1,0.7],**kwargs},\n",
    "            {'nens':10,'DA_method':'EnKF','Nx_DA':32,'R_W':75,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':20,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':40,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.58],**kwargs},\n",
    "            {'nens':80,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.5],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],**kwargs1},\n",
    "            {'nens':5,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.45],**kwargs},\n",
    "            {'nens':5,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.45],**kwargs1},\n",
    "            {'nens':10,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs},\n",
    "            {'nens':10,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs1},\n",
    "            {'nens':20,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.35],**kwargs},\n",
    "            {'nens':20,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.35],**kwargs1},\n",
    "            {'nens':40,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.35],**kwargs},\n",
    "            {'nens':40,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.35],**kwargs1},\n",
    "            {'nens':80,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs},\n",
    "            {'nens':80,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs1}]\n",
    "  \n",
    "elif exp_set=='DA64':\n",
    "  DA_paras=[{'nens':1,'DA_method':'NoDA','R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':75,**kwargs},\n",
    "            {'nens':5,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':10,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':20,'DA_method':'EnKF','R_W':100,'inflate':[1,0.6],**kwargs},\n",
    "            {'nens':40,'DA_method':'EnKF','R_W':100,'inflate':[1,0.55],**kwargs},\n",
    "            {'nens':80,'DA_method':'EnKF','R_W':100,'inflate':[1,0.5],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],**kwargs1},\n",
    "            {'nens':5,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs},\n",
    "            {'nens':5,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs1},\n",
    "            {'nens':10,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs},\n",
    "            {'nens':10,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs1},\n",
    "            {'nens':20,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs},\n",
    "            {'nens':20,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs1},\n",
    "            {'nens':40,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.4],**kwargs},\n",
    "            {'nens':40,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs1},\n",
    "            {'nens':80,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.35],**kwargs},\n",
    "            {'nens':80,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.3],**kwargs1}]\n",
    "\n",
    "elif exp_set=='3DVar':\n",
    "  DA_paras=[{'nens':1,'DA_method':'NoDA','R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':5,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':10,**kwargs}]\n",
    "              \n",
    "elif exp_set=='Compare':\n",
    "  DA_paras=[{'nens':1,'DA_method':'NoDA','R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':75,**kwargs},\n",
    "            {'nens':5,'DA_method':'EnKF','R_W':100,'inflate':[1,0.7],**kwargs},\n",
    "            {'nens':10,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':20,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':40,'DA_method':'EnKF','R_W':100,'inflate':[1,0.58],**kwargs},\n",
    "            {'nens':80,'DA_method':'EnKF','R_W':100,'inflate':[1,0.5],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],'B_alpha':0.0,'R_training':8,'R_DA':8,**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],'B_alpha':0.0,'R_training':8,'R_DA':8,**kwargs1}]\n",
    "\n",
    "elif exp_set=='Compare64':\n",
    "  DA_paras=[{'nens':1,'DA_method':'NoDA','R_W':50,**kwargs},\n",
    "            {'nens':1,'DA_method':'3DVar','R_W':75,**kwargs},\n",
    "            {'nens':5,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':10,'DA_method':'EnKF','R_W':100,'inflate':[1,0.65],**kwargs},\n",
    "            {'nens':20,'DA_method':'EnKF','R_W':100,'inflate':[1,0.6],**kwargs},\n",
    "            {'nens':40,'DA_method':'EnKF','R_W':100,'inflate':[1,0.55],**kwargs},\n",
    "            {'nens':80,'DA_method':'EnKF','R_W':100,'inflate':[1,0.5],**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],'B_alpha':0.0,'R_training':12,'R_DA':12,**kwargs},\n",
    "            {'nens':1,'DA_method':'UnetKF','R_W':100,'inflate':[1,0.0],'B_alpha':0.0,'R_training':12,'R_DA':12,**kwargs1}]\n",
    "   \n",
    "DA_exps=[]\n",
    "for DA_para in DA_paras:\n",
    "    DA_exps.append(DA.DA_exp(**DA_para))\n",
    "    print(DA_exps[-1].file_name())\n",
    "colors=sns.color_palette('bright',len(DA_exps))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "importlib.reload(DA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/scratch/cimes/feiyul/PyQG/data//Obs_Nx128_freq10_nobs50_50_err1E-5_5E-7.nc\n",
      "/scratch/cimes/feiyul/PyQG/data/NoDA/EnsMean_NoDA_Nx64_from_Nx128_nobs50_50_err1E-5_5E-7.nc\n",
      "NoDA_Nx64_from_Nx128_nobs50_50_err1E-5_5E-7\n",
      "[1.4344871e-05 2.8877182e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/3DVar/EnsMean_3DVar_Nx64_from_Nx128_freq10_R75_nobs50_50_err1E-5_5E-7.nc\n",
      "3DVar_Nx64_from_Nx128_freq10_R75_nobs50_50_err1E-5_5E-7\n",
      "[1.0328374e-05 2.2366008e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/EnKF/EnsMean_EnKF_Nx64_from_Nx128_ens5_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "EnKF_Nx64_from_Nx128_ens5_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7\n",
      "[9.5033611e-06 2.0118999e-07]\n",
      "[5.0668359e-06 1.1607084e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/EnKF/EnsMean_EnKF_Nx64_from_Nx128_ens10_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "EnKF_Nx64_from_Nx128_ens10_freq10_relax0.65_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.6997061e-06 1.8129361e-07]\n",
      "[6.3776920e-06 1.4549246e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/EnKF/EnsMean_EnKF_Nx64_from_Nx128_ens20_freq10_relax0.6_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "EnKF_Nx64_from_Nx128_ens20_freq10_relax0.6_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.2841989e-06 1.7084426e-07]\n",
      "[6.8510808e-06 1.5788389e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/EnKF/EnsMean_EnKF_Nx64_from_Nx128_ens40_freq10_relax0.55_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "EnKF_Nx64_from_Nx128_ens40_freq10_relax0.55_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.0757463e-06 1.6503638e-07]\n",
      "[7.0264432e-06 1.6149644e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/EnKF/EnsMean_EnKF_Nx64_from_Nx128_ens80_freq10_relax0.5_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "EnKF_Nx64_from_Nx128_ens80_freq10_relax0.5_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.0186310e-06 1.6338143e-07]\n",
      "[7.0206238e-06 1.6248092e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[9.8684450e-06 2.1588448e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens1_freq10_relax0.0_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[1.0120240e-05 2.2322412e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.6901955e-06 1.8080512e-07]\n",
      "[5.8268251e-06 1.3664734e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens5_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.8791667e-06 1.8441034e-07]\n",
      "[6.2126487e-06 1.4115899e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.3473451e-06 1.7094082e-07]\n",
      "[6.4288015e-06 1.5011842e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens10_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.5361789e-06 1.7577291e-07]\n",
      "[6.5665881e-06 1.5027847e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.1666085e-06 1.6660832e-07]\n",
      "[6.6731300e-06 1.5555611e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens20_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.3776622e-06 1.7209643e-07]\n",
      "[6.8103182e-06 1.5591415e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.4_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.0942345e-06 1.6442647e-07]\n",
      "[6.7922178e-06 1.5813582e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens40_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.2836768e-06 1.6937604e-07]\n",
      "[6.9489379e-06 1.5896796e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f/EnsMean_UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.35_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.35_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.0578748e-06 1.6331808e-07]\n",
      "[6.7642677e-06 1.5806123e-07]\n",
      "/scratch/cimes/feiyul/PyQG/data/UnetKF_Nx64_128_ens80_2L32f_interp/EnsMean_UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7.nc\n",
      "UnetKF_Nx64_from_Nx128_ens80_freq10_relax0.3_hybrid0.0_R100_nobs50_50_err1E-5_5E-7\n",
      "[8.2601882e-06 1.6862418e-07]\n",
      "[7.0211659e-06 1.6041035e-07]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ds_truth=DA_exps[-1].read_truth(years=10,interp=True)\n",
    "# q_truth=ds_truth.q.squeeze()\n",
    "B_ds=xr.open_dataset('{}/B_Nx{}_100years_2lev.nc'.format(data_dir,DA_exps[0].Nx_DA))\n",
    "ds_obs=DA_exps[-1].read_obs()\n",
    "\n",
    "days_compute=slice(365*2,365*20)\n",
    "plot_start=365\n",
    "fig, axes=plt.subplots(2,1,figsize=(12,8))\n",
    "rmse=np.empty((len(DA_exps),2))\n",
    "rmse_std=np.empty((len(DA_exps),2))\n",
    "spread=np.empty((len(DA_exps),2))\n",
    "spread_std=np.empty((len(DA_exps),2))\n",
    "\n",
    "for DA_exp,cl,i in zip(DA_exps,colors,np.arange(len(DA_exps))):\n",
    "    ds_truth=DA_exp.read_truth(years=20,interp=True)\n",
    "    q_truth=ds_truth.q.squeeze()\n",
    "    if DA_exp.DA_method=='UnetKF':\n",
    "        folder='{}{}'.format(DA_paras[i]['training_exp'].nens,DA_exp.training_var)\n",
    "    else:\n",
    "        folder=''\n",
    "    ds_mean=DA_exp.read_mean(folder=folder)\n",
    "    if DA_exp.nens>1:  \n",
    "        ds_std=DA_exp.read_std(folder=folder)\n",
    "    q_analysis_mean=ds_mean.q_post\n",
    "    rmse_lev=np.sqrt(((q_analysis_mean-q_truth)**2).mean(axis=(-1,-2)))\n",
    "    rmse_lev_mean=rmse_lev[days_compute,:].mean(axis=0)\n",
    "    rmse_lev_std=rmse_lev[days_compute,:].std(axis=0)\n",
    "    print(DA_exp.file_name())\n",
    "    print(rmse_lev_mean.data.astype(np.single()))\n",
    "    rmse[i,:]=rmse_lev_mean.data\n",
    "    rmse_std[i,:]=rmse_lev_std.data\n",
    "    # rmse_xy=np.sqrt(((q_analysis_mean-q_truth)**2).mean(axis=(0)))\n",
    "    if DA_exp.nens>1:\n",
    "        spread_lev=ds_std.q.mean(('x','y'))\n",
    "        spread_lev_mean=spread_lev[days_compute,:].mean('time')\n",
    "        spread_lev_std=spread_lev[days_compute,:].std('time')\n",
    "        print(spread_lev_mean.data.astype(np.single()))\n",
    "        spread[i,:]=spread_lev_mean.data\n",
    "        spread_std[i,:]=spread_lev_std.data\n",
    "        # spread_xy=ds_std.q.mean(('time'))\n",
    "    else:\n",
    "        spread[i,:]=np.nan\n",
    "\n",
    "    # for i,clim in enumerate([1.5E-5,1.0E-6]):\n",
    "        # axes[i].plot(np.arange(plot_start,len(q_analysis_mean.time)),rmse_lev[plot_start:,i],'-',color=cl,\n",
    "        #                label='{:.2E} {}'.format(rmse_lev_mean[i].values,DA_exp.file_name_short()))\n",
    "        # axes[i].set_title('RMSE (Lev {})'.format(i))\n",
    "        # im=axes[1,i].contourf(ds_mean.x/1.0E3,ds_mean.y/1.0E3,rmse_xy[i,:,:],extend='max',\n",
    "        #                       cmap='hot_r',levels=np.linspace(0,clim,11,endpoint=True))\n",
    "        # axes[1,i].set_title('{} RMSE (Lev {})'.format(DA_exp.DA_method,i))\n",
    "        # plt.colorbar(im,ax=axes[1,i])\n",
    "\n",
    "        # if DA_exp.nens>1:\n",
    "            # axes[i].plot(np.arange(len(q_analysis_mean.time)),spread_lev.isel(lev=i),'--',color=cl,\n",
    "            #                label='Spread {:.2E} {}'.format(spread_lev_mean[i].values,DA_exp.file_name_short()))\n",
    "            # im=axes[2,i].contourf(ds_mean.x/1.0E3,ds_mean.y/1.0E3,spread_xy.isel(lev=i),extend='max',\n",
    "            #                       cmap='hot_r',levels=np.linspace(0,clim,11,endpoint=True))\n",
    "            # axes[0,i].set_title('{} Spread (Lev {})'.format(DA_exp.DA_method,i))\n",
    "            # plt.colorbar(im,ax=axes[2,i])\n",
    "    \n",
    "    # axes[0].legend()\n",
    "    # axes[1].legend()\n",
    "\n",
    "width=0.35\n",
    "x=np.arange(len(DA_exps))\n",
    "labels0=[]\n",
    "labels1=[]\n",
    "for i,DA_exp in enumerate(DA_exps):\n",
    "    labels0.append(DA_exp.file_name_label()+'\\n{:3.2F}'.format(rmse[i,0]*10E5,DA_exp.training_var[1:]))\n",
    "    labels1.append(DA_exp.file_name_label()+'\\n{:3.2F}'.format(rmse[i,1]*10E6,DA_exp.training_var[1:]))\n",
    "for i,ax in enumerate(axes):\n",
    "    if exp_set=='DA32':\n",
    "        ax.bar(x[0] - width/2,rmse[0,i],width, label='Control',edgecolor=colors[0],color=colors[0])\n",
    "        # ax.errorbar(x[0] - width/2, rmse[0,i], yerr=rmse_std[0,i], ecolor='black',elinewidth=3,capsize=5)\n",
    "        \n",
    "        ax.bar(x[1] - width/2,rmse[1,i],width, label='3DVar',edgecolor=colors[1],color=colors[1])\n",
    "        # ax.errorbar(x[1] - width/2, rmse[1,i], yerr=rmse_std[1,i], ecolor='black',elinewidth=3,capsize=5)\n",
    "\n",
    "        ax.bar(x[2:7] - width/2,rmse[2:7,i],width, label='EnKF',edgecolor=colors[2],color=colors[2])\n",
    "        # ax.errorbar(x[2:7] - width/2, rmse[2:7,i], yerr=rmse_std[2:7,i],ls='',ecolor='black',elinewidth=2,capsize=4)\n",
    "                \n",
    "        ind1=[i for i in range(7,18,2)]\n",
    "        ax.bar(x[ind1] - width/2,rmse[ind1,i],width, label='UNetKF: UNet_L1',edgecolor=colors[3],color=colors[3])\n",
    "        ax.bar(x[ind1] + width/2,spread[ind1,i],width,edgecolor=colors[3],color=colors[3],fill=False,hatch='/')\n",
    "        \n",
    "        ind2=[i for i in range(8,19,2)]\n",
    "        ax.bar(x[ind2] - width/2,rmse[ind2,i],width, label='UNetKF: UNet_L2',edgecolor=colors[4],color=colors[4])\n",
    "        ax.bar(x[ind2] + width/2,spread[ind2,i],width,edgecolor=colors[4],color=colors[4],fill=False,hatch='/')\n",
    "\n",
    "        ax.bar(x[2:7] + width/2,spread[2:7,i],width, label='Spread',edgecolor='k',color='k',fill=False,hatch='/')\n",
    "        ax.bar(x[2:7] + width/2,spread[2:7,i],width,edgecolor=colors[2],color=colors[2],fill=False,hatch='/')\n",
    "        \n",
    "        print(rmse[1,i], rmse_std[1,i], rmse[7,i], rmse_std[7,i])\n",
    "        print(stats.ttest_ind_from_stats(rmse[1,i], rmse_std[1,i], 365*18, rmse[7,i], rmse_std[7,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[1,i], rmse_std[1,i], 365*18, rmse[8,i], rmse_std[8,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[2,i], rmse_std[2,i], 365*18, rmse[9,i], rmse_std[9,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[2,i], rmse_std[2,i], 365*18, rmse[10,i], rmse_std[10,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[3,i], rmse_std[3,i], 365*18, rmse[11,i], rmse_std[11,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[3,i], rmse_std[3,i], 365*18, rmse[12,i], rmse_std[12,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[4,i], rmse_std[4,i], 365*18, rmse[13,i], rmse_std[13,i], 365*18))\n",
    "        print(stats.ttest_ind_from_stats(rmse[4,i], rmse_std[4,i], 365*18, rmse[14,i], rmse_std[14,i], 365*18))\n",
    "    elif exp_set=='DA64':\n",
    "        ax.bar(x[0] - width/2,rmse[0,i],width, label='Control',edgecolor=colors[0],color=colors[0])\n",
    "\n",
    "        ax.bar(x[1] - width/2,rmse[1,i],width, label='3DVar',edgecolor=colors[1],color=colors[1])\n",
    "\n",
    "        ax.bar(x[2:7] - width/2,rmse[2:7,i],width, label='EnKF',edgecolor=colors[2],color=colors[2])\n",
    "        \n",
    "        ind1=[i for i in range(7,18,2)]        \n",
    "        ax.bar(x[ind1] - width/2,rmse[ind1,i],width, label='UNetKF: UNet_H',edgecolor=colors[3],color=colors[3])\n",
    "        ax.bar(x[ind1] + width/2,spread[ind1,i],width,edgecolor=colors[3],color=colors[3],fill=False,hatch='/')\n",
    "        \n",
    "        ind2=[i for i in range(8,19,2)]\n",
    "        ax.bar(x[ind2] - width/2,rmse[ind2,i],width, label='UNetKF: UNet_L2',edgecolor=colors[4],color=colors[4])\n",
    "        ax.bar(x[ind2] + width/2,spread[ind2,i],width,edgecolor=colors[4],color=colors[4],fill=False,hatch='/')\n",
    "        \n",
    "        ax.bar(x[2:7] + width/2,spread[2:7,i],width, label='Spread',edgecolor='k',color='k',fill=False,hatch='/')\n",
    "        ax.bar(x[2:7] + width/2,spread[2:7,i],width,edgecolor=colors[2],color=colors[2],fill=False,hatch='/')\n",
    "    # elif exp_set=='UnetCompare':\n",
    "    #     ax.bar(x[[0,4,8]] - width/2,rmse[[0,4,8],i],width, label=DA_exps[0].training_var[1:],edgecolor=colors[0],color=colors[0])\n",
    "    #     ax.bar(x[[0,4,8]] + width/2,spread[[0,4,8],i],width,edgecolor=colors[0],color=colors[0],fill=False,hatch='/')\n",
    "        \n",
    "    #     ax.bar(x[[1,5,9]] - width/2,rmse[[1,5,9],i],width, label=DA_exps[1].training_var[1:],edgecolor=colors[1],color=colors[1])\n",
    "    #     ax.bar(x[[1,5,9]] + width/2,spread[[1,5,9],i],width,edgecolor=colors[1],color=colors[1],fill=False,hatch='/')\n",
    "        \n",
    "    #     ax.bar(x[[2,6,10]] - width/2,rmse[[2,6,10],i],width, label=DA_exps[2].training_var[1:],edgecolor=colors[2],color=colors[2])\n",
    "    #     ax.bar(x[[2,6,10]] + width/2,spread[[2,6,10],i],width,edgecolor=colors[2],color=colors[2],fill=False,hatch='/')\n",
    "        \n",
    "    #     ax.bar(x[[3,7,11]] - width/2,rmse[[3,7,11],i],width, label=DA_exps[3].training_var[1:],edgecolor=colors[3],color=colors[3])\n",
    "    #     ax.bar(x[[3,7,11]] + width/2,spread[[3,7,11],i],width,edgecolor=colors[3],color=colors[3],fill=False,hatch='/')        \n",
    "    else:\n",
    "        ax.bar(x - width/2,rmse[:,i],width)\n",
    "        ax.bar(x + width/2,spread[:,i],width)\n",
    "        \n",
    "    ax.errorbar(x[0:len(rmse_std[:,i])] - width/2, rmse[:,i], yerr=rmse_std[:,i],ls='',ecolor='black',elinewidth=2,capsize=4)\n",
    "    ax.grid(which='both',axis='y',linestyle=':')\n",
    "    if i==0:\n",
    "        ax.legend() \n",
    "        \n",
    "axes[0].set_title('(b) QG-H DA Experiments: Upper Layer RMSE')\n",
    "axes[1].set_title('Lower Layer')\n",
    "axes[0].set_xticks(x, labels0)\n",
    "axes[1].set_xticks(x, labels1)\n",
    "        \n",
    "fig.tight_layout()\n",
    "fig.savefig('./plots/RMSE_Spread_{}_TrainEns{}{}_DA{}_Truth{}_nobs{}.png'.\\\n",
    "    format(exp_set,kwargs['training_exp'].nens,kwargs['training_var'],DA_exps[-1].Nx_DA,DA_exps[-1].Nx_truth,'_'.join(map(str,DA_exps[-1].nobs))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'DA_training1' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb Cell 6\u001b[0m in \u001b[0;36m<cell line: 3>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=0'>1</a>\u001b[0m DA\u001b[39m.\u001b[39mread_data_dir\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39m/scratch/cimes/feiyul/PyQG/data\u001b[39m\u001b[39m'\u001b[39m\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=1'>2</a>\u001b[0m DA\u001b[39m.\u001b[39msave_data_dir\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39m/scratch/cimes/feiyul/PyQG/data\u001b[39m\u001b[39m'\u001b[39m\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=2'>3</a>\u001b[0m DA_training_paras\u001b[39m=\u001b[39m{\u001b[39m'\u001b[39m\u001b[39mnens\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m1\u001b[39m,\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=3'>4</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mDA_method\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m'\u001b[39m\u001b[39mUnetKF\u001b[39m\u001b[39m'\u001b[39m,\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=4'>5</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mNx_DA\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m64\u001b[39m,\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=5'>6</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mNx_truth\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m128\u001b[39m,\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=6'>7</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mobs_freq\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m10\u001b[39m,\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=7'>8</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mobs_err\u001b[39m\u001b[39m'\u001b[39m:[\u001b[39m1\u001b[39m,\u001b[39m-\u001b[39m\u001b[39m5\u001b[39m,\u001b[39m5\u001b[39m,\u001b[39m-\u001b[39m\u001b[39m7\u001b[39m],\n\u001b[1;32m      <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=8'>9</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mDA_freq\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m10\u001b[39m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=9'>10</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39msave_B\u001b[39m\u001b[39m'\u001b[39m:\u001b[39mFalse\u001b[39;00m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=10'>11</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mnobs\u001b[39m\u001b[39m'\u001b[39m:[\u001b[39m50\u001b[39m,\u001b[39m50\u001b[39m],\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=11'>12</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mR_W\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m100\u001b[39m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=12'>13</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39minflate\u001b[39m\u001b[39m'\u001b[39m:[\u001b[39m1\u001b[39m,\u001b[39m0.45\u001b[39m],\n\u001b[0;32m---> <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=13'>14</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mtraining_exp\u001b[39m\u001b[39m'\u001b[39m:DA_training1,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=14'>15</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mtraining_var\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m'\u001b[39m\u001b[39m_2L32f\u001b[39m\u001b[39m'\u001b[39m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=15'>16</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mB_alpha\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m0.0\u001b[39m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=16'>17</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mR_training\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m12\u001b[39m,\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=17'>18</a>\u001b[0m                    \u001b[39m'\u001b[39m\u001b[39mR_DA\u001b[39m\u001b[39m'\u001b[39m:\u001b[39m12\u001b[39m}\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=18'>19</a>\u001b[0m DA_training\u001b[39m=\u001b[39mDA\u001b[39m.\u001b[39mDA_exp(\u001b[39m*\u001b[39m\u001b[39m*\u001b[39mDA_training_paras)\n\u001b[1;32m     <a href='vscode-notebook-cell://ssh-remote%2Bstellar-vis/home/feiyul/PyQG/pyqg_experiments/examples/DA/DA_analysis.ipynb#W5sdnNjb2RlLXJlbW90ZQ%3D%3D?line=20'>21</a>\u001b[0m ds_truth\u001b[39m=\u001b[39mDA_training\u001b[39m.\u001b[39mread_truth(years\u001b[39m=\u001b[39m\u001b[39m10\u001b[39m,interp\u001b[39m=\u001b[39m\u001b[39mTrue\u001b[39;00m)\n",
      "\u001b[0;31mNameError\u001b[0m: name 'DA_training1' is not defined"
     ]
    }
   ],
   "source": [
    "DA.read_data_dir='/scratch/cimes/feiyul/PyQG/data/training'\n",
    "DA.save_data_dir='/scratch/cimes/feiyul/PyQG/data/training'\n",
    "DA_training_paras={'nens':1280,\n",
    "                   'DA_method':'EnKF',\n",
    "                   'Nx_DA':64,\n",
    "                   'Nx_truth':128,\n",
    "                   'obs_freq':10,\n",
    "                   'obs_err':[1,-5,5,-7],\n",
    "                   'DA_freq':10,\n",
    "                   'save_B':False,\n",
    "                   'nobs':[50,50],\n",
    "                   'R_W':100,\n",
    "                   'inflate':[1,0.45],\n",
    "                   'training_exp':DA_training1,\n",
    "                   'training_var':'_3L8f',\n",
    "                   'B_alpha':0.0,\n",
    "                   'R_training':8,\n",
    "                   'R_DA':8}\n",
    "DA_training=DA.DA_exp(**DA_training_paras)\n",
    "\n",
    "ds_truth=DA_training.read_truth(years=10,interp=True)\n",
    "q_truth=ds_truth.q.squeeze()\n",
    "if DA_training.DA_method=='UnetKF':\n",
    "    folder='{}{}'.format(DA_training_paras['training_exp'].nens,DA_training.training_var)\n",
    "else:\n",
    "    folder=''\n",
    "ds_mean=DA_training.read_mean(folder=folder)\n",
    "rmse_lev=np.sqrt(((ds_mean.q_post-q_truth)**2).mean(axis=(-1,-2)))\n",
    "print(rmse_lev.isel(time=slice(730,None)).mean('time'))\n",
    "\n",
    "if DA_training_paras['nens']>1:\n",
    "    ds_std=DA_training.read_std()\n",
    "    spread_lev=ds_std.q.mean(('x','y'))\n",
    "    print(spread_lev.isel(time=slice(73,None)).mean('time'))\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nx_truth 64\n",
    "NoDA [1.2792072e-05 2.8608397e-07]\n",
    "\n",
    "3DVar \n",
    "R25 [6.0717052e-06 1.6883095e-07]\n",
    "R50 [5.1855136e-06 1.5625071e-07]\n",
    "R100 [5.8226915e-06 1.6983651e-07]\n",
    "R150 [6.0276134e-06 1.7305297e-07]\n",
    "\n",
    "EnKF\n",
    "50\n",
    "R150 0.7\n",
    "[2.9184437e-06 7.3589426e-08]\n",
    "[3.2488447e-06 7.0400063e-08]\n",
    "\n",
    "500\n",
    "R150 0.4\n",
    "[1.9683214e-06 4.0464720e-08]\n",
    "[2.2011977e-06 4.5205457e-08]\n",
    "\n",
    "UnetKF\n",
    "[4.8861693e-06 1.2738903e-07]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "B_ens_mean_ds=xr.open_dataset('{}/{}/B_ens_mean.nc'.format(data_dir,DA_exp.file_name()))\n",
    "B_ens_mean=B_ens_mean_ds.B_ens\n",
    "\n",
    "i_x,i_y=30,20\n",
    "B_loc=int((len(B_ens_mean.x_d)-1)/2)\n",
    "\n",
    "fig, axes=plt.subplots(2,3,figsize=(15,8))\n",
    "xy0=i_x+i_y*DA_exp.Nx_DA\n",
    "xy1=i_x+i_y*DA_exp.Nx_DA+DA_exp.Nx_DA*DA_exp.Nx_DA\n",
    "cov_xy0=B_ds.cov.data[xy0,:].reshape(2,DA_exp.Nx_DA,DA_exp.Nx_DA)\n",
    "cov_xy1=B_ds.cov.data[xy1,:].reshape(2,DA_exp.Nx_DA,DA_exp.Nx_DA)\n",
    "x_loc=np.arange(-B_loc,B_loc+1)\n",
    "y_loc=np.arange(-B_loc,B_loc+1)\n",
    "\n",
    "im=axes[0,0].contourf(x_loc,y_loc,cov_xy0[0,i_y-B_loc:i_y+B_loc+1,i_x-B_loc:i_x+B_loc+1],**B_kws)\n",
    "plt.colorbar(im,ax=axes[0,0])\n",
    "axes[0,0].set_title('Lev 0-0 cov (y={},x={})'.format(i_y,i_x))\n",
    "\n",
    "im=axes[0,1].contourf(x_loc,y_loc,cov_xy0[1,i_y-B_loc:i_y+B_loc+1,i_x-B_loc:i_x+B_loc+1],**B_kws0)\n",
    "plt.colorbar(im,ax=axes[0,1])\n",
    "axes[0,1].set_title('Lev 0-1 cov (y={},x={})'.format(i_y,i_x))\n",
    "\n",
    "im=axes[0,2].contourf(x_loc,y_loc,cov_xy1[1,i_y-B_loc:i_y+B_loc+1,i_x-B_loc:i_x+B_loc+1],**B_kws1)\n",
    "plt.colorbar(im,ax=axes[0,2])\n",
    "axes[0,2].set_title('Lev 1-1 cov (y={},x={})'.format(i_y,i_x))\n",
    "\n",
    "im=axes[1,0].contourf(x_loc,y_loc,B_ens_mean.isel(x=i_x,y=i_y,lev=0,lev_d=0).squeeze(),**B_ens_kws)\n",
    "plt.colorbar(im,ax=axes[1,0])\n",
    "axes[1,0].set_title('Lev 0-0 ens cov (y={},x={})'.format(i_y,i_x))\n",
    "\n",
    "im=axes[1,1].contourf(x_loc,y_loc,B_ens_mean.isel(x=i_x,y=i_y,lev=0,lev_d=1).squeeze(),**B_ens_kws0)\n",
    "plt.colorbar(im,ax=axes[1,1])\n",
    "axes[1,1].set_title('Lev 0-1 ens cov (y={},x={})'.format(i_y,i_x))\n",
    "\n",
    "im=axes[1,2].contourf(x_loc,y_loc,B_ens_mean.isel(x=i_x,y=i_y,lev=1,lev_d=1).squeeze(),**B_ens_kws1)\n",
    "plt.colorbar(im,ax=axes[1,2])\n",
    "axes[1,2].set_title('Lev 1-1 ens cov (y={},x={})'.format(i_y,i_x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_B(t,axes):\n",
    "    t_DA=int((t-9)/10)\n",
    "    # im=[[0,0,0],[0,0,0]]\n",
    "    im=[[0,0],[0,0]]\n",
    "    xy=i_x+i_y*DA_exp.Nx_DA+i_l*DA_exp.Nx_DA*DA_exp.Nx_DA\n",
    "    cov_xy=xr.DataArray(B_ds.cov.data[xy,:].reshape(2,DA_exp.Nx_DA,DA_exp.Nx_DA),\n",
    "                        coords=[B_ens_ds.lev,B_ens_ds.y,B_ens_ds.x])\n",
    "    B_R=int((len(B_ens_ds.x_d)-1)/2)\n",
    "    \n",
    "    # cov_loc=DA.localize_q(cov_xy,i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    # im[1][1]=axes[1,1].contourf(B_ens_ds.x_d,B_ens_ds.y_d,\n",
    "    #                             cov_loc[i_l,:,:],**B_kws)\n",
    "    # axes[1,1].set_title('Static B{}'.format(i_l))\n",
    "\n",
    "    im[0][1]=axes[0,1].contourf(B_ens_ds.x_d,B_ens_ds.y_d,\n",
    "                            B_ens_ds.B_ens.isel(time=t_DA,x=i_x,y=i_y,lev=i_l,lev_d=i_l),\n",
    "                            **B_ens_kws)\n",
    "    axes[0,1].set_title('Ensemble Covariance'.format(i_l))\n",
    "\n",
    "    t_day=t\n",
    "    q_loc=DA.localize_q(ds_mean.q_post.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    im[0][0]=axes[0,0].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_loc,**q_kws)\n",
    "    axes[0,0].set_title('Analysis q (lev={},y={},x={})'.format(i_l,i_y,i_x))\n",
    "    \n",
    "    q_truth=DA.localize_q(ds_truth.q.isel(time=t_day,lev=i_l).squeeze(),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    im[1][0]=axes[1,0].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_truth,**q_kws)\n",
    "    axes[1,0].set_title('True q (subsampled)')\n",
    "\n",
    "#     q_spread=DA.localize_q(ds_std.q.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "#     im[1][2]=axes[1,2].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_spread,**spread_kws)\n",
    "#     axes[1,2].set_title('Spread q')\n",
    "    \n",
    "    q_inc=DA.localize_q(ds_mean.q_post.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)-\\\n",
    "        DA.localize_q(ds_mean.q.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    im[1][1]=axes[1,1].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_inc,**inc_kws)\n",
    "    axes[1,1].set_title('Increments')\n",
    "    \n",
    "#     for ax in axes.flatten():\n",
    "#         xt_lim=int(B_R/4)*4\n",
    "#         ax.set_xticks(np.arange(-xt_lim,xt_lim+1,4).astype(np.int_))\n",
    "#         ax.set_yticks(np.arange(-xt_lim,xt_lim+1,4).astype(np.int_))\n",
    "#         ax.grid(linestyle=':')\n",
    "        \n",
    "    obs_day=t\n",
    "    # obs_li=ds_obs.li.isel(time=obs_day).data\n",
    "    obs_x=ds_obs.x[ds_obs.xi.sel(day=obs_day).data]\n",
    "    obs_y=ds_obs.y[ds_obs.yi.sel(day=obs_day).data]\n",
    "    x_DA,y_DA=ds_mean.x,ds_mean.y\n",
    "    x_d=x_DA[1]-x_DA[0]\n",
    "    y_d=y_DA[1]-y_DA[0]\n",
    "    x_DA_ex=np.concatenate((np.array([x_DA[0]-x_d]),x_DA,np.array([x_DA[-1]+x_d])))\n",
    "    y_DA_ex=np.concatenate((np.array([y_DA[0]-y_d]),y_DA,np.array([y_DA[-1]+y_d])))\n",
    "    obs_xi=np.interp(obs_x,x_DA_ex,np.arange(-1,DA_exp.Nx_DA+1))\n",
    "    obs_yi=np.interp(obs_y,y_DA_ex,np.arange(-1,DA_exp.Nx_DA+1))\n",
    "    obs_domain=np.logical_and(np.logical_and(obs_xi >= i_x-B_R,obs_xi <= i_x+B_R),\n",
    "                              np.logical_and(obs_yi >= i_y-B_R,obs_yi <= i_y+B_R))\n",
    "    if np.any(obs_domain):\n",
    "        axes[1,1].scatter(obs_xi[obs_domain]-i_x,obs_yi[obs_domain]-i_y,c='k')\n",
    "        \n",
    "    return im\n",
    "\n",
    "def plot_q(t,axes):\n",
    "    im=[[0,0],[0,0]]\n",
    "    \n",
    "    q_truth0=ds_truth.q.isel(time=t,lev=0).squeeze()\n",
    "    im[0][0]=axes[0,0].contourf(q_truth.x/1000,q_truth.y/1000,q_truth0,**q_kws)\n",
    "    axes[0,0].set_title('True q0')\n",
    "           \n",
    "    q_truth1=ds_truth.q.isel(time=t,lev=1).squeeze()\n",
    "    im[0][1]=axes[0,1].contourf(q_truth.x/1000,q_truth.y/1000,q_truth1,**q_kws1)\n",
    "    axes[0,1].set_title('True q1')\n",
    "\n",
    "    q_mean0=ds_mean.q.isel(time=t,lev=0).squeeze()\n",
    "    im[1][0]=axes[1,0].contourf(q_truth.x/1000,q_truth.y/1000,q_mean0,**q_kws)\n",
    "    axes[1,0].set_title('Analysis q0')\n",
    "           \n",
    "    q_mean1=ds_mean.q.isel(time=t,lev=1).squeeze()\n",
    "    im[1][1]=axes[1,1].contourf(q_truth.x/1000,q_truth.y/1000,q_mean1,**q_kws1)\n",
    "    axes[1,1].set_title('Analysis q1')\n",
    "    \n",
    "    return im\n",
    "\n",
    "def compare_B(t,axes):\n",
    "    im=[[0,0,0],[0,0,0],[0,0,0]]\n",
    "    t_DA=int((t-9)/10)\n",
    "    B_ens=B_ens_ds.B_ens.isel(time=t_DA,x=i_x,y=i_y,lev=i_l,lev_d=i_l)\n",
    "    im[0][0]=axes[0,0].contourf(B_ens_ds.x_d,B_ens_ds.y_d,B_ens,**B_ens_kws)\n",
    "    axes[0,0].set_title('Ens Cov {}'.format(i_l))\n",
    "\n",
    "    t_day=t\n",
    "    q_inst=DA.localize_q(ds_mean.q_post.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    im[1][0]=axes[1,0].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_inst,**q_kws)\n",
    "    axes[1,0].set_title('Ensemble-mean q')\n",
    "\n",
    "    q_inc=DA.localize_q(ds_mean.q_post.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)-\\\n",
    "        DA.localize_q(ds_mean.q.isel(time=t_day,lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "    im[2][0]=axes[2,0].contourf(B_ens_ds.x_d,B_ens_ds.y_d,q_inc,**inc_kws)\n",
    "    axes[2,0].set_title('q inc')\n",
    "\n",
    "    # obs_day=t_day\n",
    "    # # obs_li=ds_obs.li.isel(time=obs_day).data\n",
    "    # obs_x=ds_obs.x[ds_obs.xi.sel(day=obs_day).data]\n",
    "    # obs_y=ds_obs.y[ds_obs.yi.sel(day=obs_day).data]\n",
    "    # x_DA,y_DA=ds_mean.x,ds_mean.y\n",
    "    # x_d=x_DA[1]-x_DA[0]\n",
    "    # y_d=y_DA[1]-y_DA[0]\n",
    "    # x_DA_ex=np.concatenate((np.array([x_DA[0]-x_d]),x_DA,np.array([x_DA[-1]+x_d])))\n",
    "    # y_DA_ex=np.concatenate((np.array([y_DA[0]-y_d]),y_DA,np.array([y_DA[-1]+y_d])))\n",
    "    # obs_xi=np.interp(obs_x,x_DA_ex,np.arange(-1,Nx_DA+1))\n",
    "    # obs_yi=np.interp(obs_y,y_DA_ex,np.arange(-1,Nx_DA+1))\n",
    "    # obs_domain=np.logical_and(np.logical_and(obs_xi >= i_x-B_R,obs_xi <= i_x+B_R),\n",
    "    #                           np.logical_and(obs_yi >= i_y-B_R,obs_yi <= i_y+B_R))\n",
    "    # if np.any(obs_domain):\n",
    "    #     axes[2,0].scatter(obs_xi[obs_domain]-i_x,obs_yi[obs_domain]-i_y,c='k')\n",
    "\n",
    "    for i,lag in enumerate([30,20,10]):\n",
    "        q_truth=DA.localize_q(ds_truth.q.isel(time=slice(t_day-lag,t_day+1),lev=i_l).squeeze(),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "        q_mean=DA.localize_q(ds_mean.q.isel(time=slice(t_day-lag,t_day+1),lev=i_l),i_y,i_x,DA_exp.Nx_DA,B_R)\n",
    "\n",
    "        cov_truth=((q_truth.isel(x=B_R,y=B_R)-q_truth.isel(x=B_R,y=B_R).mean('time'))*(q_truth-q_truth.mean('time'))).sum('time')/(lag-1)\n",
    "        im[i][1]=axes[i,1].contourf(B_ens_ds.x_d,B_ens_ds.y_d,cov_truth,**B_ens_kws)\n",
    "        axes[i,1].set_title('Cov Truth (Trailing {}d)'.format(lag))\n",
    "\n",
    "        cov_mean=((q_mean.isel(x=B_R,y=B_R)-q_mean.isel(x=B_R,y=B_R).mean('time'))*(q_mean-q_mean.mean('time'))).sum('time')/(lag-1)\n",
    "        im[i][2]=axes[i,2].contourf(B_ens_ds.x_d,B_ens_ds.y_d,cov_mean,**B_ens_kws)\n",
    "        axes[i,2].set_title('Cov Analysis (Trailing {}d)'.format(lag))\n",
    "        \n",
    "    for ax in axes.flatten():\n",
    "        xt_lim=int(B_R/4)*4\n",
    "        ax.set_xticks(np.arange(-xt_lim,xt_lim+1,4).astype(np.int_))\n",
    "        ax.set_yticks(np.arange(-xt_lim,xt_lim+1,4).astype(np.int_))\n",
    "        ax.grid(linestyle=':')\n",
    "\n",
    "    return im\n",
    "\n",
    "def animate_B(t,axes,func):\n",
    "    for ax in axes.flatten():\n",
    "        ax.clear()\n",
    "        ax.set_title('')\n",
    "    im=func(t,axes)\n",
    "    \n",
    "    return im\n",
    "\n",
    "def colorbar_list_of_lists(im,axes):\n",
    "    for i,sub_list in enumerate(im):\n",
    "            for j,sub_item in enumerate(sub_list):\n",
    "                    if sub_item!=0:\n",
    "                            plt.colorbar(sub_item,ax=axes[i,j])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# DA_exp=DA_exps[1]\n",
    "paras={'nens':1280,'DA_method':'EnKF','Nx_DA':32,'Nx_truth':128,'obs_freq':10,'obs_err':[1,-6,5,-8],'nobs':[100,20],'R_W':100,'DA_freq':10,'inflate':[1,0.7]}\n",
    "DA_exp=DA.DA_exp(**paras)\n",
    "B_ens_ds=xr.open_dataset('{}/training/{}/B_ens.nc'.format(data_dir,DA_exp.file_name()))\n",
    "B_ds=xr.open_dataset('{}/B_Nx{}_100years_2lev.nc'.format(data_dir,DA_exp.Nx_DA))\n",
    "ds_truth=DA_exp.read_truth(years=10,interp=True)\n",
    "ds_mean=DA_exp.read_mean()\n",
    "ds_std=DA_exp.read_std()\n",
    "ds_obs=DA_exp.read_obs()\n",
    "fig, axes=plt.subplots(2,2,figsize=(10,8))\n",
    "\n",
    "i_x=15\n",
    "i_y=10\n",
    "i_l=0\n",
    "im=plot_B(2009,axes)\n",
    "colorbar_list_of_lists(im,axes)\n",
    "\n",
    "ani = animation.FuncAnimation(fig, animate_B, frames=np.arange(2009,3000,10), \n",
    "    fargs=(axes,plot_B,), interval=200)\n",
    "\n",
    "ani.save(\"./plots/B_small_x{}y{}l{}_{}.gif\".format(i_x,i_y,i_l,DA_exp.file_name()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-04-13T18:10:31.068599Z",
     "iopub.status.busy": "2022-04-13T18:10:31.068100Z",
     "iopub.status.idle": "2022-04-13T18:10:31.186669Z",
     "shell.execute_reply": "2022-04-13T18:10:31.185103Z",
     "shell.execute_reply.started": "2022-04-13T18:10:31.068536Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "DA_days=np.arange(9,3650,10)\n",
    "B_ens_ds=xr.open_mfdataset(['{}/{}/B_ens_day{:04d}.nc'.format(data_dir,DA_exp.file_name(),day) for day in np.arange(9,3650,10)])\n",
    "Path('{}/plots/{}'.format(work_dir,DA_exp.file_name())).mkdir(exist_ok=True)\n",
    "print(B_ens_ds)\n",
    "func=plot_B\n",
    "func_str='ens'\n",
    "\n",
    "for i_x,i_y,i_l in zip([30,30,30],[12,24,36],[0,0,0]):\n",
    "    B_R=int((len(B_ens_ds.x_d)-1)/2)\n",
    "\n",
    "    fig, axes=plt.subplots(2,3,figsize=(15,8))\n",
    "\n",
    "    im=func(729,axes)\n",
    "    colorbar_list_of_lists(im,axes)\n",
    "\n",
    "    ani = animation.FuncAnimation(fig, animate_B, frames=DA_days, \n",
    "        fargs=(axes,func,), interval=100)\n",
    "\n",
    "    ani.save(\"./plots/{}/B_{}_x{}_y{}_l{}.gif\".format(DA_exp.file_name(),func_str,i_x,i_y,i_l))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/2L16f/slurm-801854.out\n",
      "454\n",
      "[0.7153, 0.4767, 0.4393, 0.4151, 0.3945, 0.381, 0.3663, 0.3507, 0.3449, 0.3359, 0.3295, 0.3241, 0.3219, 0.3159, 0.3119, 0.3104, 0.3074, 0.3048, 0.3066, 0.3018, 0.3103, 0.3004, 0.2969, 0.3032, 0.2955, 0.2912, 0.2925, 0.2895, 0.2887, 0.2969, 0.2856, 0.2866, 0.284, 0.2982, 0.2824, 0.2832, 0.2822, 0.2814, 0.2802, 0.286, 0.2789, 0.2782, 0.2791, 0.2806, 0.2774, 0.2941, 0.2782, 0.2777, 0.2817, 0.2758, 0.2757, 0.2745, 0.2848, 0.2739, 0.2846, 0.2735, 0.2744, 0.2781, 0.2777, 0.2739, 0.2724, 0.2742, 0.2754, 0.2737, 0.2729, 0.273, 0.2733, 0.2725, 0.2717, 0.2732, 0.2727, 0.2725, 0.2728, 0.2754, 0.2711, 0.2752, 0.2713, 0.2728, 0.2711, 0.2704, 0.2707, 0.2728, 0.2723, 0.2714, 0.2727, 0.2706, 0.2703, 0.2714, 0.2693, 0.2734, 0.27, 0.2703, 0.2699, 0.2726, 0.27, 0.2724, 0.2708, 0.2717, 0.2692, 0.272, 0.2719, 0.2715, 0.27, 0.273, 0.2698, 0.2695, 0.2704, 0.2724, 0.2712, 0.2727, 0.2691, 0.27, 0.2692, 0.2748, 0.2711, 0.2708, 0.2712, 0.2698, 0.2747, 0.2718, 0.2703, 0.2703, 0.2682, 0.2695, 0.2693, 0.2691, 0.2691, 0.2709, 0.2759, 0.2708, 0.2681, 0.271, 0.2699, 0.2697, 0.2719, 0.2695, 0.2709, 0.2706, 0.2688, 0.2687, 0.2708, 0.2743, 0.2698, 0.271, 0.2723, 0.2698, 0.2707, 0.2679, 0.2668, 0.2695, 0.2688, 0.2689, 0.2712, 0.2708, 0.2711, 0.2687, 0.2675, 0.2692, 0.2683, 0.2689, 0.2728, 0.2679, 0.2679, 0.2732, 0.2681, 0.2681, 0.2688, 0.2695, 0.2679, 0.2692, 0.2696, 0.2669, 0.2686, 0.2667, 0.2688, 0.2687, 0.2673, 0.267, 0.2685, 0.2694, 0.2686, 0.2721, 0.2671, 0.2677, 0.2729, 0.2674, 0.2676, 0.2689, 0.2665, 0.2683, 0.2671, 0.2673, 0.2703, 0.2685, 0.2692, 0.2658, 0.2656, 0.2693, 0.2686, 0.2695]\n",
      "0.2656 197 0.191\n",
      "0.2695 0.1914\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/2L32f/slurm-801858.out\n",
      "454\n",
      "[1.1677, 0.5673, 0.4471, 0.4207, 0.3948, 0.364, 0.3471, 0.3311, 0.3168, 0.3227, 0.3042, 0.2957, 0.294, 0.2891, 0.2857, 0.2853, 0.3266, 0.2831, 0.2829, 0.2905, 0.2864, 0.2751, 0.2741, 0.2757, 0.2709, 0.2717, 0.2753, 0.2688, 0.2689, 0.2657, 0.2682, 0.2674, 0.2668, 0.2639, 0.2739, 0.265, 0.2656, 0.2655, 0.2612, 0.268, 0.2703, 0.26, 0.2688, 0.265, 0.2593, 0.259, 0.2602, 0.2601, 0.2602, 0.2737, 0.2601, 0.2561, 0.2566, 0.2561, 0.2605, 0.259, 0.2614, 0.2565, 0.2559, 0.2629, 0.2575, 0.2586, 0.2567, 0.2548, 0.256, 0.2536, 0.2549, 0.2627, 0.257, 0.2576, 0.2547, 0.2549, 0.2594, 0.2576, 0.2546, 0.2556, 0.2543, 0.2553, 0.2615, 0.2547, 0.2544, 0.2538, 0.2566, 0.2575, 0.2524, 0.2596, 0.2523, 0.2533, 0.2554, 0.255, 0.2586, 0.2535, 0.253, 0.258, 0.2529, 0.2553, 0.2549, 0.2657, 0.2548, 0.2677, 0.2531, 0.2526, 0.2545, 0.2545, 0.2516, 0.2537, 0.2531, 0.2536, 0.2521, 0.2532, 0.2535, 0.2578, 0.2532, 0.2579, 0.2553, 0.2558, 0.2535, 0.2545, 0.2591, 0.2543, 0.2534, 0.2534, 0.256, 0.254, 0.2543, 0.2532, 0.2588, 0.2527, 0.2541, 0.2569, 0.2552, 0.2565, 0.2522, 0.2538, 0.2532, 0.2548, 0.2569, 0.2662, 0.2544, 0.2541, 0.2547, 0.2584, 0.2547, 0.2579, 0.2522, 0.2606, 0.2548, 0.2537, 0.2548, 0.2549, 0.2559, 0.2548, 0.2628, 0.2549, 0.2557, 0.2553, 0.2552, 0.2549, 0.2542, 0.255, 0.2602, 0.2598, 0.2557, 0.255, 0.2553, 0.2558, 0.2552, 0.2549, 0.256, 0.2567, 0.2565, 0.2565, 0.2551, 0.255, 0.2554, 0.2549, 0.2584, 0.2532, 0.2617, 0.2553, 0.2555, 0.2597, 0.256, 0.2556, 0.2556, 0.2566, 0.2563, 0.258, 0.2567, 0.2565, 0.2574, 0.2559, 0.2551, 0.2573, 0.2577, 0.2569, 0.2597, 0.2555, 0.2568, 0.2564]\n",
      "0.2516 105 0.166\n",
      "0.2564 0.1449\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/2L64f/slurm-801859.out\n",
      "454\n",
      "[1.2013, 1.1823, 0.7338, 0.4702, 0.4349, 0.4051, 0.3877, 0.3575, 0.3347, 0.323, 0.3209, 0.3042, 0.2942, 0.3353, 0.2993, 0.2837, 0.2818, 0.2762, 0.2749, 0.2799, 0.274, 0.2704, 0.2681, 0.2691, 0.2679, 0.2673, 0.264, 0.2664, 0.2686, 0.2644, 0.2605, 0.2692, 0.2711, 0.2722, 0.2621, 0.2635, 0.2629, 0.2644, 0.2653, 0.2693, 0.2619, 0.2592, 0.2573, 0.2642, 0.2596, 0.2587, 0.2595, 0.2614, 0.2584, 0.2603, 0.2574, 0.258, 0.2687, 0.2594, 0.2601, 0.2584, 0.2609, 0.2577, 0.2614, 0.2597, 0.258, 0.2601, 0.2594, 0.2615, 0.2606, 0.2625, 0.2577, 0.2567, 0.2556, 0.2615, 0.2588, 0.2584, 0.2581, 0.2674, 0.2564, 0.2583, 0.2591, 0.2572, 0.2579, 0.2617, 0.2604, 0.2607, 0.2604, 0.2601, 0.2584, 0.2583, 0.2575, 0.2611, 0.2578, 0.2611, 0.259, 0.2607, 0.2599, 0.2587, 0.2572, 0.2609, 0.2603, 0.2582, 0.2629, 0.2601, 0.2626, 0.26, 0.2593, 0.2603, 0.2692, 0.2627, 0.2649, 0.2581, 0.2608, 0.2673, 0.2635, 0.27, 0.2598, 0.2662, 0.2636, 0.2617, 0.2732, 0.2635, 0.2606, 0.2602, 0.2616, 0.2614, 0.2627, 0.2597, 0.2596, 0.2622, 0.262, 0.259, 0.2664, 0.2627, 0.262, 0.2649, 0.2607, 0.2635, 0.2593, 0.2619, 0.2627, 0.2624, 0.2615, 0.2674, 0.2599, 0.2619, 0.2611, 0.2619, 0.2632, 0.2593, 0.2582, 0.2621, 0.2633, 0.2612, 0.2659, 0.2659, 0.2601, 0.2638, 0.2629, 0.2647, 0.2614, 0.2608, 0.2617, 0.263, 0.2648, 0.2633, 0.2604, 0.261, 0.2661, 0.2604, 0.2619, 0.2625, 0.2668, 0.2622, 0.2648, 0.2724, 0.2617, 0.2656, 0.2622, 0.2652, 0.2623, 0.2644, 0.2669, 0.2617, 0.2629, 0.2643, 0.2612, 0.2629, 0.2663, 0.2619, 0.2634, 0.2662, 0.2626, 0.2619, 0.2627, 0.2637, 0.266, 0.2626, 0.2619, 0.262, 0.2641, 0.2634, 0.2646, 0.265]\n",
      "0.2556 69 0.1661\n",
      "0.265 0.1352\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/3L16f/slurm-802520.out\n",
      "464\n",
      "[1.1522, 0.5199, 0.472, 0.4528, 0.439, 0.4279, 0.4061, 0.4581, 0.3908, 0.3773, 0.3633, 0.352, 0.3434, 0.4331, 0.3487, 0.3343, 0.329, 0.3245, 0.3241, 0.3186, 0.3162, 0.3125, 0.3203, 0.3105, 0.3085, 0.3084, 0.3043, 0.3036, 0.3055, 0.3017, 0.3002, 0.299, 0.2978, 0.2989, 0.2981, 0.2951, 0.2961, 0.2942, 0.2931, 0.2928, 0.2932, 0.2967, 0.3065, 0.292, 0.2907, 0.2965, 0.2888, 0.2928, 0.2914, 0.2944, 0.2884, 0.2909, 0.2925, 0.2875, 0.3072, 0.285, 0.2856, 0.2934, 0.2873, 0.284, 0.301, 0.2838, 0.2823, 0.2861, 0.2827, 0.2812, 0.282, 0.2869, 0.2824, 0.293, 0.2796, 0.282, 0.2856, 0.2799, 0.2843, 0.2799, 0.2787, 0.2827, 0.2788, 0.2832, 0.2787, 0.2785, 0.2797, 0.2778, 0.2777, 0.2778, 0.281, 0.2775, 0.2775, 0.2802, 0.2778, 0.2784, 0.2784, 0.2782, 0.2802, 0.2763, 0.277, 0.2768, 0.2782, 0.2783, 0.2766, 0.2827, 0.2785, 0.2773, 0.2771, 0.2784, 0.2773, 0.2768, 0.2769, 0.2761, 0.2751, 0.2762, 0.281, 0.2807, 0.2782, 0.2773, 0.2806, 0.2762, 0.2793, 0.2804, 0.2774, 0.2796, 0.2765, 0.2775, 0.2759, 0.2759, 0.2777, 0.2786, 0.2772, 0.2739, 0.2774, 0.2768, 0.2775, 0.2769, 0.2775, 0.2746, 0.2795, 0.2753, 0.2776, 0.2761, 0.2756, 0.2762, 0.2757, 0.2748, 0.2853, 0.2783, 0.2777, 0.276, 0.2764, 0.2784, 0.277, 0.2772, 0.2761, 0.2757, 0.2765, 0.2781, 0.2759, 0.28, 0.2747, 0.2768, 0.2781, 0.276, 0.2797, 0.276, 0.2759, 0.2756, 0.2779, 0.2767, 0.2767, 0.2767, 0.2792, 0.2755, 0.2767, 0.2793, 0.2748, 0.2775, 0.2757, 0.2787, 0.2765, 0.278, 0.275, 0.2771, 0.2773, 0.2797, 0.2755, 0.279, 0.276, 0.2766, 0.2757, 0.2815, 0.2771, 0.2757, 0.2782, 0.2757, 0.2751, 0.2765, 0.2777, 0.2772, 0.2758, 0.2765]\n",
      "0.2739 130 0.1893\n",
      "0.2765 0.1692\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/3L32f/slurm-802546.out\n",
      "464\n",
      "[1.1583, 1.1822, 0.5799, 0.4709, 0.4189, 0.3985, 0.3858, 0.3416, 0.3303, 0.3199, 0.3048, 0.3016, 0.3097, 0.2947, 0.2966, 0.2909, 0.287, 0.2835, 0.2806, 0.2785, 0.278, 0.278, 0.332, 0.2759, 0.2739, 0.2712, 0.2734, 0.2695, 0.2697, 0.2673, 0.27, 0.2683, 0.2676, 0.2696, 0.2665, 0.266, 0.2648, 0.2732, 0.2654, 0.2714, 0.2725, 0.263, 0.2711, 0.263, 0.2636, 0.2665, 0.2632, 0.2637, 0.2705, 0.2654, 0.2721, 0.2663, 0.2641, 0.2619, 0.2629, 0.264, 0.263, 0.2624, 0.2618, 0.2614, 0.2673, 0.2664, 0.2637, 0.2749, 0.2665, 0.2603, 0.2647, 0.2723, 0.2693, 0.2614, 0.2657, 0.2634, 0.2623, 0.2632, 0.2624, 0.2621, 0.2621, 0.2622, 0.2614, 0.2666, 0.2608, 0.2598, 0.2604, 0.2604, 0.2648, 0.2616, 0.2615, 0.2666, 0.2609, 0.263, 0.262, 0.2594, 0.2643, 0.2611, 0.2603, 0.2613, 0.2619, 0.2604, 0.2634, 0.2611, 0.2652, 0.2599, 0.2626, 0.2598, 0.2618, 0.2662, 0.2606, 0.2606, 0.2609, 0.26, 0.2635, 0.2672, 0.2602, 0.2684, 0.2632, 0.2605, 0.2607, 0.2623, 0.261, 0.2621, 0.2632, 0.2596, 0.2612, 0.2618, 0.2614, 0.2642, 0.26, 0.26, 0.2646, 0.269, 0.2671, 0.2664, 0.2641, 0.261, 0.2611, 0.2607, 0.2682, 0.262, 0.2661, 0.2618, 0.2612, 0.2606, 0.2659, 0.2613, 0.2646, 0.2653, 0.262, 0.2637, 0.2621, 0.2598, 0.2624, 0.2634, 0.2614, 0.2614, 0.2607, 0.2634, 0.2643, 0.2611, 0.2629, 0.2599, 0.2625, 0.2651, 0.2688, 0.2632, 0.2658, 0.2618, 0.2663, 0.2645, 0.2632, 0.2638, 0.2645, 0.2602, 0.2592, 0.2595, 0.2625, 0.2615, 0.263, 0.2628, 0.2626, 0.2616, 0.2621, 0.2666, 0.2619, 0.2626, 0.2661, 0.2641, 0.2658, 0.264, 0.2617, 0.2656, 0.2618, 0.2607, 0.2671, 0.2613, 0.263, 0.2637, 0.2671, 0.2618, 0.2614, 0.2624]\n",
      "0.2592 173 0.1503\n",
      "0.2624 0.1461\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/3L64f/slurm-807447.out\n",
      "464\n",
      "[1.1937, 1.0063, 0.5044, 0.4516, 0.4194, 0.3997, 0.3704, 0.3548, 0.3354, 0.3313, 0.3229, 0.3163, 0.3226, 0.3068, 0.302, 0.3063, 0.2979, 0.3115, 0.2985, 0.2971, 0.2958, 0.2942, 0.2927, 0.2957, 0.2901, 0.2916, 0.2877, 0.2881, 0.288, 0.2889, 0.2859, 0.2868, 0.2923, 0.2847, 0.2855, 0.2925, 0.2845, 0.286, 0.2838, 0.2852, 0.2854, 0.2842, 0.2834, 0.2867, 0.2875, 0.2866, 0.2867, 0.2836, 0.2846, 0.2844, 0.2861, 0.2847, 0.2856, 0.287, 0.286, 0.2915, 0.287, 0.285, 0.2928, 0.2846, 0.291, 0.29, 0.288, 0.2895, 0.285, 0.2851, 0.2915, 0.2858, 0.2875, 0.2866, 0.2869, 0.2863, 0.2858, 0.2911, 0.2868, 0.2933, 0.287, 0.2865, 0.2871, 0.2865, 0.3013, 0.2903, 0.2888, 0.2873, 0.2908, 0.2896, 0.2877, 0.2869, 0.2994, 0.2918, 0.2866, 0.2876, 0.2894, 0.2895, 0.2894, 0.2882, 0.2966, 0.2872, 0.2895, 0.2894, 0.2898, 0.2897, 0.2909, 0.2913, 0.292, 0.29, 0.2899, 0.2887, 0.2889, 0.2903, 0.2917, 0.2924, 0.2919, 0.2903, 0.2912, 0.3061, 0.2916, 0.2908, 0.2957, 0.2913, 0.2908, 0.2927, 0.2928, 0.2983, 0.2968, 0.2955, 0.2911, 0.2917, 0.2959, 0.2946, 0.2936, 0.2943, 0.2908, 0.2924, 0.2901, 0.2932, 0.2922, 0.2914, 0.2946, 0.2926, 0.2915, 0.3027, 0.292, 0.2942, 0.2923, 0.2926, 0.2979, 0.2956, 0.2958, 0.2927, 0.2936, 0.293, 0.2933, 0.2921, 0.2949, 0.2923, 0.2936, 0.2929, 0.2929, 0.293, 0.2978, 0.2937, 0.297, 0.2928, 0.2934, 0.2948, 0.2929, 0.2996, 0.2932, 0.2925, 0.2939, 0.2972, 0.2967, 0.2963, 0.2933, 0.2955, 0.295, 0.2934, 0.2958, 0.2957, 0.2941, 0.2946, 0.2937, 0.2985, 0.3009, 0.2987, 0.297, 0.2939, 0.3032, 0.2964, 0.2989, 0.2966, 0.2946, 0.2983, 0.296, 0.295, 0.3001, 0.2979, 0.2974, 0.2953]\n",
      "0.2834 43 0.1844\n",
      "0.2953 0.1273\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for Ulevels,features in zip([2,2,2,3,3,3],[16,32,64,16,32,64]):\n",
    "    DA_training_paras={'nens':1280,\n",
    "                    'DA_method':'EnKF',\n",
    "                    'Nx_DA':32,\n",
    "                    'Nx_truth':128,\n",
    "                    'obs_freq':10,\n",
    "                    'obs_err':[1,-5,5,-7],\n",
    "                    'DA_freq':10,\n",
    "                    'save_B':False,\n",
    "                    'nobs':[50,50],\n",
    "                    'R_W':100,\n",
    "                    'inflate':[1,0.45]}\n",
    "    DA_training=DA.DA_exp(**DA_training_paras)\n",
    "    ml_dir='/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/'+DA_training.file_name()\n",
    "    stdout_file=glob('{}/{}L{}f/slurm-??????.out'.format(ml_dir,Ulevels,features))[0]\n",
    "    print(stdout_file)\n",
    "    with open(stdout_file) as f:\n",
    "        for count, line in enumerate(f):\n",
    "            pass\n",
    "    print(count)\n",
    "    with open(stdout_file) as f:\n",
    "        train_loss=[float(line.split()[5][7:-1]) for line in f.readlines()[count-399::2]]\n",
    "    with open(stdout_file) as f:\n",
    "        valid_loss=[float(line.split()[2][7:-1]) for line in f.readlines()[count-398::2]]\n",
    "\n",
    "    # with open('/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens80_freq10_relax0.72_R100_nobs100_20_err1E-6_5E-8/slurm-219462.out') as f:\n",
    "    #     train_loss=train_loss+[float(line.split()[5][7:-1]) for line in f.readlines()[70:120:2]]\n",
    "    # with open('/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens80_freq10_relax0.72_R100_nobs100_20_err1E-6_5E-8/slurm-219462.out') as f:\n",
    "    #     valid_loss=valid_loss+[float(line.split()[2][7:-1]) for line in f.readlines()[71:121:2]]\n",
    "\n",
    "    print(valid_loss)\n",
    "    fig=plt.figure(figsize=(6,4))\n",
    "    ax=plt.axes()\n",
    "    ax.plot(range(1,len(train_loss)+1),train_loss,linewidth=3,label='Training Loss')\n",
    "    ax.plot(range(1,len(train_loss)+1),valid_loss,linewidth=3,label=\"Validation Loss\")\n",
    "    min_valid=min(valid_loss)\n",
    "    max_valid=max(valid_loss)\n",
    "    min_index=valid_loss.index(min_valid)\n",
    "    print(min_valid,min_index+1,train_loss[min_index])\n",
    "    print(valid_loss[-1],train_loss[-1])\n",
    "    ax.text(20,max_valid-0.2*(max_valid-min_valid),'Min:{}\\nEpoch:{}'.format(min_valid,min_index+1),fontsize=12)\n",
    "    ax.set_xlabel('Epoch')\n",
    "    ax.grid()\n",
    "    ax.legend()\n",
    "    ax.set_title('U-Net Training and Validation Loss')\n",
    "    fig.tight_layout()\n",
    "    fig.savefig('./plots/loss/Nx{}_{}_{}L{}F.png'.format(DA_training.Nx_DA,DA_training.file_name_short(),Ulevels,features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/3L16f/slurm-802520.out\n",
      "464\n",
      "[1.1522, 0.5199, 0.472, 0.4528, 0.439, 0.4279, 0.4061, 0.4581, 0.3908, 0.3773, 0.3633, 0.352, 0.3434, 0.4331, 0.3487, 0.3343, 0.329, 0.3245, 0.3241, 0.3186, 0.3162, 0.3125, 0.3203, 0.3105, 0.3085, 0.3084, 0.3043, 0.3036, 0.3055, 0.3017, 0.3002, 0.299, 0.2978, 0.2989, 0.2981, 0.2951, 0.2961, 0.2942, 0.2931, 0.2928, 0.2932, 0.2967, 0.3065, 0.292, 0.2907, 0.2965, 0.2888, 0.2928, 0.2914, 0.2944, 0.2884, 0.2909, 0.2925, 0.2875, 0.3072, 0.285, 0.2856, 0.2934, 0.2873, 0.284, 0.301, 0.2838, 0.2823, 0.2861, 0.2827, 0.2812, 0.282, 0.2869, 0.2824, 0.293, 0.2796, 0.282, 0.2856, 0.2799, 0.2843, 0.2799, 0.2787, 0.2827, 0.2788, 0.2832, 0.2787, 0.2785, 0.2797, 0.2778, 0.2777, 0.2778, 0.281, 0.2775, 0.2775, 0.2802, 0.2778, 0.2784, 0.2784, 0.2782, 0.2802, 0.2763, 0.277, 0.2768, 0.2782, 0.2783, 0.2766, 0.2827, 0.2785, 0.2773, 0.2771, 0.2784, 0.2773, 0.2768, 0.2769, 0.2761, 0.2751, 0.2762, 0.281, 0.2807, 0.2782, 0.2773, 0.2806, 0.2762, 0.2793, 0.2804, 0.2774, 0.2796, 0.2765, 0.2775, 0.2759, 0.2759, 0.2777, 0.2786, 0.2772, 0.2739, 0.2774, 0.2768, 0.2775, 0.2769, 0.2775, 0.2746, 0.2795, 0.2753, 0.2776, 0.2761, 0.2756, 0.2762, 0.2757, 0.2748, 0.2853, 0.2783, 0.2777, 0.276, 0.2764, 0.2784, 0.277, 0.2772, 0.2761, 0.2757, 0.2765, 0.2781, 0.2759, 0.28, 0.2747, 0.2768, 0.2781, 0.276, 0.2797, 0.276, 0.2759, 0.2756, 0.2779, 0.2767, 0.2767, 0.2767, 0.2792, 0.2755, 0.2767, 0.2793, 0.2748, 0.2775, 0.2757, 0.2787, 0.2765, 0.278, 0.275, 0.2771, 0.2773, 0.2797, 0.2755, 0.279, 0.276, 0.2766, 0.2757, 0.2815, 0.2771, 0.2757, 0.2782, 0.2757, 0.2751, 0.2765, 0.2777, 0.2772, 0.2758, 0.2765]\n",
      "0.2739 130 0.1893\n",
      "0.2765 0.1692\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/3L32f/slurm-802546.out\n",
      "464\n",
      "[1.1583, 1.1822, 0.5799, 0.4709, 0.4189, 0.3985, 0.3858, 0.3416, 0.3303, 0.3199, 0.3048, 0.3016, 0.3097, 0.2947, 0.2966, 0.2909, 0.287, 0.2835, 0.2806, 0.2785, 0.278, 0.278, 0.332, 0.2759, 0.2739, 0.2712, 0.2734, 0.2695, 0.2697, 0.2673, 0.27, 0.2683, 0.2676, 0.2696, 0.2665, 0.266, 0.2648, 0.2732, 0.2654, 0.2714, 0.2725, 0.263, 0.2711, 0.263, 0.2636, 0.2665, 0.2632, 0.2637, 0.2705, 0.2654, 0.2721, 0.2663, 0.2641, 0.2619, 0.2629, 0.264, 0.263, 0.2624, 0.2618, 0.2614, 0.2673, 0.2664, 0.2637, 0.2749, 0.2665, 0.2603, 0.2647, 0.2723, 0.2693, 0.2614, 0.2657, 0.2634, 0.2623, 0.2632, 0.2624, 0.2621, 0.2621, 0.2622, 0.2614, 0.2666, 0.2608, 0.2598, 0.2604, 0.2604, 0.2648, 0.2616, 0.2615, 0.2666, 0.2609, 0.263, 0.262, 0.2594, 0.2643, 0.2611, 0.2603, 0.2613, 0.2619, 0.2604, 0.2634, 0.2611, 0.2652, 0.2599, 0.2626, 0.2598, 0.2618, 0.2662, 0.2606, 0.2606, 0.2609, 0.26, 0.2635, 0.2672, 0.2602, 0.2684, 0.2632, 0.2605, 0.2607, 0.2623, 0.261, 0.2621, 0.2632, 0.2596, 0.2612, 0.2618, 0.2614, 0.2642, 0.26, 0.26, 0.2646, 0.269, 0.2671, 0.2664, 0.2641, 0.261, 0.2611, 0.2607, 0.2682, 0.262, 0.2661, 0.2618, 0.2612, 0.2606, 0.2659, 0.2613, 0.2646, 0.2653, 0.262, 0.2637, 0.2621, 0.2598, 0.2624, 0.2634, 0.2614, 0.2614, 0.2607, 0.2634, 0.2643, 0.2611, 0.2629, 0.2599, 0.2625, 0.2651, 0.2688, 0.2632, 0.2658, 0.2618, 0.2663, 0.2645, 0.2632, 0.2638, 0.2645, 0.2602, 0.2592, 0.2595, 0.2625, 0.2615, 0.263, 0.2628, 0.2626, 0.2616, 0.2621, 0.2666, 0.2619, 0.2626, 0.2661, 0.2641, 0.2658, 0.264, 0.2617, 0.2656, 0.2618, 0.2607, 0.2671, 0.2613, 0.263, 0.2637, 0.2671, 0.2618, 0.2614, 0.2624]\n",
      "0.2592 173 0.1503\n",
      "0.2624 0.1461\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(4,3))\n",
    "ax=plt.axes()\n",
    "colors=sns.color_palette()\n",
    "for i,features in enumerate([16,32]):\n",
    "    Ulevels=3\n",
    "    DA_training_paras={'nens':1280,\n",
    "                    'DA_method':'EnKF',\n",
    "                    'Nx_DA':32,\n",
    "                    'Nx_truth':128,\n",
    "                    'obs_freq':10,\n",
    "                    'obs_err':[1,-5,5,-7],\n",
    "                    'DA_freq':10,\n",
    "                    'save_B':False,\n",
    "                    'nobs':[50,50],\n",
    "                    'R_W':100,\n",
    "                    'inflate':[1,0.45]}\n",
    "    DA_training=DA.DA_exp(**DA_training_paras)\n",
    "    ml_dir='/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/'+DA_training.file_name()\n",
    "    stdout_file=glob('{}/{}L{}f/slurm-??????.out'.format(ml_dir,Ulevels,features))[0]\n",
    "    print(stdout_file)\n",
    "    with open(stdout_file) as f:\n",
    "        for count, line in enumerate(f):\n",
    "            pass\n",
    "    print(count)\n",
    "    with open(stdout_file) as f:\n",
    "        train_loss=[float(line.split()[5][7:-1]) for line in f.readlines()[count-399::2]]\n",
    "    with open(stdout_file) as f:\n",
    "        valid_loss=[float(line.split()[2][7:-1]) for line in f.readlines()[count-398::2]]\n",
    "\n",
    "    print(valid_loss)\n",
    "    \n",
    "    ax.plot(range(1,len(train_loss)+1),train_loss,color=colors[i],linestyle='--',linewidth=3,\n",
    "            label='{}L{}f'.format(Ulevels,features)+' Training')\n",
    "    ax.plot(range(1,len(train_loss)+1),valid_loss,color=colors[i],linestyle='-',linewidth=3,\n",
    "            label='{}L{}f'.format(Ulevels,features)+' Validation')\n",
    "    min_valid=min(valid_loss)\n",
    "    max_valid=max(valid_loss)\n",
    "    min_index=valid_loss.index(min_valid)\n",
    "    print(min_valid,min_index+1,train_loss[min_index])\n",
    "    print(valid_loss[-1],train_loss[-1])\n",
    "    # ax.text(20,max_valid-0.2*(max_valid-min_valid),'Min:{}\\nEpoch:{}'.format(min_valid,min_index+1),fontsize=12)\n",
    "\n",
    "ax.set_xlim([0,None])\n",
    "ax.set_xlabel('Epoch')\n",
    "# ax.set_ylim([0.17,0.4])\n",
    "ax.grid()\n",
    "ax.legend()\n",
    "ax.set_title('U-Net Training and Validation Loss')\n",
    "fig.tight_layout()\n",
    "fig.savefig('./plots/loss/Nx{}_{}_{}L{}F_200.png'.format(DA_training.Nx_DA,DA_training.file_name_short(),Ulevels,features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens1280_freq10_relax0.45_R100_nobs50_50_err1E-5_5E-7/2L32f/slurm-801858.out\n",
      "454\n",
      "[1.1677, 0.5673, 0.4471, 0.4207, 0.3948, 0.364, 0.3471, 0.3311, 0.3168, 0.3227, 0.3042, 0.2957, 0.294, 0.2891, 0.2857, 0.2853, 0.3266, 0.2831, 0.2829, 0.2905, 0.2864, 0.2751, 0.2741, 0.2757, 0.2709, 0.2717, 0.2753, 0.2688, 0.2689, 0.2657, 0.2682, 0.2674, 0.2668, 0.2639, 0.2739, 0.265, 0.2656, 0.2655, 0.2612, 0.268, 0.2703, 0.26, 0.2688, 0.265, 0.2593, 0.259, 0.2602, 0.2601, 0.2602, 0.2737, 0.2601, 0.2561, 0.2566, 0.2561, 0.2605, 0.259, 0.2614, 0.2565, 0.2559, 0.2629, 0.2575, 0.2586, 0.2567, 0.2548, 0.256, 0.2536, 0.2549, 0.2627, 0.257, 0.2576, 0.2547, 0.2549, 0.2594, 0.2576, 0.2546, 0.2556, 0.2543, 0.2553, 0.2615, 0.2547, 0.2544, 0.2538, 0.2566, 0.2575, 0.2524, 0.2596, 0.2523, 0.2533, 0.2554, 0.255, 0.2586, 0.2535, 0.253, 0.258, 0.2529, 0.2553, 0.2549, 0.2657, 0.2548, 0.2677, 0.2531, 0.2526, 0.2545, 0.2545, 0.2516, 0.2537, 0.2531, 0.2536, 0.2521, 0.2532, 0.2535, 0.2578, 0.2532, 0.2579, 0.2553, 0.2558, 0.2535, 0.2545, 0.2591, 0.2543, 0.2534, 0.2534, 0.256, 0.254, 0.2543, 0.2532, 0.2588, 0.2527, 0.2541, 0.2569, 0.2552, 0.2565, 0.2522, 0.2538, 0.2532, 0.2548, 0.2569, 0.2662, 0.2544, 0.2541, 0.2547, 0.2584, 0.2547, 0.2579, 0.2522, 0.2606, 0.2548, 0.2537, 0.2548, 0.2549, 0.2559, 0.2548, 0.2628, 0.2549, 0.2557, 0.2553, 0.2552, 0.2549, 0.2542, 0.255, 0.2602, 0.2598, 0.2557, 0.255, 0.2553, 0.2558, 0.2552, 0.2549, 0.256, 0.2567, 0.2565, 0.2565, 0.2551, 0.255, 0.2554, 0.2549, 0.2584, 0.2532, 0.2617, 0.2553, 0.2555, 0.2597, 0.256, 0.2556, 0.2556, 0.2566, 0.2563, 0.258, 0.2567, 0.2565, 0.2574, 0.2559, 0.2551, 0.2573, 0.2577, 0.2569, 0.2597, 0.2555, 0.2568, 0.2564]\n",
      "0.2516 105 0.166\n",
      "0.2564 0.1449\n",
      "/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/EnKF_Nx32_from_Nx128_ens80_freq10_relax0.5_R100_nobs50_50_err1E-5_5E-7/2L32f/slurm-802547.out\n",
      "454\n",
      "[0.9042, 0.728, 0.704, 0.6856, 0.689, 0.6734, 0.6652, 0.6612, 0.6665, 0.6563, 0.6525, 0.6527, 0.65, 0.65, 0.6498, 0.6464, 0.6473, 0.6496, 0.6455, 0.6466, 0.6478, 0.644, 0.6461, 0.6454, 0.6434, 0.6437, 0.6462, 0.645, 0.6446, 0.645, 0.6452, 0.645, 0.6492, 0.6447, 0.6466, 0.6461, 0.645, 0.645, 0.6448, 0.6545, 0.6445, 0.644, 0.6459, 0.6485, 0.6464, 0.6465, 0.6467, 0.6475, 0.6487, 0.6454, 0.6484, 0.6492, 0.6477, 0.6488, 0.6469, 0.6503, 0.6531, 0.6485, 0.65, 0.6505, 0.6488, 0.655, 0.6508, 0.6513, 0.6535, 0.6541, 0.6549, 0.6541, 0.6582, 0.6554, 0.6544, 0.6575, 0.6572, 0.6586, 0.6542, 0.6538, 0.6604, 0.6595, 0.6619, 0.6574, 0.6571, 0.6626, 0.6561, 0.6558, 0.6589, 0.6582, 0.6574, 0.658, 0.6584, 0.6642, 0.6626, 0.6643, 0.6616, 0.6621, 0.661, 0.666, 0.6618, 0.6631, 0.66, 0.6627, 0.6652, 0.6644, 0.6594, 0.6621, 0.6654, 0.6687, 0.6624, 0.6641, 0.6684, 0.6626, 0.6655, 0.668, 0.6661, 0.6663, 0.6737, 0.6679, 0.6732, 0.6716, 0.6674, 0.6765, 0.6672, 0.6656, 0.6711, 0.6687, 0.6691, 0.6655, 0.6691, 0.6686, 0.6685, 0.6688, 0.6747, 0.6693, 0.6688, 0.6693, 0.6739, 0.6706, 0.6692, 0.6696, 0.6734, 0.674, 0.6723, 0.6729, 0.6697, 0.6712, 0.6751, 0.6739, 0.6686, 0.6786, 0.6712, 0.6733, 0.6694, 0.6772, 0.6784, 0.6739, 0.6754, 0.6753, 0.6741, 0.6739, 0.6736, 0.6741, 0.68, 0.6844, 0.6722, 0.6762, 0.68, 0.6732, 0.6715, 0.6706, 0.6758, 0.6733, 0.6776, 0.6784, 0.6769, 0.6788, 0.6748, 0.6786, 0.6748, 0.6766, 0.6749, 0.6753, 0.6788, 0.6756, 0.6852, 0.6778, 0.6765, 0.6772, 0.674, 0.6816, 0.6798, 0.6874, 0.6783, 0.6802, 0.6806, 0.6819, 0.6811, 0.6775, 0.6785, 0.6844, 0.6753, 0.6775]\n",
      "0.6434 25 0.5826\n",
      "0.6775 0.4984\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(6,4))\n",
    "ax=plt.axes()\n",
    "colors=sns.color_palette()\n",
    "obs_kwargs={'Nx_truth':128,'obs_freq':10,'obs_err':[1,-5,5,-7],'DA_freq':10,'save_B':False,'nobs':[50,50]}\n",
    "DA_training_paras1={'nens':1280,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.45],**obs_kwargs}\n",
    "DA_training_paras2={'nens':80,'DA_method':'EnKF','Nx_DA':32,'R_W':100,'inflate':[1,0.5],**obs_kwargs}\n",
    "\n",
    "for i,(features,DA_paras,exp_str) in enumerate(zip([32,32],\n",
    "                                                   [DA_training_paras1,DA_training_paras2],\n",
    "                                                   ['UNet_L1','UNet_L2'])):\n",
    "    Ulevels=2\n",
    "    DA_training=DA.DA_exp(**DA_paras)\n",
    "    ml_dir='/home/feiyul/PyQG/pyqg_experiments/examples/DA/ML/'+DA_training.file_name()\n",
    "    stdout_file=glob('{}/{}L{}f/slurm-??????.out'.format(ml_dir,Ulevels,features))[0]\n",
    "    print(stdout_file)\n",
    "    with open(stdout_file) as f:\n",
    "        for count, line in enumerate(f):\n",
    "            pass\n",
    "    print(count)\n",
    "    with open(stdout_file) as f:\n",
    "        train_loss=[float(line.split()[5][7:-1]) for line in f.readlines()[count-399::2]]\n",
    "    with open(stdout_file) as f:\n",
    "        valid_loss=[float(line.split()[2][7:-1]) for line in f.readlines()[count-398::2]]\n",
    "\n",
    "    print(valid_loss)\n",
    "    \n",
    "    ax.plot(range(1,len(train_loss)+1),train_loss,color=colors[i],linestyle='--',linewidth=3,\n",
    "            label=exp_str+' Training')\n",
    "    ax.plot(range(1,len(train_loss)+1),valid_loss,color=colors[i],linestyle='-',linewidth=3,\n",
    "            label=exp_str+' Validation')\n",
    "    min_valid=min(valid_loss)\n",
    "    max_valid=max(valid_loss)\n",
    "    min_index=valid_loss.index(min_valid)\n",
    "    print(min_valid,min_index+1,train_loss[min_index])\n",
    "    print(valid_loss[-1],train_loss[-1])\n",
    "    # ax.text(20,max_valid-0.2*(max_valid-min_valid),'Min:{}\\nEpoch:{}'.format(min_valid,min_index+1),fontsize=12)\n",
    "\n",
    "# ax.set_xlim([20,None])\n",
    "ax.set_xlabel('Epoch')\n",
    "# ax.set_ylim([0.17,0.4])\n",
    "ax.grid()\n",
    "ax.legend()\n",
    "ax.set_title('U-Net Training and Validation Loss')\n",
    "fig.tight_layout()\n",
    "fig.savefig('./plots/loss/Nx{}_{}_{}L{}F_compare.png'.format(DA_training.Nx_DA,DA_training.file_name_short(),Ulevels,features))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pyqg",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  },
  "vscode": {
   "interpreter": {
    "hash": "6973bb065886c44b729be95666fc7914cdffcda5c95a8d198f96362135dd92cd"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}