test/test_KMT.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os, sys, re\n",
    "sys.path.append(os.path.dirname(os.getcwd()))\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm.notebook import tqdm\n",
    "import torch\n",
    "import torchcde\n",
    "import h5py\n",
    "from scipy.signal import find_peaks\n",
    "import MulensModel as mm\n",
    "import corner\n",
    "from model.utils import getfsfb\n",
    "from matplotlib.offsetbox import AnchoredText\n",
    "\n",
    "from scipy.signal import medfilt\n",
    "from PyAstronomy.pyasl import binningx0dt\n",
    "\n",
    "from model.locator import Locator\n",
    "from model.cde_mdn import CDE_MDN\n",
    "\n",
    "torch.random.manual_seed(42)\n",
    "np.random.seed(42)\n",
    "plt.rcParams[\"font.family\"] = \"serif\"\n",
    "plt.rcParams['mathtext.fontset'] = 'cm'\n",
    "plt.rcParams['mathtext.rm'] = 'serif'\n",
    "plt.rc('font', size=14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "gap = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "if gap:\n",
    "    tel = pd.read_csv('./KMT/kb190371_aligned.csv', skiprows=0, usecols=[0, 2, 5, 6])\n",
    "    tel = tel[(tel['Tel'] != 'KMTS01') & (tel['Tel'] != 'KMTS41')]\n",
    "    x_orig = tel[['HJD', 'mag_aligned', 'e_mag_aligned']].to_numpy()\n",
    "else:\n",
    "    x_orig = np.loadtxt('./KMT/kb190371_aligned.csv', delimiter=',', skiprows=1, usecols=(2, 5, 6))\n",
    "order = np.argsort(x_orig[:, 0])\n",
    "x_orig = x_orig[order]\n",
    "err = x_orig[:, -1]\n",
    "x_orig = x_orig[:, :-1]\n",
    "x = x_orig\n",
    "err = err[x[:, 0] > 8550]\n",
    "x = x[x[:, 0] > 8550]\n",
    "if gap:\n",
    "    err = err[x[:, 0] < 8650]\n",
    "    x = x[x[:, 0] < 8650]\n",
    "else:\n",
    "    err = err[x[:, 0] < 8640]\n",
    "    x = x[x[:, 0] < 8640]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3770\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": [
    "plt.scatter(x[:, 0], x[:, 1])\n",
    "print(len(x))\n",
    "plt.show()\n",
    "# plt.xlim(8590, 8595)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "436\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": [
    "# x[:, 1] = medfilt(x[:, 1], 7)\n",
    "threshold = 0.1\n",
    "ind = err < threshold\n",
    "x = x[ind]\n",
    "err = err[ind]\n",
    "x[:, 1] -= 5\n",
    "x, dt = binningx0dt(x[:, 0], x[:, 1], reduceBy=5, yvalFunc=np.median)\n",
    "print(len(x))\n",
    "plt.scatter(x[:, 0], x[:, 1])\n",
    "plt.show()\n",
    "x = x[:, :2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_coeffs = torchcde.hermite_cubic_coefficients_with_backward_differences(torch.tensor(x).unsqueeze(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Locator(\n",
       "  (prefilter): Sequential(\n",
       "    (0): Conv1d(1, 16, kernel_size=(3,), stride=(1,), padding=(1,), padding_mode=reflect)\n",
       "    (1): PReLU(num_parameters=1)\n",
       "    (2): Conv1d(16, 16, kernel_size=(15,), stride=(1,), padding=(7,), padding_mode=reflect)\n",
       "    (3): PReLU(num_parameters=1)\n",
       "    (4): Conv1d(16, 1, kernel_size=(1,), stride=(1,), padding_mode=reflect)\n",
       "    (5): PReLU(num_parameters=1)\n",
       "  )\n",
       "  (unet): UNET_1D(\n",
       "    (AvgPool1D1): AvgPool1d(kernel_size=(1,), stride=(5,), padding=(0,))\n",
       "    (AvgPool1D2): AvgPool1d(kernel_size=(1,), stride=(25,), padding=(0,))\n",
       "    (AvgPool1D3): AvgPool1d(kernel_size=(1,), stride=(125,), padding=(0,))\n",
       "    (layer1): Sequential(\n",
       "      (0): conbr_block(\n",
       "        (conv1): Conv1d(1, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "        (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "        (relu): PReLU(num_parameters=1)\n",
       "      )\n",
       "      (1): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(128, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(128, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(128, 16, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(16, 128, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=16, bias=True)\n",
       "          (fc2): Linear(in_features=16, out_features=128, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "      (2): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(128, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(128, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(128, 16, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(16, 128, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=16, bias=True)\n",
       "          (fc2): Linear(in_features=16, out_features=128, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (layer2): Sequential(\n",
       "      (0): conbr_block(\n",
       "        (conv1): Conv1d(128, 256, kernel_size=(7,), stride=(5,), padding=(3,))\n",
       "        (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "        (relu): PReLU(num_parameters=1)\n",
       "      )\n",
       "      (1): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(256, 256, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(256, 256, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(256, 32, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(32, 256, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=32, bias=True)\n",
       "          (fc2): Linear(in_features=32, out_features=256, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "      (2): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(256, 256, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(256, 256, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(256, 32, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(32, 256, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=32, bias=True)\n",
       "          (fc2): Linear(in_features=32, out_features=256, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (layer3): Sequential(\n",
       "      (0): conbr_block(\n",
       "        (conv1): Conv1d(257, 384, kernel_size=(7,), stride=(5,), padding=(3,))\n",
       "        (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "        (relu): PReLU(num_parameters=1)\n",
       "      )\n",
       "      (1): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(384, 384, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(384, 384, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(384, 48, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(48, 384, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=48, bias=True)\n",
       "          (fc2): Linear(in_features=48, out_features=384, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "      (2): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(384, 384, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(384, 384, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(384, 48, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(48, 384, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=48, bias=True)\n",
       "          (fc2): Linear(in_features=48, out_features=384, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (layer4): Sequential(\n",
       "      (0): conbr_block(\n",
       "        (conv1): Conv1d(385, 512, kernel_size=(7,), stride=(5,), padding=(3,))\n",
       "        (bn): InstanceNorm1d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "        (relu): PReLU(num_parameters=1)\n",
       "      )\n",
       "      (1): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(512, 512, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(512, 512, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(512, 64, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(64, 512, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=64, bias=True)\n",
       "          (fc2): Linear(in_features=64, out_features=512, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "      (2): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(512, 512, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(512, 512, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(512, 64, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(64, 512, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=64, bias=True)\n",
       "          (fc2): Linear(in_features=64, out_features=512, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (layer5): Sequential(\n",
       "      (0): conbr_block(\n",
       "        (conv1): Conv1d(513, 640, kernel_size=(7,), stride=(4,), padding=(3,))\n",
       "        (bn): InstanceNorm1d(640, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "        (relu): PReLU(num_parameters=1)\n",
       "      )\n",
       "      (1): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(640, 640, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(640, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(640, 640, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(640, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(640, 80, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(80, 640, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=80, bias=True)\n",
       "          (fc2): Linear(in_features=80, out_features=640, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "      (2): re_block(\n",
       "        (cbr1): conbr_block(\n",
       "          (conv1): Conv1d(640, 640, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(640, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (cbr2): conbr_block(\n",
       "          (conv1): Conv1d(640, 640, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "          (bn): InstanceNorm1d(640, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "        )\n",
       "        (seblock): se_block(\n",
       "          (conv1): Conv1d(640, 80, kernel_size=(1,), stride=(1,))\n",
       "          (conv2): Conv1d(80, 640, kernel_size=(1,), stride=(1,))\n",
       "          (fc): Linear(in_features=1, out_features=80, bias=True)\n",
       "          (fc2): Linear(in_features=80, out_features=640, bias=True)\n",
       "          (relu): PReLU(num_parameters=1)\n",
       "          (sigmoid): Sigmoid()\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (cbr_up1): conbr_block(\n",
       "      (conv1): Conv1d(896, 384, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "      (bn): InstanceNorm1d(384, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "      (relu): PReLU(num_parameters=1)\n",
       "    )\n",
       "    (cbr_up2): conbr_block(\n",
       "      (conv1): Conv1d(640, 256, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "      (bn): InstanceNorm1d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "      (relu): PReLU(num_parameters=1)\n",
       "    )\n",
       "    (cbr_up3): conbr_block(\n",
       "      (conv1): Conv1d(384, 128, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "      (bn): InstanceNorm1d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)\n",
       "      (relu): PReLU(num_parameters=1)\n",
       "    )\n",
       "    (upsample): Upsample(scale_factor=5.0, mode=nearest)\n",
       "    (upsample1): Upsample(scale_factor=5.0, mode=nearest)\n",
       "    (outcov): Conv1d(128, 1, kernel_size=(7,), stride=(1,), padding=(3,))\n",
       "    (sig): Sigmoid()\n",
       "  )\n",
       "  (loss): DiceLoss()\n",
       ")"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "k = 2; method='avg'\n",
    "checkpt = torch.load('/work/hmzhao/experiments/locator/experiment_smallte.ckpt', map_location='cpu')\n",
    "# checkpt = torch.load('/work/hmzhao/experiments/locator/experiment_54125.ckpt', map_location='cpu')\n",
    "device = torch.device(\"cuda:2\" if torch.cuda.is_available() else \"cpu\")\n",
    "\n",
    "ckpt_args = checkpt['args']\n",
    "state_dict = checkpt['state_dict']\n",
    "\n",
    "model = Locator(device, k=k, method=method).to(device)\n",
    "model_dict = model.state_dict()\n",
    "\n",
    "# 1. filter out unnecessary keys\n",
    "state_dict = {k: v for k, v in state_dict.items() if k in model_dict}\n",
    "# 2. overwrite entries in the existing state dict\n",
    "model_dict.update(state_dict) \n",
    "# 3. load the new state dict\n",
    "model.load_state_dict(state_dict)\n",
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "with torch.no_grad():\n",
    "    model.eval()\n",
    "    model.animate = True\n",
    "    model.soft_threshold = False\n",
    "    res = model(train_coeffs.float().to(device), torch.zeros((1, 2)).to(device))\n",
    "    pred_t = res[0].detach().cpu().numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mask = res[2].detach().cpu().numpy()[0]\n",
    "plt.plot(mask[:, 0], mask[:, 1])\n",
    "plt.twinx()\n",
    "plt.scatter(x[:, 0], x[:, 1])\n",
    "# plt.xlim(8580, 8610)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[8592.390625, 6.6822509765625]]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_t.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "if gap:\n",
    "    tel = pd.read_csv('./KMT/kb190371_aligned.csv', skiprows=0, usecols=[0, 2, 5, 6])\n",
    "    tel = tel[(tel['Tel'] != 'KMTS01') & (tel['Tel'] != 'KMTS41')]\n",
    "    x_orig = tel[['HJD', 'mag_aligned', 'e_mag_aligned']].to_numpy()\n",
    "else:\n",
    "    x_orig = np.loadtxt('./KMT/kb190371_aligned.csv', delimiter=',', skiprows=1, usecols=(2, 5, 6))\n",
    "order = np.argsort(x_orig[:, 0])\n",
    "x_orig = x_orig[order]\n",
    "err = x_orig[:, -1]\n",
    "x_orig = x_orig[:, :-1]\n",
    "# t0 = 8592.392; tE = 6.55\n",
    "t0, tE = pred_t[0]\n",
    "x_orig[:, 0] = (x_orig[:, 0] - t0)/tE\n",
    "ind = (x_orig[:, 0] > -2) * (x_orig[:, 0] < 2)\n",
    "mbase = x_orig[~ind, 1].mean()\n",
    "x_orig = x_orig[ind]\n",
    "err = err[ind]\n",
    "# ind = (x_orig[:, 0] < -0.03) * (x_orig[:, 0] > -0.045)\n",
    "# x_orig = x_orig[~ind]\n",
    "# err = err[~ind]\n",
    "x = x_orig.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "if gap:\n",
    "    x[:, 1] = medfilt(x[:, 1], 7)\n",
    "else:\n",
    "    x[:, 1] = medfilt(x[:, 1], 5)\n",
    "# x[x[:, 0] < -0.1, 1] = medfilt(x[x[:, 0] < -0.1, 1], 5)\n",
    "# x[x[:, 0] > 0.1, 1] = medfilt(x[x[:, 0] > 0.1, 1], 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "threshold = 0.06\n",
    "plt.scatter(x[:, 0], x[:, 1])\n",
    "plt.scatter(x[err<threshold, 0], x[err<threshold, 1])\n",
    "plt.xlim(-0.3, 0.3)\n",
    "len(x[err<threshold, 0])\n",
    "ind = err < threshold\n",
    "x = x[ind]\n",
    "err = err[ind]\n",
    "x[:, 1] = (x[:, 1] - mbase) / 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = torch.tensor(x.reshape(1, *x.shape))\n",
    "# depth = 3; window_length = 4\n",
    "depth = 3; window_length = 8\n",
    "data = torchcde.logsig_windows(data, depth, window_length=window_length)\n",
    "train_coeffs = torchcde.hermite_cubic_coefficients_with_backward_differences(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CDE_MDN(\n",
       "  (cde_func): CDEFunc(\n",
       "    (linear1): Linear(in_features=32, out_features=1024, bias=True)\n",
       "    (relu1): PReLU(num_parameters=1)\n",
       "    (resblocks): Sequential(\n",
       "      (0): ResBlock(\n",
       "        (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "        (nonlinear1): PReLU(num_parameters=1)\n",
       "        (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      )\n",
       "      (1): ResBlock(\n",
       "        (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "        (nonlinear1): PReLU(num_parameters=1)\n",
       "        (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      )\n",
       "      (2): ResBlock(\n",
       "        (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "        (nonlinear1): PReLU(num_parameters=1)\n",
       "        (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      )\n",
       "    )\n",
       "    (relu2): PReLU(num_parameters=1)\n",
       "    (linear2): Linear(in_features=1024, out_features=160, bias=True)\n",
       "    (tanh): Tanh()\n",
       "    (linear3): Linear(in_features=160, out_features=160, bias=True)\n",
       "  )\n",
       "  (initial): Sequential(\n",
       "    (0): Sequential(\n",
       "      (0): Linear(in_features=5, out_features=1024, bias=True)\n",
       "      (1): ReLU()\n",
       "      (2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (1): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (2): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (3): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (4): Sequential(\n",
       "      (0): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (1): ReLU()\n",
       "      (2): Linear(in_features=1024, out_features=32, bias=True)\n",
       "    )\n",
       "  )\n",
       "  (readout): Sequential(\n",
       "    (0): Sequential(\n",
       "      (0): Linear(in_features=32, out_features=1024, bias=True)\n",
       "      (1): ReLU()\n",
       "      (2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (1): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (2): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (3): ResBlock(\n",
       "      (linear1): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "      (nonlinear1): ReLU()\n",
       "      (linear2): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "    )\n",
       "    (4): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "  )\n",
       "  (mdn): MixtureDensityNetwork(\n",
       "    (pi_network): CategoricalNetwork(\n",
       "      (network): Sequential(\n",
       "        (0): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "        (1): ELU(alpha=1.0)\n",
       "        (2): Linear(in_features=1024, out_features=12, bias=True)\n",
       "      )\n",
       "    )\n",
       "    (normal_network): MixtureDiagNormalNetwork(\n",
       "      (network): Sequential(\n",
       "        (0): Linear(in_features=1024, out_features=1024, bias=True)\n",
       "        (1): ELU(alpha=1.0)\n",
       "        (2): Linear(in_features=1024, out_features=120, bias=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "checkpt = torch.load('/work/hmzhao/experiments/cde_mdn/experiment_l32nG12diag.ckpt', map_location='cpu')\n",
    "ckpt_args = checkpt['args']\n",
    "state_dict = checkpt['state_dict']\n",
    "\n",
    "output_dim = 5\n",
    "input_dim = data.shape[-1]\n",
    "latent_dim = ckpt_args.latents\n",
    "\n",
    "model = CDE_MDN(input_dim, latent_dim, output_dim).to(device)\n",
    "model_dict = model.state_dict()\n",
    "\n",
    "# 1. filter out unnecessary keys\n",
    "state_dict = {k: v for k, v in state_dict.items() if k in model_dict}\n",
    "# 2. overwrite entries in the existing state dict\n",
    "model_dict.update(state_dict) \n",
    "# 3. load the new state dict\n",
    "model.load_state_dict(state_dict)\n",
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "with torch.no_grad():\n",
    "    pi, normal = model(train_coeffs.float().to(device))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 849.6x849.6 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n = int(1e6)\n",
    "\n",
    "pi_ = pi.probs.cpu(); loc_ = normal.loc.cpu(); scale_ = normal.scale.cpu()\n",
    "pi_ = torch.tile(pi_, (n, 1)); loc_ = torch.tile(loc_, (n, 1, 1)); scale_ = torch.tile(scale_, (n, 1, 1))\n",
    "normal_dist = torch.distributions.Normal(loc_, scale_)\n",
    "pi_dist = torch.distributions.OneHotCategorical(probs=pi_)\n",
    "sample = model.sample(pi_dist, normal_dist).numpy()\n",
    "\n",
    "\n",
    "test_p = [[0.140, np.log10(6.46)-3, np.log10(0.079), np.log10(0.826), 2-1.542/np.pi, 0],\n",
    "        [0.141, np.log10(6.99)-3, np.log10(0.123), np.log10(1.575), 2-1.544/np.pi, 0]]\n",
    "fs, fb = getfsfb(x[:, 0], 10**(x[:, 1]/5/(-2.5)), err/(2.5*np.log(10))*10**(x[:, 1]/5/(-2.5)), 0, 1, *test_p[0][:-1])[1:3]\n",
    "test_p[0][-1] = np.log10(fs / (fs + fb))\n",
    "fs, fb = getfsfb(x[:, 0], 10**(x[:, 1]/5/(-2.5)), err/(2.5*np.log(10))*10**(x[:, 1]/5/(-2.5)), 0, 1, *test_p[1][:-1])[1:3]\n",
    "test_p[1][-1] = np.log10(fs / (fs + fb))\n",
    "truths_full = np.array(test_p)\n",
    "truths = np.delete(truths_full, 1, axis=-1)\n",
    "\n",
    "range_p = [(0, 1), (-4, 0), (-0.6, 0.6), (0, 2), (-1, 0)]\n",
    "# range_p = None\n",
    "sigma_level = 1-np.exp(-0.5)\n",
    "fig = corner.corner(sample, labels=[r\"$u_0$\", r\"$\\lg q$\", r\"$\\lg s$\", r\"$\\alpha/180$\", r\"$\\lg f_s$\"],\n",
    "            smooth=1,\n",
    "            bins=50,\n",
    "            range=range_p,\n",
    "            show_titles=True, title_kwargs={\"fontsize\": 12},\n",
    "            truths=truths[0], truth_color='C1', \n",
    "            fill_contours=False, color='blue', no_fill_contours=True,\n",
    "            plot_datapoints=False, plot_density=False,\n",
    "            )\n",
    "corner.overplot_lines(fig, truths[1], color=\"C2\")\n",
    "corner.overplot_points(fig, truths[1][None], marker=\"s\", color=\"C2\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.optimize import fmin\n",
    "import VBBinaryLensing\n",
    "\n",
    "def get_fsfb(amp, flux, ferr):\n",
    "    sig2 = ferr**2\n",
    "    wght = flux/sig2\n",
    "    d = np.ones(2)\n",
    "    d[0] = np.sum(wght*amp)\n",
    "    d[1] = np.sum(wght)\n",
    "    b = np.zeros((2,2))\n",
    "    b[0,0] = np.sum(amp**2/sig2)\n",
    "    b[0,1] = np.sum(amp/sig2)\n",
    "    b[1,0] = b[0,1]\n",
    "    b[1,1] = np.sum(1./sig2)\n",
    "    c = np.linalg.inv(b)\n",
    "    fs = np.sum(c[0]*d)\n",
    "    fb = np.sum(c[1]*d)\n",
    "    fserr = np.sqrt(c[0,0])\n",
    "    fberr = np.sqrt(c[1,1])\n",
    "    fmod = fs*amp+fb\n",
    "    chi2 = np.sum((flux-fmod)**2/sig2)\n",
    "    return chi2,fs,fb,fserr,fberr\n",
    "\n",
    "def compute_model_lc(time_array, fitting_parameters, VBBL, rho=None):\n",
    "    u0, lgq, lgs, ad180 = fitting_parameters[:4]\n",
    "    q, s = 10**lgq, 10**lgs\n",
    "    alpha = ad180 * np.pi # convert to radian\n",
    "    t0, te = 0, 1\n",
    "    if rho == None:\n",
    "        rho = 1e-3\n",
    "    if len(fitting_parameters) == 5:\n",
    "        rho = 10**fitting_parameters[-1]\n",
    "    if len(fitting_parameters) == 6:\n",
    "        t0 = fitting_parameters[-2]\n",
    "        te = fitting_parameters[-1]\n",
    "    if len(fitting_parameters) == 7:\n",
    "        rho = 10**fitting_parameters[-3]\n",
    "        t0 = fitting_parameters[-2]\n",
    "        te = fitting_parameters[-1]\n",
    "    tau = (time_array-t0)/te\n",
    "    xs = tau*np.cos(alpha) - u0*np.sin(alpha)\n",
    "    ys = tau*np.sin(alpha) + u0*np.cos(alpha)\n",
    "    magnifications = np.array([VBBL.BinaryMag2(s, q, xs[i], ys[i], rho) for i in range(len(xs))])\n",
    "    return magnifications\n",
    "\n",
    "def compute_chisq(fitting_parameters, time, flux, ferr, VBBL, return_model=False, rho=None, return_lc=False):\n",
    "        if len(fitting_parameters) == 5:\n",
    "            rho = 10**fitting_parameters[-1]\n",
    "            fitting_parameters = fitting_parameters[:-1]\n",
    "        magnifications = compute_model_lc(time, fitting_parameters, VBBL, rho)\n",
    "        chi2, fs, fb, fserr, fberr = get_fsfb(magnifications, flux, ferr)\n",
    "        if return_lc:\n",
    "            time_model = np.arange(time.min(), time.max(), 0.001)\n",
    "            magnifications = compute_model_lc(time_model, fitting_parameters, VBBL, rho)\n",
    "            mag_model = 18 - 2.5*np.log10(magnifications*fs + fb)\n",
    "            model = np.vstack((time_model, mag_model))\n",
    "            return chi2, fs, fb, model\n",
    "        if return_model:\n",
    "            return chi2, fs, fb\n",
    "        return chi2\n",
    "\n",
    "def perform_optimization(time, flux, ferr, para_initial, verbose=True):\n",
    "    VBBL = VBBinaryLensing.VBBinaryLensing()\n",
    "\n",
    "    para_best, chi2_min, iter, funcalls, warnflag, allevcs = fmin(compute_chisq, para_initial, args=(time, flux, ferr, VBBL), full_output=True, retall=True, maxiter=1000, maxfun=5000, disp=verbose)\n",
    "\n",
    "    chi2_min, fs, fb = compute_chisq(para_initial, time, flux, ferr, VBBL, return_model=True)\n",
    "    if verbose:\n",
    "        print('initial chisq: ', chi2_min)\n",
    "    chi2_min, fs, fb = compute_chisq(para_best, time, flux, ferr, VBBL, return_model=True)\n",
    "    if verbose:\n",
    "        print('best chisq & (fs, fb): ', chi2_min, fs, fb)\n",
    "    time_model = np.arange(time.min(), time.max(), 0.001)\n",
    "    magnifications = compute_model_lc(time_model, para_best, VBBL)\n",
    "    mag_model = 18 - 2.5*np.log10(magnifications*fs + fb)\n",
    "    model = np.vstack((time_model, mag_model))\n",
    "    return para_best, chi2_min, model, warnflag\n",
    "\n",
    "def prepare_lc_mdn(X, pis, locs):\n",
    "    mag = X[:, :, 1] / 5 + 18\n",
    "    flux = 10 ** (0.4 * (18 - mag))\n",
    "    merr = torch.ones_like(mag) * 0.033\n",
    "    ferr = merr*flux*np.log(10)/2.5\n",
    "    # times, mag, flux, ferr\n",
    "    lc = torch.stack([X[:, :, 0], mag, flux, ferr], dim=-1)\n",
    "    first_indices = torch.arange(len(pis))[:, None]\n",
    "    order = torch.argsort(pis, dim=-1, descending=True)\n",
    "    pis = pis[first_indices, order]\n",
    "    locs = locs[first_indices, order]\n",
    "    return lc.numpy(), pis.numpy(), locs.numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lc = torch.tensor(x_orig.reshape(1, *x_orig.shape))\n",
    "# lc[:, :, 1] = (lc[:, :, 1] - mbase) / 0.2\n",
    "\n",
    "if gap:\n",
    "    tel = pd.read_csv('./KMT/kb190371_aligned.csv', skiprows=0, usecols=[0, 2, 5, 6])\n",
    "    tel = tel[(tel['Tel'] != 'KMTS01') & (tel['Tel'] != 'KMTS41')]\n",
    "    lc = tel[['HJD', 'mag_aligned', 'e_mag_aligned']].to_numpy()[:, :-1].reshape(1, -1, 2)\n",
    "    lc[:, :, 1] = (lc[:, :, 1] - mbase) / 0.2\n",
    "else:\n",
    "    lc = np.loadtxt('./KMT/kb190371_aligned.csv', delimiter=',', skiprows=1, usecols=(2, 5)).reshape(1, -1, 2)\n",
    "    lc[:, :, 1] = (lc[:, :, 1] - mbase) / 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f0de9d15c10>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(lc[0, :, 0], lc[0, :, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "lc, pis_sort, locs_sort = prepare_lc_mdn(torch.tensor(lc), pi.probs.cpu(), normal.loc.cpu())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f0de0965d00>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(lc[0, :, 0], lc[0, :, 1]-18)\n",
    "plt.scatter(data[0, :, 0] * pred_t[0][1] + pred_t[0][0], data[0, :, 1]/5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "dof = lc.shape[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9606f161b293453d906722e74a8588b3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.13316278159618378, -1.138269066810608, -0.07768447697162628, 1.514648199081421, -2, 8592.391, 6.682251]\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 180511.029570\n",
      "         Iterations: 695\n",
      "         Function evaluations: 1099\n",
      "initial chisq:  185294.8626020858\n",
      "best chisq & (fs, fb):  180511.02956962617 0.8492246167160573 0.12902535943228344\n",
      "fs 0.8492246167160573 fb 0.12902535943228344\n",
      "[0.08196108788251877, -1.0764445066452026, 0.15759719908237457, 1.504199504852295, -2, 8592.391, 6.682251]\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 180510.126938\n",
      "         Iterations: 423\n",
      "         Function evaluations: 686\n",
      "initial chisq:  187092.07301437866\n",
      "best chisq & (fs, fb):  180510.1269376604 0.9078577435511193 0.07030168240265766\n",
      "fs 0.9078577435511193 fb 0.07030168240265766\n",
      "best params: [array([ 1.80511030e+05,  0.00000000e+00,  1.66004226e-01, -2.17830934e+00,\n",
      "       -1.15653836e+00, -5.66534259e-02,  1.50776405e+00,  8.59239205e+03,\n",
      "        5.92456888e+00, -6.14272713e-02]), array([ 1.80510127e+05,  0.00000000e+00,  7.79152620e-02, -2.13532856e+00,\n",
      "       -9.52431520e-01,  1.79496598e-01,  1.50692431e+00,  8.59240408e+03,\n",
      "        5.99073566e+00, -3.23918421e-02])]\n",
      "[ 1.66004226e-01 -2.17830934e+00 -1.15653836e+00 -5.66534259e-02\n",
      "  1.50776405e+00  8.59239205e+03  5.92456888e+00 -6.14272713e-02]\n",
      "[ 7.79152620e-02 -2.13532856e+00 -9.52431520e-01  1.79496598e-01\n",
      "  1.50692431e+00  8.59240408e+03  5.99073566e+00 -3.23918421e-02]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x864 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "i = 0\n",
    "verbose = True\n",
    "\n",
    "fig = plt.figure(1, (12, 12))\n",
    "\n",
    "ax_lc = plt.subplot2grid(shape=(3, 3), loc=(0, 0), rowspan=1, colspan=3)\n",
    "\n",
    "# plt.xlabel(r'$\\tau=(t - t_0)/t_E$', fontsize=20)\n",
    "# plt.ylabel(r'$m$ - $m_{\\rm base}$', fontsize=20)\n",
    "plt.xlabel(r'HJD-2450000', fontsize=20)\n",
    "plt.ylabel(r'$m$', fontsize=20)\n",
    "\n",
    "best_parameters = []\n",
    "colors = ['orange', 'green', 'black']\n",
    "labels = ['close', 'wide', 'black']\n",
    "linestyle = ['-', 'dashed', '-']\n",
    "minimum_chi2 = np.inf\n",
    "for index in tqdm(range(2)):\n",
    "    para_initial = locs_sort[i, index, :-1]\n",
    "    para_initial = para_initial.tolist()\n",
    "    if index == 0 and gap == True:\n",
    "        para_initial.insert(4, -1.7) # lgrho\n",
    "    else:\n",
    "        para_initial.insert(4, -2)\n",
    "    para_initial.insert(5, pred_t[0][0]) # t0\n",
    "    para_initial.insert(6, pred_t[0][1]) # te\n",
    "    if verbose:\n",
    "        print(para_initial)\n",
    "    para_best, chi2_min, model, warnflag = perform_optimization(lc[i, :, 0], lc[i, :, 2], lc[i, :, 3], para_initial, verbose=verbose)\n",
    "    chi2_min, fs, fb = compute_chisq(para_best, lc[i, :, 0], lc[i, :, 2], lc[i, :, 3], VBBinaryLensing.VBBinaryLensing(), return_model=True)\n",
    "    print('fs', fs, 'fb', fb)\n",
    "    if chi2_min < minimum_chi2:\n",
    "        minimum_chi2 = chi2_min\n",
    "    if len(para_best) == 5:\n",
    "        lgrho = para_best[-1]\n",
    "        para_best = np.delete(para_best, -1).tolist()\n",
    "        para_best.insert(1, lgrho)\n",
    "    if len(para_best) == 7:\n",
    "        lgrho = para_best[-3]\n",
    "        para_best = np.delete(para_best, -3).tolist()\n",
    "        para_best.insert(1, lgrho)\n",
    "    best_param = np.hstack((chi2_min, warnflag, para_best, np.array([np.log10(fs/(fs+fb))])))\n",
    "    best_parameters.append(best_param)\n",
    "    # plt.plot(model[0] * pred_t[0][1] + pred_t[0][0], model[1] - 18 + mbase, label=r'%s, $\\chi^2$=%.1f'%(labels[index], chi2_min), color=colors[index], linestyle=linestyle[index])\n",
    "    plt.plot(model[0], model[1] - 18 + mbase, label=r'%s, $\\chi^2$=%.1f'%(labels[index], chi2_min / minimum_chi2 * dof), color=colors[index], linestyle=linestyle[index])\n",
    "\n",
    "print('best params:', best_parameters)\n",
    "\n",
    "\n",
    "tru_wide = truths_full[1].tolist()\n",
    "delta = 10**truths_full[1][2]/(1+10**truths_full[1][2])*(10**truths_full[1][3]-1/10**truths_full[1][3])\n",
    "tru_wide[0] -= delta * np.cos(tru_wide[-2]*np.pi-3*np.pi/2)\n",
    "\n",
    "\n",
    "tru_close = truths_full[0].tolist()\n",
    "tru_time_close = (lc[i, :, 0] * pred_t[0][1] + pred_t[0][0] - 8592.392) / 6.50\n",
    "tru_time_wide = (lc[i, :, 0] * pred_t[0][1] + pred_t[0][0] - 8592.391) / 6.57\n",
    "\n",
    "# chi2_min, fs, fb, model = compute_chisq(np.array(tru_close)[[0, 2, 3, 4]], tru_time_close, lc[i, :, 2], lc[i, :, 3], VBBinaryLensing.VBBinaryLensing(), rho=10**tru_close[1], return_model=True, return_lc=True)\n",
    "# truths[0][-1] = np.log10(fs/(fs+fb))\n",
    "# plt.plot(model[0], model[1], label=r'close, $\\chi^2$=%.1f'%(chi2_min), color='red', linestyle='dashed')\n",
    "\n",
    "# chi2_min, fs, fb, model = compute_chisq(np.array(tru_wide)[[0, 2, 3, 4]], tru_time_wide, lc[i, :, 2], lc[i, :, 3], VBBinaryLensing.VBBinaryLensing(), rho=10**tru_wide[1], return_model=True, return_lc=True)\n",
    "# truths[1][-1] = np.log10(fs/(fs+fb))\n",
    "# plt.plot(model[0], model[1], label=r'wide, $\\chi^2$=%.1f'%(chi2_min), color='blue', linestyle='dashed')\n",
    "\n",
    "\n",
    "# plt.scatter(lc[0, :, 0] * pred_t[0][1] + pred_t[0][0], lc[0, :, 1] - 18 + mbase, color='black', alpha=0.5, rasterized=True)\n",
    "plt.scatter(lc[0, :, 0], lc[0, :, 1] - 18 + mbase, color='black', alpha=0.5, rasterized=True)\n",
    "# plt.xlim(-0.3 * pred_t[0][1] + pred_t[0][0], 0.3 * pred_t[0][1] + pred_t[0][0])\n",
    "plt.xlim(8588, 8596)\n",
    "plt.ylim(16.5, 19.2)\n",
    "plt.legend(loc='upper right', prop={'size': 16})\n",
    "plt.gca().invert_yaxis()\n",
    "\n",
    "param_list = [tru_close, best_parameters[0][2:]]\n",
    "traj_color = ['red', 'orange']\n",
    "cau_color = traj_color\n",
    "\n",
    "ax_geo = plt.subplot2grid(shape=(3, 3), loc=(1, 1), rowspan=1, colspan=2)\n",
    "plt.axis('equal')\n",
    "plt.xlim(-1, 1)\n",
    "plt.ylim(-0.3, 0.3)\n",
    "plt.yticks([-0.5, 0, 0.5])\n",
    "for j, params in enumerate(param_list):\n",
    "    if j != 0:\n",
    "        print(params)\n",
    "        if len(params) == 5:\n",
    "            u_0, lgq, lgs, alpha_180, lgfs = params\n",
    "            lgrho = -3\n",
    "            t0, te = 0, 1\n",
    "        if len(params) == 8:\n",
    "            u_0, lgrho, lgq, lgs, alpha_180, t0, te, lgfs = params\n",
    "        else:\n",
    "            u_0, lgrho, lgq, lgs, alpha_180, lgfs = params\n",
    "            t0, te = 0, 1\n",
    "        parameters = {\n",
    "                    't_0': t0,\n",
    "                    't_E': te,\n",
    "                    'u_0': u_0,\n",
    "                    'rho': 10**lgrho, \n",
    "                    'q': 10**lgq, \n",
    "                    's': 10**lgs, \n",
    "                    'alpha': alpha_180*180,\n",
    "                }\n",
    "        modelmm = mm.Model(parameters, coords=None)\n",
    "        if j == 0 or j == 1:\n",
    "            modelmm.plot_trajectory(t_range=(t0-2*te, t0+2*te), caustics=False, arrow=True, color=traj_color[j], linestyle='dashed')\n",
    "        else:\n",
    "            modelmm.plot_trajectory(t_range=(t0-2*te, t0+2*te), caustics=False, arrow=False, color=traj_color[j])\n",
    "        modelmm.plot_caustics(color=cau_color[j], s=3)\n",
    "\n",
    "param_list = [tru_wide, best_parameters[1][2:]]\n",
    "traj_color = ['blue', 'green']\n",
    "cau_color = traj_color\n",
    "\n",
    "ax_geo = plt.subplot2grid(shape=(3, 3), loc=(2, 1), rowspan=1, colspan=2)\n",
    "plt.axis('equal')\n",
    "plt.xlim(-1, 1)\n",
    "plt.ylim(-0.3, 0.3)\n",
    "plt.yticks([-0.5, 0, 0.5])\n",
    "for j, params in enumerate(param_list):\n",
    "    if j != 0:\n",
    "        print(params)\n",
    "        if len(params) == 5:\n",
    "            u_0, lgq, lgs, alpha_180, lgfs = params\n",
    "            lgrho = -3\n",
    "            t0, te = 0, 1\n",
    "        if len(params) == 8:\n",
    "            u_0, lgrho, lgq, lgs, alpha_180, t0, te, lgfs = params\n",
    "        else:\n",
    "            u_0, lgrho, lgq, lgs, alpha_180, lgfs = params\n",
    "            t0, te = 0, 1\n",
    "        parameters = {\n",
    "                    't_0': t0,\n",
    "                    't_E': te,\n",
    "                    'u_0': u_0,\n",
    "                    'rho': 10**lgrho, \n",
    "                    'q': 10**lgq, \n",
    "                    's': 10**lgs, \n",
    "                    'alpha': alpha_180*180,\n",
    "                }\n",
    "        modelmm = mm.Model(parameters, coords=None)\n",
    "        if j == 0 or j == 1:\n",
    "            modelmm.plot_trajectory(t_range=(t0-2*te, t0+2*te), caustics=False, arrow=True, color=traj_color[j], linestyle='dashed')\n",
    "        else:\n",
    "            modelmm.plot_trajectory(t_range=(t0-2*te, t0+2*te), caustics=False, arrow=False, color=traj_color[j])\n",
    "        modelmm.plot_caustics(color=cau_color[j], s=3)\n",
    "\n",
    "# ax = plt.subplot2grid((3, 3), (2, 1), rowspan=1, colspan=2)\n",
    "# ax.axis('off')\n",
    "# plt.text(-0.5, 0.7, r'Truth close:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%.2f$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "#     tru_close[0], tru_close[2], tru_close[3], tru_close[1], tru_close[4]*180, tru_close[5]\n",
    "# ), fontsize=20, color='red')\n",
    "# plt.text(0.2, 0.7, r'Truth wide:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%.2f$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "#     tru_wide[0], tru_wide[2], tru_wide[3], tru_wide[1], tru_wide[4]*180, tru_wide[5]\n",
    "# ), fontsize=20, color='blue')\n",
    "# # plt.text(0.2, 0.5, r'Closest Peak:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%d$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "# #     param_pred_gap[0], param_pred_gap[2], param_pred_gap[3], param_pred_gap[1], param_pred_gap[4]*180, param_pred_gap[5]\n",
    "# # ), fontsize=20)\n",
    "# # plt.text(0.2, 0.1, r'Global Peak:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%d$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "# #     param_pred_gap_g[0], param_pred_gap_g[2], param_pred_gap_g[3], param_pred_gap_g[1], param_pred_gap_g[4]*180, param_pred_gap_g[5]\n",
    "# # ), fontsize=20)\n",
    "# plt.text(-0.5, 0.3, r'Predicted close:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%d$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "#     best_parameters[0][2:][0], best_parameters[0][2:][1], best_parameters[0][2:][2], -3, best_parameters[0][2:][3]*180, best_parameters[0][2:][4]\n",
    "# ), fontsize=20, color='orange')\n",
    "# plt.text(0.2, 0.3, r'Predicted wide:' '\\n' r'$u_0=%.3f$, $\\lg q=%.3f$, $\\lg s=%.3f$,' '\\n' r'$\\lg\\rho=%d$, $\\alpha=%.3f\\degree$, $\\lg f_{\\rm S}=%.3f$' % (\n",
    "#     best_parameters[1][2:][0], best_parameters[1][2:][1], best_parameters[1][2:][2], -3, best_parameters[1][2:][3]*180, best_parameters[1][2:][4]\n",
    "# ), fontsize=20, color='green')\n",
    "plt.tight_layout()\n",
    "if gap:\n",
    "    plt.savefig('./lc_kmt_gap.pdf')\n",
    "else:\n",
    "    plt.savefig('./lc_kmt.pdf')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8592.391578646471, 5.289018166473633, 271.397529),\n",
       " (8592.400288583975, 5.196841372248413, 271.2463758))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "((1.12957975e-04 / 7.91502472e-01 * pred_t[0][1] + pred_t[0][0],\n",
    "    7.91502472e-01 * pred_t[0][1],\n",
    "    1.50776405 * 180),\n",
    "(1.12468815e-03 / 7.77708199e-01 * pred_t[0][1] + pred_t[0][0],\n",
    "    7.77708199e-01 * pred_t[0][1],\n",
    "    1.50692431 * 180))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8592.395377578581, 5.858886109434448, 270.0385686),\n",
       " (8592.395035698297, 6.184700598906372, 269.32772159999996))"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "((6.23589503e-04 / 8.76783307e-01 * pred_t[0][1] + pred_t[0][0],\n",
    "    8.76783307e-01 * pred_t[0][1],\n",
    "    1.50021427e+00 * 180),\n",
    "(6.10914546e-04 / 9.25541501e-01 * pred_t[0][1] + pred_t[0][0],\n",
    "    9.25541501e-01 * pred_t[0][1],\n",
    "    1.49626512e+00 * 180))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.76511616 0.21410677 0.00878714 0.00530975]\n",
      "[[ 0.14       -1.10237291 -0.08301995  1.50916616]\n",
      " [ 0.141      -0.91009489  0.19728056  1.50852954]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 612x612 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "checkpt = torch.load('/work/hmzhao/experiments/cde_mdn/experiment_l32nG12diag.ckpt', map_location='cpu')\n",
    "ckpt_args = checkpt['args']\n",
    "state_dict = checkpt['state_dict']\n",
    "\n",
    "output_dim = 5\n",
    "input_dim = data.shape[-1]\n",
    "latent_dim = ckpt_args.latents\n",
    "\n",
    "model = CDE_MDN(input_dim, latent_dim, output_dim).to(device)\n",
    "model_dict = model.state_dict()\n",
    "\n",
    "# 1. filter out unnecessary keys\n",
    "state_dict = {k: v for k, v in state_dict.items() if k in model_dict}\n",
    "# 2. overwrite entries in the existing state dict\n",
    "model_dict.update(state_dict) \n",
    "# 3. load the new state dict\n",
    "model.load_state_dict(state_dict)\n",
    "model.to(device)\n",
    "with torch.no_grad():\n",
    "    pi, normal = model(train_coeffs.float().to(device))\n",
    "plt.rc('font', size=12)\n",
    "import plot_triangle\n",
    "## first sort the gaussians according to the weight ##\n",
    "i = 0\n",
    "pi = pi.probs.cpu().numpy()[i]\n",
    "mu = normal.loc.cpu().numpy()[i]\n",
    "scale = normal.scale.cpu().numpy()[i]\n",
    "order = np.argsort(pi)[::-1]\n",
    "pi = pi[order][:4]\n",
    "mu = mu[order][:4][:, :-1]\n",
    "scale = scale[order][:4][:, :-1]\n",
    "print(pi)\n",
    "## the ground truth ##\n",
    "# param_true = [0.34033019, -0.35541759, 0.01886, 1.18859352, -0.8133]\n",
    "param_true = truths[:, :-1]\n",
    "print(param_true)\n",
    "#param_true = [0.247, -0.074, 0.227, 291.236/180, -0.094]\n",
    "param_pred = np.zeros_like(param_true)\n",
    "norm_fac = np.zeros_like(param_true)\n",
    "\n",
    "param_best = best_parameters.copy()\n",
    "for i in range(len(param_best)):\n",
    "    param_best[i] = np.delete(param_best[i][2:], 1)\n",
    "\n",
    "n_param = len(mu[0])\n",
    "# labels = [r'$u_0$', r'$\\lg q$', r'$\\lg s$', r'$\\alpha/180\\degree$', r'$\\lg f_{\\rm s}$']\n",
    "labels = [r'$u_0$', r'$\\lg q$', r'$\\lg s$', r'$\\alpha/180\\degree$']\n",
    "\n",
    "# param_ranges = [[0, 0.3], [-1.55, -0.8], [-0.2, 0.3], [1.4, 1.6], [-0.2, 0.01]]\n",
    "param_ranges = [[0, 0.3], [-1.55, -0.8], [-0.2, 0.3], [1.4, 1.6]]\n",
    "for j_gauss in range(len(pi)):\n",
    "    weight = pi[j_gauss]/pi[0] # normalize the weight by the highest value, so that the color appears better\n",
    "    cov_mat = np.zeros((n_param, n_param))\n",
    "    for i in range(len(mu[j_gauss])):\n",
    "        cov_mat[i, i] = scale[j_gauss, i]**2\n",
    "    norm_fac += pi[j_gauss] #/np.linalg.det(cov_mat)\n",
    "    param_pred += mu[j_gauss] * pi[j_gauss] #/np.linalg.det(cov_mat) * mu[j_gauss]\n",
    "    if j_gauss == 0:\n",
    "        fig, axes = plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, weight=weight, ground_truth_color='red', ground_truth_ls='dashed')\n",
    "        # fig, axes = plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, weight=weight, ground_truth=param_true[0], ground_truth_color='red', ground_truth_ls='dashed')\n",
    "        # plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, fig=fig, axes=axes, weight=weight, ground_truth=param_true[1], ground_truth_color='blue', ground_truth_ls='dashed')\n",
    "    if j_gauss == len(pi)-1:\n",
    "        param_pred /= norm_fac\n",
    "        # plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, fig=fig, axes=axes, weight=weight)\n",
    "        plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, fig=fig, axes=axes, weight=weight, extents=param_ranges, ground_truth=np.array(param_best)[0], ground_truth_color='orange', ground_truth_ls='-')\n",
    "        plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, fig=fig, axes=axes, weight=weight, extents=param_ranges, ground_truth=np.array(param_best)[1], ground_truth_color='green', ground_truth_ls='-')\n",
    "    else:\n",
    "        plot_triangle.plot_covariance(mu[j_gauss], labels, cov_mat, fig=fig, axes=axes, weight=weight)\n",
    "\n",
    "# mu_truth = param_true\n",
    "# cov_mat_truth = [np.diag((0.007, 0.002/0.079/np.log(10), 0.011/0.826/np.log(10), 0.008/np.pi, 0.492/0.190/np.log(10)*np.sqrt(3)*0.01))**2, \n",
    "#                 np.diag((0.007, 0.003/0.079/np.log(10), 0.014/0.826/np.log(10), 0.007/np.pi, 0.535/0.206/np.log(10)*np.sqrt(3)*0.01))**2]\n",
    "# plot_triangle.plot_covariance(mu_truth[0], labels, cov_mat_truth[0], fig=fig, axes=axes, weight=0.5, extents=param_ranges)\n",
    "# plot_triangle.plot_covariance(mu_truth[1], labels, cov_mat_truth[1], fig=fig, axes=axes, weight=0.5, extents=param_ranges)\n",
    "if gap:\n",
    "    plt.savefig('./corner_kmt_gap.pdf')\n",
    "else:\n",
    "    plt.savefig('./corner_kmt.pdf')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "274e8ed069bc503cc01fa1aa16068ccb03180744fbf761618fb0dbde6179c71c"
  },
  "kernelspec": {
   "display_name": "Python 3.9.12 ('ode-rnn')",
   "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": "3.9.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}