7.model_trackster_graph_net.classification.multiparticle.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6c07744a",
   "metadata": {},
   "source": [
    "# Graph net for node classification\n",
    "\n",
    "Question: does including the neighborhood provide a benefit over the pairwise classification?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e3962d51",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import sys\n",
    "\n",
    "import torch.nn as nn\n",
    "from torch.optim import Adam, SGD\n",
    "from torch_cluster import knn_graph\n",
    "from torch.optim.lr_scheduler import CosineAnnealingLR\n",
    "\n",
    "from torch.utils.data import random_split\n",
    "from torch_geometric.loader import DataLoader\n",
    "import torch_geometric.transforms as T\n",
    "\n",
    "from reco.model import DynamicEdgeConvBlock\n",
    "\n",
    "from reco.training import precision_recall_curve, roc_auc\n",
    "from reco.loss import GraphClassificationLoss, GraphClassificationLossSingleClass\n",
    "from reco.datasetPU import TracksterGraph\n",
    "\n",
    "import torch_geometric.nn as gnn\n",
    "\n",
    "\n",
    "# TODO: for PU, I trained on 10GeV and evaluated on 5Gev\n",
    "# Retrain when larger dataset is ready (100E)\n",
    "ds_name = \"MultiParticle\"\n",
    "\n",
    "data_root = \"/mnt/ceph/users/ecuba/processed\"\n",
    "raw_dir = f\"/mnt/ceph/users/ecuba/{ds_name}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "02f59304-76e4-4132-a379-978a46a042fd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using device: cuda\n"
     ]
    }
   ],
   "source": [
    "# CUDA Setup\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else \"cpu\")\n",
    "print(f\"Using device: {device}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "012b3572",
   "metadata": {},
   "outputs": [],
   "source": [
    "def knn_transform(data):\n",
    "    # pos coordinates are on position 3:6\n",
    "    data.edge_index = knn_graph(data.pos, k=4, loop=False)\n",
    "    return data\n",
    "\n",
    "def prepend_pos(data):\n",
    "    # GravNet: add S to the front\n",
    "    data.x = torch.hstack((data.pos, data.x))\n",
    "    return data\n",
    "\n",
    "transforms = T.Compose([prepend_pos])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ba68cd4a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TracksterGraph(graphs=39548, nodes=2622390, radius=30, bigT_e_th=10)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ds = TracksterGraph(\n",
    "    ds_name,\n",
    "    data_root,\n",
    "    raw_dir,\n",
    "    N_FILES=10,\n",
    "    pileup=False,\n",
    "    radius=30,\n",
    "    bigT_e_th=10,\n",
    ")\n",
    "ds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "604298f4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train graphs: 37571, Test graphs: 1977, total: 39548\n"
     ]
    }
   ],
   "source": [
    "ds_size = len(ds)\n",
    "test_set_size = ds_size // 20\n",
    "train_set_size = ds_size - test_set_size\n",
    "train_set, test_set = random_split(ds, [train_set_size, test_set_size])\n",
    "print(f\"Train graphs: {len(train_set)}, Test graphs: {len(test_set)}, total: {ds_size}\")\n",
    "\n",
    "train_dl = DataLoader(train_set, batch_size=8, shuffle=True)\n",
    "test_dl = DataLoader(test_set, batch_size=8, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "3f4a0b99-bef3-4c82-a0ac-d88e1521382b",
   "metadata": {},
   "outputs": [],
   "source": [
    "class TracksterGraphNet(nn.Module):\n",
    "    def __init__(self, input_dim, output_dim=2, dropout=0.2):\n",
    "        super(TracksterGraphNet, self).__init__()\n",
    "\n",
    "        self.norm = gnn.LayerNorm(input_dim)\n",
    "        \n",
    "        hdim1 = 64\n",
    "        hdim2 = 64\n",
    "        hdim3 = 64\n",
    "\n",
    "        hdim_fc = 256\n",
    "\n",
    "        self.edgeconv1 = DynamicEdgeConvBlock(input_dim, hdim1)\n",
    "        self.edgeconv2 = DynamicEdgeConvBlock(hdim1, hdim2)\n",
    "        self.edgeconv3 = DynamicEdgeConvBlock(hdim2, hdim3)\n",
    "        \n",
    "        # Edge features from node embeddings for classification        \n",
    "        self.nodenetwork = nn.Sequential(\n",
    "            nn.Linear(hdim3, hdim_fc),\n",
    "            nn.ReLU(),\n",
    "            nn.Dropout(dropout),\n",
    "            nn.Linear(hdim_fc, output_dim),\n",
    "            # nn.Sigmoid(),\n",
    "            nn.Softmax(dim=1),\n",
    "        )\n",
    "\n",
    "    def forward(self, X, batch=None):\n",
    "        H0 = self.norm(X, batch)\n",
    "        H1 = self.edgeconv1(H0, batch)\n",
    "        H2 = self.edgeconv2(H1, batch)\n",
    "        H3 = self.edgeconv3(H2, batch)\n",
    "        return self.nodenetwork(H3).squeeze(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "2539cf65",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_graph_classification(model, device, optimizer, loss_func, train_dl):\n",
    "    train_loss = 0.\n",
    "    model.train()\n",
    "\n",
    "    for data in train_dl:\n",
    "        data = data.to(device)\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        seg_pred = model(data.x, data.batch)\n",
    "        loss = loss_func(seg_pred, data)\n",
    "        loss.backward()\n",
    "\n",
    "        optimizer.step()\n",
    "        train_loss += loss.item()\n",
    "\n",
    "    return train_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "259ee8b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "@torch.no_grad()\n",
    "def test_graph_classification(model, device, loss_func, test_dl):\n",
    "    test_loss = 0.\n",
    "    model.eval()\n",
    "    for data in test_dl:\n",
    "        data = data.to(device)\n",
    "        seg_pred = model(data.x, data.batch)\n",
    "        loss = loss_func(seg_pred, data)\n",
    "        test_loss += loss.item()\n",
    "    return test_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "b85638de",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = TracksterGraphNet(input_dim=ds.data.x.shape[1], output_dim=2)\n",
    "epochs = 101\n",
    "model_path = f\"models/TracksterGraphNet.ParticleNetLight.ns.{epochs}e-{ds_name}.r{ds.RADIUS}.e{ds.bigT_e_th}.f{ds.N_FILES}.pt\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "2d3f1e82-74c5-4629-9b20-f05d7662aaf1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 0:\t train loss: 1702.978\t validation loss: 71.179\n",
      "Epoch 5:\t train loss: 1195.045\t validation loss: 64.787\n",
      "Epoch 10:\t train loss: 1115.874\t validation loss: 57.540\n",
      "Epoch 15:\t train loss: 1065.958\t validation loss: 63.503\n",
      "Epoch 20:\t train loss: 1047.639\t validation loss: 53.832\n",
      "Epoch 25:\t train loss: 1025.324\t validation loss: 53.450\n",
      "Epoch 30:\t train loss: 1004.154\t validation loss: 55.199\n",
      "Epoch 35:\t train loss: 984.669\t validation loss: 53.552\n",
      "Epoch 40:\t train loss: 968.501\t validation loss: 52.091\n",
      "Epoch 45:\t train loss: 954.370\t validation loss: 52.649\n",
      "Epoch 50:\t train loss: 937.071\t validation loss: 52.465\n",
      "Epoch 55:\t train loss: 919.584\t validation loss: 52.176\n",
      "Epoch 60:\t train loss: 900.123\t validation loss: 53.355\n",
      "Epoch 65:\t train loss: 886.754\t validation loss: 52.823\n",
      "Epoch 70:\t train loss: 876.600\t validation loss: 53.248\n",
      "Epoch 75:\t train loss: 860.779\t validation loss: 52.732\n",
      "Epoch 80:\t train loss: 846.350\t validation loss: 54.108\n",
      "Epoch 85:\t train loss: 841.386\t validation loss: 54.298\n",
      "Epoch 90:\t train loss: 835.622\t validation loss: 53.274\n",
      "Epoch 95:\t train loss: 830.832\t validation loss: 54.409\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "models/TracksterGraphNet.ParticleNetLight.ns.101e-MultiParticle.r30.e10.f10.pt\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 100:\t train loss: 832.613\t validation loss: 53.189\n"
     ]
    }
   ],
   "source": [
    "#%%script echo skipping\n",
    "# alpha - percentage of negative edges\n",
    "loss_func = GraphClassificationLoss()\n",
    "\n",
    "model = model.to(device)\n",
    "optimizer = Adam(model.parameters(), lr=1e-3)\n",
    "scheduler = CosineAnnealingLR(optimizer, epochs, eta_min=1e-5)\n",
    "\n",
    "for epoch in range(epochs):\n",
    "\n",
    "    train_loss = train_graph_classification(\n",
    "        model,\n",
    "        device,\n",
    "        optimizer,\n",
    "        loss_func,\n",
    "        train_dl\n",
    "    )\n",
    "    \n",
    "    scheduler.step()\n",
    "\n",
    "    if epoch % 5 == 0:\n",
    "        test_loss = test_graph_classification(model, device, loss_func, test_dl)\n",
    "        print(\n",
    "            f\"Epoch {epoch}:\\t train loss: {train_loss:.3f}\\t validation loss: {test_loss:.3f}\",\n",
    "            file=sys.stderr\n",
    "        )\n",
    "        \n",
    "torch.save(model.state_dict(), model_path)\n",
    "print(model_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "0b9cbaa6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "skipping\n"
     ]
    }
   ],
   "source": [
    "%%script echo skipping\n",
    "model.load_state_dict(torch.load(\n",
    "    model_path,\n",
    "    map_location=device\n",
    "))\n",
    "model = model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "2eb0507a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Future exception was never retrieved\n",
      "future: <Future finished exception=BrokenPipeError(32, 'Broken pipe')>\n",
      "Traceback (most recent call last):\n",
      "  File \"/mnt/sw/nix/store/z6v265ivx5w6xbajay41vdfc0la8gla5-python-3.9.12-view/lib/python3.9/asyncio/unix_events.py\", line 665, in write\n",
      "    n = os.write(self._fileno, data)\n",
      "BrokenPipeError: [Errno 32] Broken pipe\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9240373746507635\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TP: 13971, TN: 102170, FP: 3699, FN: 10666\n",
      "TH: 0.52 precision: 0.791 recall: 0.567 fbeta: 0.733 b_acc: 0.766\n"
     ]
    }
   ],
   "source": [
    "#%%script echo skipping - only evaluating foreground\n",
    "print(roc_auc(model, device, test_dl))\n",
    "precision_recall_curve(model, device, test_dl, focus_metric=\"fbeta\", beta=0.5, step=3, truth_threshold=0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb8ce4e2",
   "metadata": {},
   "source": [
    "## Eval"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "fb3f12d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import uproot\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from reco.data import get_event_data\n",
    "from reco.evaluation import model_evaluation\n",
    "from reco.dummy import DummyPleaser\n",
    "\n",
    "file_name = f\"{raw_dir}/new_ntuples_15163313_1.root\"\n",
    "cluster_data, trackster_data, simtrackster_data, assoc_data = get_event_data(file_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "6b5ffc46",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " --- Radius threshold: 5 ---\n",
      "Event 0:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.148 F: 0.464\n",
      "\ttarget_to_sim:\tP: 0.995 R: 0.148 F: 0.463\n",
      "\treco_to_sim:\tP: 0.991 R: 0.159 F: 0.484\n",
      "\t|S| = 10 |T| = 16 |R| = 12\n",
      "Event 1:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.093 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.949 R: 0.166 F: 0.488\n",
      "\treco_to_sim:\tP: 0.921 R: 0.173 F: 0.495\n",
      "\t|S| = 10 |T| = 22 |R| = 21\n",
      "Event 2:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.189 F: 0.536\n",
      "\ttarget_to_sim:\tP: 0.993 R: 0.198 F: 0.550\n",
      "\treco_to_sim:\tP: 0.991 R: 0.203 F: 0.558\n",
      "\t|S| = 10 |T| = 14 |R| = 12\n",
      "Event 3:\n",
      "\tclue3d_to_sim:\tP: 0.975 R: 0.177 F: 0.513\n",
      "\ttarget_to_sim:\tP: 0.978 R: 0.246 F: 0.612\n",
      "\treco_to_sim:\tP: 0.967 R: 0.254 F: 0.620\n",
      "\t|S| = 10 |T| = 18 |R| = 18\n",
      "Event 4:\n",
      "\tclue3d_to_sim:\tP: 0.963 R: 0.095 F: 0.340\n",
      "\ttarget_to_sim:\tP: 0.960 R: 0.145 F: 0.452\n",
      "\treco_to_sim:\tP: 0.949 R: 0.156 F: 0.470\n",
      "\t|S| = 10 |T| = 21 |R| = 17\n",
      "Event 5:\n",
      "\tclue3d_to_sim:\tP: 0.979 R: 0.136 F: 0.437\n",
      "\ttarget_to_sim:\tP: 0.979 R: 0.168 F: 0.498\n",
      "\treco_to_sim:\tP: 0.977 R: 0.171 F: 0.502\n",
      "\t|S| = 10 |T| = 16 |R| = 16\n",
      "Event 6:\n",
      "\tclue3d_to_sim:\tP: 0.982 R: 0.199 F: 0.550\n",
      "\ttarget_to_sim:\tP: 0.982 R: 0.245 F: 0.613\n",
      "\treco_to_sim:\tP: 0.962 R: 0.248 F: 0.610\n",
      "\t|S| = 10 |T| = 15 |R| = 13\n",
      "Event 7:\n",
      "\tclue3d_to_sim:\tP: 0.990 R: 0.171 F: 0.505\n",
      "\ttarget_to_sim:\tP: 0.992 R: 0.191 F: 0.540\n",
      "\treco_to_sim:\tP: 0.987 R: 0.198 F: 0.550\n",
      "\t|S| = 10 |T| = 12 |R| = 10\n",
      "Event 8:\n",
      "\tclue3d_to_sim:\tP: 0.984 R: 0.121 F: 0.405\n",
      "\ttarget_to_sim:\tP: 0.987 R: 0.152 F: 0.470\n",
      "\treco_to_sim:\tP: 0.982 R: 0.170 F: 0.503\n",
      "\t|S| = 10 |T| = 15 |R| = 12\n",
      "Event 9:\n",
      "\tclue3d_to_sim:\tP: 0.986 R: 0.195 F: 0.544\n",
      "\ttarget_to_sim:\tP: 0.989 R: 0.224 F: 0.587\n",
      "\treco_to_sim:\tP: 0.979 R: 0.235 F: 0.599\n",
      "\t|S| = 10 |T| = 21 |R| = 18\n",
      "Event 10:\n",
      "\tclue3d_to_sim:\tP: 0.974 R: 0.184 F: 0.524\n",
      "\ttarget_to_sim:\tP: 0.973 R: 0.238 F: 0.601\n",
      "\treco_to_sim:\tP: 0.965 R: 0.238 F: 0.599\n",
      "\t|S| = 10 |T| = 17 |R| = 17\n",
      "Event 11:\n",
      "\tclue3d_to_sim:\tP: 0.973 R: 0.183 F: 0.521\n",
      "\ttarget_to_sim:\tP: 0.977 R: 0.252 F: 0.620\n",
      "\treco_to_sim:\tP: 0.968 R: 0.263 F: 0.630\n",
      "\t|S| = 10 |T| = 25 |R| = 22\n",
      "Event 12:\n",
      "\tclue3d_to_sim:\tP: 0.958 R: 0.148 F: 0.457\n",
      "\ttarget_to_sim:\tP: 0.951 R: 0.233 F: 0.589\n",
      "\treco_to_sim:\tP: 0.874 R: 0.297 F: 0.630\n",
      "\t|S| = 10 |T| = 19 |R| = 14\n",
      "Event 13:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.070 F: 0.273\n",
      "\ttarget_to_sim:\tP: 0.978 R: 0.090 F: 0.330\n",
      "\treco_to_sim:\tP: 0.959 R: 0.104 F: 0.362\n",
      "\t|S| = 10 |T| = 25 |R| = 21\n",
      "Event 14:\n",
      "\tclue3d_to_sim:\tP: 0.961 R: 0.129 F: 0.419\n",
      "\ttarget_to_sim:\tP: 0.962 R: 0.166 F: 0.492\n",
      "\treco_to_sim:\tP: 0.935 R: 0.182 F: 0.511\n",
      "\t|S| = 10 |T| = 19 |R| = 15\n",
      "Event 15:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.236 F: 0.594\n",
      "\ttarget_to_sim:\tP: 0.948 R: 0.308 F: 0.669\n",
      "\treco_to_sim:\tP: 0.930 R: 0.318 F: 0.671\n",
      "\t|S| = 10 |T| = 21 |R| = 18\n",
      "Event 16:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.156 F: 0.476\n",
      "\ttarget_to_sim:\tP: 0.973 R: 0.190 F: 0.533\n",
      "\treco_to_sim:\tP: 0.954 R: 0.202 F: 0.546\n",
      "\t|S| = 10 |T| = 16 |R| = 14\n",
      "Event 17:\n",
      "\tclue3d_to_sim:\tP: 0.976 R: 0.092 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.976 R: 0.137 F: 0.438\n",
      "\treco_to_sim:\tP: 0.967 R: 0.138 F: 0.440\n",
      "\t|S| = 10 |T| = 23 |R| = 23\n",
      "Event 18:\n",
      "\tclue3d_to_sim:\tP: 0.980 R: 0.133 F: 0.431\n",
      "\ttarget_to_sim:\tP: 0.982 R: 0.141 F: 0.447\n",
      "\treco_to_sim:\tP: 0.964 R: 0.156 F: 0.474\n",
      "\t|S| = 10 |T| = 27 |R| = 23\n",
      "Event 19:\n",
      "\tclue3d_to_sim:\tP: 0.968 R: 0.136 F: 0.434\n",
      "\ttarget_to_sim:\tP: 0.972 R: 0.185 F: 0.525\n",
      "\treco_to_sim:\tP: 0.917 R: 0.224 F: 0.566\n",
      "\t|S| = 10 |T| = 16 |R| = 12\n",
      "-----\n",
      "mean clue3d_to_sim:\tP: 0.974 R: 0.150 F: 0.455\n",
      "mean target_to_sim:\tP: 0.975 R: 0.191 F: 0.526\n",
      "mean reco_to_sim:\tP: 0.957 R: 0.204 F: 0.541\n",
      "mean n_tracksters:\tP: 10.000 R: 18.900 F: 16.400\n",
      " --- Radius threshold: 10 ---\n",
      "Event 0:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.148 F: 0.464\n",
      "\ttarget_to_sim:\tP: 0.990 R: 0.168 F: 0.501\n",
      "\treco_to_sim:\tP: 0.986 R: 0.190 F: 0.536\n",
      "\t|S| = 10 |T| = 13 |R| = 11\n",
      "Event 1:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.093 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.943 R: 0.220 F: 0.569\n",
      "\treco_to_sim:\tP: 0.894 R: 0.208 F: 0.539\n",
      "\t|S| = 10 |T| = 15 |R| = 14\n",
      "Event 2:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.189 F: 0.536\n",
      "\ttarget_to_sim:\tP: 0.987 R: 0.262 F: 0.635\n",
      "\treco_to_sim:\tP: 0.976 R: 0.257 F: 0.626\n",
      "\t|S| = 10 |T| = 11 |R| = 11\n",
      "Event 3:\n",
      "\tclue3d_to_sim:\tP: 0.975 R: 0.177 F: 0.513\n",
      "\ttarget_to_sim:\tP: 0.969 R: 0.307 F: 0.677\n",
      "\treco_to_sim:\tP: 0.899 R: 0.328 F: 0.667\n",
      "\t|S| = 10 |T| = 14 |R| = 13\n",
      "Event 4:\n",
      "\tclue3d_to_sim:\tP: 0.963 R: 0.095 F: 0.340\n",
      "\ttarget_to_sim:\tP: 0.954 R: 0.217 F: 0.568\n",
      "\treco_to_sim:\tP: 0.904 R: 0.212 F: 0.547\n",
      "\t|S| = 10 |T| = 14 |R| = 14\n",
      "Event 5:\n",
      "\tclue3d_to_sim:\tP: 0.979 R: 0.136 F: 0.437\n",
      "\ttarget_to_sim:\tP: 0.972 R: 0.197 F: 0.544\n",
      "\treco_to_sim:\tP: 0.967 R: 0.245 F: 0.608\n",
      "\t|S| = 10 |T| = 12 |R| = 12\n",
      "Event 6:\n",
      "\tclue3d_to_sim:\tP: 0.982 R: 0.199 F: 0.550\n",
      "\ttarget_to_sim:\tP: 0.977 R: 0.354 F: 0.723\n",
      "\treco_to_sim:\tP: 0.954 R: 0.322 F: 0.685\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "Event 7:\n",
      "\tclue3d_to_sim:\tP: 0.990 R: 0.171 F: 0.505\n",
      "\ttarget_to_sim:\tP: 0.987 R: 0.211 F: 0.569\n",
      "\treco_to_sim:\tP: 0.973 R: 0.253 F: 0.620\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "Event 8:\n",
      "\tclue3d_to_sim:\tP: 0.984 R: 0.121 F: 0.405\n",
      "\ttarget_to_sim:\tP: 0.981 R: 0.248 F: 0.617\n",
      "\treco_to_sim:\tP: 0.977 R: 0.244 F: 0.610\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 9:\n",
      "\tclue3d_to_sim:\tP: 0.986 R: 0.195 F: 0.544\n",
      "\ttarget_to_sim:\tP: 0.980 R: 0.272 F: 0.644\n",
      "\treco_to_sim:\tP: 0.955 R: 0.297 F: 0.662\n",
      "\t|S| = 10 |T| = 13 |R| = 11\n",
      "Event 10:\n",
      "\tclue3d_to_sim:\tP: 0.974 R: 0.184 F: 0.524\n",
      "\ttarget_to_sim:\tP: 0.966 R: 0.322 F: 0.690\n",
      "\treco_to_sim:\tP: 0.928 R: 0.335 F: 0.685\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 11:\n",
      "\tclue3d_to_sim:\tP: 0.973 R: 0.183 F: 0.521\n",
      "\ttarget_to_sim:\tP: 0.950 R: 0.332 F: 0.693\n",
      "\treco_to_sim:\tP: 0.908 R: 0.349 F: 0.688\n",
      "\t|S| = 10 |T| = 19 |R| = 12\n",
      "Event 12:\n",
      "\tclue3d_to_sim:\tP: 0.958 R: 0.148 F: 0.457\n",
      "\ttarget_to_sim:\tP: 0.934 R: 0.286 F: 0.643\n",
      "\treco_to_sim:\tP: 0.874 R: 0.377 F: 0.691\n",
      "\t|S| = 10 |T| = 13 |R| = 11\n",
      "Event 13:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.070 F: 0.273\n",
      "\ttarget_to_sim:\tP: 0.965 R: 0.137 F: 0.437\n",
      "\treco_to_sim:\tP: 0.904 R: 0.149 F: 0.450\n",
      "\t|S| = 10 |T| = 17 |R| = 12\n",
      "Event 14:\n",
      "\tclue3d_to_sim:\tP: 0.961 R: 0.129 F: 0.419\n",
      "\ttarget_to_sim:\tP: 0.952 R: 0.232 F: 0.587\n",
      "\treco_to_sim:\tP: 0.898 R: 0.271 F: 0.614\n",
      "\t|S| = 10 |T| = 13 |R| = 12\n",
      "Event 15:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.236 F: 0.594\n",
      "\ttarget_to_sim:\tP: 0.946 R: 0.462 F: 0.782\n",
      "\treco_to_sim:\tP: 0.808 R: 0.459 F: 0.702\n",
      "\t|S| = 10 |T| = 12 |R| = 8\n",
      "Event 16:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.156 F: 0.476\n",
      "\ttarget_to_sim:\tP: 0.964 R: 0.232 F: 0.591\n",
      "\treco_to_sim:\tP: 0.936 R: 0.275 F: 0.632\n",
      "\t|S| = 10 |T| = 13 |R| = 9\n",
      "Event 17:\n",
      "\tclue3d_to_sim:\tP: 0.976 R: 0.092 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.971 R: 0.179 F: 0.515\n",
      "\treco_to_sim:\tP: 0.916 R: 0.189 F: 0.518\n",
      "\t|S| = 10 |T| = 17 |R| = 17\n",
      "Event 18:\n",
      "\tclue3d_to_sim:\tP: 0.980 R: 0.133 F: 0.431\n",
      "\ttarget_to_sim:\tP: 0.967 R: 0.191 F: 0.534\n",
      "\treco_to_sim:\tP: 0.820 R: 0.215 F: 0.525\n",
      "\t|S| = 10 |T| = 19 |R| = 10\n",
      "Event 19:\n",
      "\tclue3d_to_sim:\tP: 0.968 R: 0.136 F: 0.434\n",
      "\ttarget_to_sim:\tP: 0.956 R: 0.249 F: 0.609\n",
      "\treco_to_sim:\tP: 0.879 R: 0.277 F: 0.612\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "-----\n",
      "mean clue3d_to_sim:\tP: 0.974 R: 0.150 F: 0.455\n",
      "mean target_to_sim:\tP: 0.966 R: 0.254 F: 0.606\n",
      "mean reco_to_sim:\tP: 0.918 R: 0.273 F: 0.611\n",
      "mean n_tracksters:\tP: 10.000 R: 13.250 F: 11.250\n",
      " --- Radius threshold: 20 ---\n",
      "Event 0:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.148 F: 0.464\n",
      "\ttarget_to_sim:\tP: 0.985 R: 0.177 F: 0.516\n",
      "\treco_to_sim:\tP: 0.975 R: 0.212 F: 0.567\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 1:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.093 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.941 R: 0.249 F: 0.606\n",
      "\treco_to_sim:\tP: 0.848 R: 0.292 F: 0.614\n",
      "\t|S| = 10 |T| = 13 |R| = 15\n",
      "Event 2:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.189 F: 0.536\n",
      "\ttarget_to_sim:\tP: 0.986 R: 0.296 F: 0.673\n",
      "\treco_to_sim:\tP: 0.978 R: 0.293 F: 0.667\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "Event 3:\n",
      "\tclue3d_to_sim:\tP: 0.975 R: 0.177 F: 0.513\n",
      "\ttarget_to_sim:\tP: 0.968 R: 0.305 F: 0.674\n",
      "\treco_to_sim:\tP: 0.936 R: 0.309 F: 0.666\n",
      "\t|S| = 10 |T| = 12 |R| = 11\n",
      "Event 4:\n",
      "\tclue3d_to_sim:\tP: 0.963 R: 0.095 F: 0.340\n",
      "\ttarget_to_sim:\tP: 0.948 R: 0.210 F: 0.557\n",
      "\treco_to_sim:\tP: 0.925 R: 0.206 F: 0.545\n",
      "\t|S| = 10 |T| = 11 |R| = 14\n",
      "Event 5:\n",
      "\tclue3d_to_sim:\tP: 0.979 R: 0.136 F: 0.437\n",
      "\ttarget_to_sim:\tP: 0.970 R: 0.206 F: 0.557\n",
      "\treco_to_sim:\tP: 0.883 R: 0.258 F: 0.595\n",
      "\t|S| = 10 |T| = 11 |R| = 9\n",
      "Event 6:\n",
      "\tclue3d_to_sim:\tP: 0.982 R: 0.199 F: 0.550\n",
      "\ttarget_to_sim:\tP: 0.978 R: 0.349 F: 0.719\n",
      "\treco_to_sim:\tP: 0.948 R: 0.335 F: 0.694\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 7:\n",
      "\tclue3d_to_sim:\tP: 0.990 R: 0.171 F: 0.505\n",
      "\ttarget_to_sim:\tP: 0.988 R: 0.214 F: 0.574\n",
      "\treco_to_sim:\tP: 0.953 R: 0.258 F: 0.620\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 8:\n",
      "\tclue3d_to_sim:\tP: 0.984 R: 0.121 F: 0.405\n",
      "\ttarget_to_sim:\tP: 0.981 R: 0.250 F: 0.618\n",
      "\treco_to_sim:\tP: 0.947 R: 0.259 F: 0.618\n",
      "\t|S| = 10 |T| = 10 |R| = 8\n",
      "Event 9:\n",
      "\tclue3d_to_sim:\tP: 0.986 R: 0.195 F: 0.544\n",
      "\ttarget_to_sim:\tP: 0.978 R: 0.292 F: 0.666\n",
      "\treco_to_sim:\tP: 0.955 R: 0.321 F: 0.684\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 10:\n",
      "\tclue3d_to_sim:\tP: 0.974 R: 0.184 F: 0.524\n",
      "\ttarget_to_sim:\tP: 0.966 R: 0.328 F: 0.696\n",
      "\treco_to_sim:\tP: 0.934 R: 0.323 F: 0.678\n",
      "\t|S| = 10 |T| = 9 |R| = 10\n",
      "Event 11:\n",
      "\tclue3d_to_sim:\tP: 0.973 R: 0.183 F: 0.521\n",
      "\ttarget_to_sim:\tP: 0.949 R: 0.356 F: 0.712\n",
      "\treco_to_sim:\tP: 0.914 R: 0.368 F: 0.705\n",
      "\t|S| = 10 |T| = 12 |R| = 10\n",
      "Event 12:\n",
      "\tclue3d_to_sim:\tP: 0.958 R: 0.148 F: 0.457\n",
      "\ttarget_to_sim:\tP: 0.929 R: 0.268 F: 0.622\n",
      "\treco_to_sim:\tP: 0.802 R: 0.372 F: 0.651\n",
      "\t|S| = 10 |T| = 10 |R| = 8\n",
      "Event 13:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.070 F: 0.273\n",
      "\ttarget_to_sim:\tP: 0.949 R: 0.156 F: 0.471\n",
      "\treco_to_sim:\tP: 0.904 R: 0.174 F: 0.492\n",
      "\t|S| = 10 |T| = 12 |R| = 12\n",
      "Event 14:\n",
      "\tclue3d_to_sim:\tP: 0.961 R: 0.129 F: 0.419\n",
      "\ttarget_to_sim:\tP: 0.945 R: 0.258 F: 0.616\n",
      "\treco_to_sim:\tP: 0.914 R: 0.268 F: 0.617\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 15:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.236 F: 0.594\n",
      "\ttarget_to_sim:\tP: 0.938 R: 0.478 F: 0.786\n",
      "\treco_to_sim:\tP: 0.801 R: 0.477 F: 0.705\n",
      "\t|S| = 10 |T| = 10 |R| = 7\n",
      "Event 16:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.156 F: 0.476\n",
      "\ttarget_to_sim:\tP: 0.954 R: 0.286 F: 0.650\n",
      "\treco_to_sim:\tP: 0.937 R: 0.290 F: 0.648\n",
      "\t|S| = 10 |T| = 9 |R| = 9\n",
      "Event 17:\n",
      "\tclue3d_to_sim:\tP: 0.976 R: 0.092 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.918 R: 0.200 F: 0.535\n",
      "\treco_to_sim:\tP: 0.933 R: 0.197 F: 0.534\n",
      "\t|S| = 10 |T| = 11 |R| = 12\n",
      "Event 18:\n",
      "\tclue3d_to_sim:\tP: 0.980 R: 0.133 F: 0.431\n",
      "\ttarget_to_sim:\tP: 0.949 R: 0.207 F: 0.553\n",
      "\treco_to_sim:\tP: 0.930 R: 0.202 F: 0.540\n",
      "\t|S| = 10 |T| = 11 |R| = 11\n",
      "Event 19:\n",
      "\tclue3d_to_sim:\tP: 0.968 R: 0.136 F: 0.434\n",
      "\ttarget_to_sim:\tP: 0.949 R: 0.256 F: 0.615\n",
      "\treco_to_sim:\tP: 0.913 R: 0.282 F: 0.631\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "-----\n",
      "mean clue3d_to_sim:\tP: 0.974 R: 0.150 F: 0.455\n",
      "mean target_to_sim:\tP: 0.958 R: 0.267 F: 0.621\n",
      "mean reco_to_sim:\tP: 0.916 R: 0.285 F: 0.624\n",
      "mean n_tracksters:\tP: 10.000 R: 10.550 F: 10.200\n",
      " --- Radius threshold: 30 ---\n",
      "Event 0:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.148 F: 0.464\n",
      "\ttarget_to_sim:\tP: 0.984 R: 0.221 F: 0.582\n",
      "\treco_to_sim:\tP: 0.981 R: 0.227 F: 0.590\n",
      "\t|S| = 10 |T| = 9 |R| = 9\n",
      "Event 1:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.093 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.938 R: 0.262 F: 0.619\n",
      "\treco_to_sim:\tP: 0.865 R: 0.281 F: 0.611\n",
      "\t|S| = 10 |T| = 11 |R| = 12\n",
      "Event 2:\n",
      "\tclue3d_to_sim:\tP: 0.991 R: 0.189 F: 0.536\n",
      "\ttarget_to_sim:\tP: 0.986 R: 0.296 F: 0.672\n",
      "\treco_to_sim:\tP: 0.983 R: 0.295 F: 0.670\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "Event 3:\n",
      "\tclue3d_to_sim:\tP: 0.975 R: 0.177 F: 0.513\n",
      "\ttarget_to_sim:\tP: 0.968 R: 0.307 F: 0.677\n",
      "\treco_to_sim:\tP: 0.940 R: 0.305 F: 0.664\n",
      "\t|S| = 10 |T| = 11 |R| = 13\n",
      "Event 4:\n",
      "\tclue3d_to_sim:\tP: 0.963 R: 0.095 F: 0.340\n",
      "\ttarget_to_sim:\tP: 0.948 R: 0.212 F: 0.559\n",
      "\treco_to_sim:\tP: 0.906 R: 0.192 F: 0.520\n",
      "\t|S| = 10 |T| = 10 |R| = 10\n",
      "Event 5:\n",
      "\tclue3d_to_sim:\tP: 0.979 R: 0.136 F: 0.437\n",
      "\ttarget_to_sim:\tP: 0.971 R: 0.208 F: 0.559\n",
      "\treco_to_sim:\tP: 0.841 R: 0.208 F: 0.523\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 6:\n",
      "\tclue3d_to_sim:\tP: 0.982 R: 0.199 F: 0.550\n",
      "\ttarget_to_sim:\tP: 0.978 R: 0.350 F: 0.720\n",
      "\treco_to_sim:\tP: 0.944 R: 0.321 F: 0.680\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "Event 7:\n",
      "\tclue3d_to_sim:\tP: 0.990 R: 0.171 F: 0.505\n",
      "\ttarget_to_sim:\tP: 0.986 R: 0.220 F: 0.580\n",
      "\treco_to_sim:\tP: 0.957 R: 0.220 F: 0.573\n",
      "\t|S| = 10 |T| = 9 |R| = 7\n",
      "Event 8:\n",
      "\tclue3d_to_sim:\tP: 0.984 R: 0.121 F: 0.405\n",
      "\ttarget_to_sim:\tP: 0.981 R: 0.250 F: 0.618\n",
      "\treco_to_sim:\tP: 0.946 R: 0.261 F: 0.621\n",
      "\t|S| = 10 |T| = 10 |R| = 8\n",
      "Event 9:\n",
      "\tclue3d_to_sim:\tP: 0.986 R: 0.195 F: 0.544\n",
      "\ttarget_to_sim:\tP: 0.979 R: 0.292 F: 0.666\n",
      "\treco_to_sim:\tP: 0.948 R: 0.320 F: 0.681\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 10:\n",
      "\tclue3d_to_sim:\tP: 0.974 R: 0.184 F: 0.524\n",
      "\ttarget_to_sim:\tP: 0.966 R: 0.328 F: 0.696\n",
      "\treco_to_sim:\tP: 0.928 R: 0.322 F: 0.674\n",
      "\t|S| = 10 |T| = 9 |R| = 11\n",
      "Event 11:\n",
      "\tclue3d_to_sim:\tP: 0.973 R: 0.183 F: 0.521\n",
      "\ttarget_to_sim:\tP: 0.945 R: 0.358 F: 0.712\n",
      "\treco_to_sim:\tP: 0.925 R: 0.370 F: 0.712\n",
      "\t|S| = 10 |T| = 11 |R| = 9\n",
      "Event 12:\n",
      "\tclue3d_to_sim:\tP: 0.958 R: 0.148 F: 0.457\n",
      "\ttarget_to_sim:\tP: 0.930 R: 0.269 F: 0.623\n",
      "\treco_to_sim:\tP: 0.816 R: 0.370 F: 0.658\n",
      "\t|S| = 10 |T| = 10 |R| = 8\n",
      "Event 13:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.070 F: 0.273\n",
      "\ttarget_to_sim:\tP: 0.924 R: 0.188 F: 0.518\n",
      "\treco_to_sim:\tP: 0.911 R: 0.184 F: 0.509\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 14:\n",
      "\tclue3d_to_sim:\tP: 0.961 R: 0.129 F: 0.419\n",
      "\ttarget_to_sim:\tP: 0.945 R: 0.258 F: 0.616\n",
      "\treco_to_sim:\tP: 0.903 R: 0.257 F: 0.601\n",
      "\t|S| = 10 |T| = 10 |R| = 11\n",
      "Event 15:\n",
      "\tclue3d_to_sim:\tP: 0.956 R: 0.236 F: 0.594\n",
      "\ttarget_to_sim:\tP: 0.939 R: 0.473 F: 0.784\n",
      "\treco_to_sim:\tP: 0.780 R: 0.485 F: 0.695\n",
      "\t|S| = 10 |T| = 8 |R| = 7\n",
      "Event 16:\n",
      "\tclue3d_to_sim:\tP: 0.972 R: 0.156 F: 0.476\n",
      "\ttarget_to_sim:\tP: 0.953 R: 0.287 F: 0.651\n",
      "\treco_to_sim:\tP: 0.902 R: 0.328 F: 0.668\n",
      "\t|S| = 10 |T| = 9 |R| = 9\n",
      "Event 17:\n",
      "\tclue3d_to_sim:\tP: 0.976 R: 0.092 F: 0.335\n",
      "\ttarget_to_sim:\tP: 0.901 R: 0.203 F: 0.534\n",
      "\treco_to_sim:\tP: 0.928 R: 0.203 F: 0.542\n",
      "\t|S| = 10 |T| = 10 |R| = 13\n",
      "Event 18:\n",
      "\tclue3d_to_sim:\tP: 0.980 R: 0.133 F: 0.431\n",
      "\ttarget_to_sim:\tP: 0.951 R: 0.219 F: 0.570\n",
      "\treco_to_sim:\tP: 0.939 R: 0.208 F: 0.551\n",
      "\t|S| = 10 |T| = 10 |R| = 13\n",
      "Event 19:\n",
      "\tclue3d_to_sim:\tP: 0.968 R: 0.136 F: 0.434\n",
      "\ttarget_to_sim:\tP: 0.950 R: 0.255 F: 0.615\n",
      "\treco_to_sim:\tP: 0.909 R: 0.255 F: 0.601\n",
      "\t|S| = 10 |T| = 10 |R| = 9\n",
      "-----\n",
      "mean clue3d_to_sim:\tP: 0.974 R: 0.150 F: 0.455\n",
      "mean target_to_sim:\tP: 0.956 R: 0.273 F: 0.629\n",
      "mean reco_to_sim:\tP: 0.913 R: 0.281 F: 0.617\n",
      "mean n_tracksters:\tP: 10.000 R: 9.850 F: 10.050\n"
     ]
    }
   ],
   "source": [
    "r_ranges = [5, 10, 20, 30]\n",
    "results = []\n",
    "\n",
    "max_events = 20\n",
    "clue3D_F = []\n",
    "target_F = []\n",
    "model_reco_F = []\n",
    "\n",
    "for r in r_ranges:\n",
    "    print(f\" --- Radius threshold: {r} ---\")\n",
    "\n",
    "    result = model_evaluation(\n",
    "        cluster_data,\n",
    "        trackster_data,\n",
    "        simtrackster_data,\n",
    "        assoc_data,\n",
    "        model.to(\"cpu\"),\n",
    "        decision_th=0.5,\n",
    "        radius=r,\n",
    "        max_events=max_events,\n",
    "        bigT_e_th=5,\n",
    "        pileup=False,\n",
    "        graph=True,\n",
    "        multiparticle=True,\n",
    "    )\n",
    "    \n",
    "    clue3D_F.append(np.sum(np.array(result[\"clue3d_to_sim\"])[:,2]) / max_events)\n",
    "    target_F.append(np.sum(np.array(result[\"target_to_sim\"])[:,2]) / max_events)\n",
    "    model_reco_F.append(np.sum(np.array(result[\"reco_to_sim\"])[:,2]) / max_events)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "829a354a-0b77-46d7-b502-c08d0383ac17",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.45468812367852296, 0.45468812367852296, 0.45468812367852296, 0.45468812367852296]\n",
      "[0.5259644098162075, 0.6064421196483588, 0.6208185378572347, 0.6286151306406551]\n",
      "[0.5409682937473914, 0.6108236731056269, 0.6235436162880916, 0.6171658438392684]\n"
     ]
    }
   ],
   "source": [
    "print(clue3D_F)\n",
    "print(target_F)\n",
    "print(model_reco_F)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "6c2a36a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(8, 5))\n",
    "\n",
    "ax = fig.add_subplot(111)\n",
    "\n",
    "ax.plot(r_ranges[:6], (np.array(target_F) - np.array(clue3D_F))[:6], '--s', label=\"target $\\Delta F_{0.5}$\", c=\"#D55E00\")\n",
    "ax.plot(r_ranges[:6], (np.array(model_reco_F) - np.array(clue3D_F))[:6], '-o', label=\"GNN $\\Delta F_{0.5}$\", c=\"#56B4E9\")\n",
    "# ax.plot(r_ranges[:6], (np.array(naive_reco_F) - np.array(clue3D_F))[:6], '-v', label=\"naive $\\Delta F_{0.5}$\", c=\"#E69F00\")\n",
    "\n",
    "#ax.axhline(max(np.array(naive_reco_F) - np.array(clue3D_F)), label=\"baseline\", c=\"lightgray\", linestyle=\"--\")\n",
    "\n",
    "ax.legend()\n",
    "ax.set_xlabel(\"Neighborhood radius (cm)\")\n",
    "ax.set_ylabel(\"$\\Delta F_{0.5}$\")\n",
    "# ax.set_ylim((-0.1, 0.1))\n",
    "\n",
    "fig.savefig(f'export/{ds_name}-e30-pnetl-graph-radius-plot.png', dpi=300, bbox_inches='tight', transparent=True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "799ffc4a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reco 0.16885549260956867\n",
      "Target 0.17392700696213215\n"
     ]
    }
   ],
   "source": [
    "print(\"Reco\", max(np.array(model_reco_F) - np.array(clue3D_F)))\n",
    "# print(\"Naive\", max(np.array(naive_reco_F) - np.array(clue3D_F)))\n",
    "print(\"Target\", max(np.array(target_F) - np.array(clue3D_F)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "328866f1-c862-4dfb-a201-91d2baa53d9f",
   "metadata": {},
   "source": [
    "## Results\n",
    "\n",
    "Using alpha=balance seems to converge faster (100 epochs rather than 200)\n",
    "- maybe just sticking to 0.25? (works just fine)\n",
    "\n",
    "Hyperparams\n",
    "- (64, 128, noskip), alpha=balance, roc_auc = 0.9733\n",
    "- (64, 128, 256fc, noskip, 200e), alpha=1-balance, roc_auc = 0.9771\n",
    "- (64, 128, 256fc, noskip, 200e), alpha=balance, roc_auc = 0.981\n",
    "- (64, 128, 256fc, noskip, 100e), alpha=0.25, roc_auc = 0.9796"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ve",
   "language": "python",
   "name": "ve"
  },
  "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"
  },
  "vscode": {
   "interpreter": {
    "hash": "37c238d9819f69c2c770157eac01081978c120e64661e10d7fd52c4caf977dc9"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}