inductive_BTC.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload \n",
    "\n",
    "import time\n",
    "import os\n",
    "import numpy as np\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "import pickle\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from torch_geometric.data import Data\n",
    "from torch_geometric.utils import subgraph, to_undirected\n",
    "\n",
    "from torch_geometric.loader import DataLoader\n",
    "from torch_geometric.utils import subgraph\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "Download=True\n",
    "if Download:\n",
    "  !gdown 1CIFpAquzYBA98gQCdMb92fC0w6yrYf2m\n",
    "  !gdown 1Cfh0VIXWTc8EK96WRZdyaqgA2-JwvsUG\n",
    "  !gdown 1Cfh8hA9Tl8uCPrLSmcIQI3qCbEjOFl7C"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "Process = True\n",
    "if Process: \n",
    "    # Load data from the folder\n",
    "    df_features = pd.read_csv('elliptic_txs_features.csv',header=None)\n",
    "    df_edges = pd.read_csv(\"elliptic_txs_edgelist.csv\")\n",
    "    df_classes =  pd.read_csv(\"elliptic_txs_classes.csv\")\n",
    "    # Add class names for easy understanding\n",
    "    # reformat classes 0:licit, 1:illicit, 2:unknow \n",
    "    df_classes['class'] = df_classes['class'].map({'unknown': 2, '1':1, '2':0}) \n",
    "    # See repartition of nodes per class\n",
    "    group_class = df_classes.groupby('class').count()\n",
    "    # merge dataset\n",
    "    df_merge = df_features.merge(df_classes, how='left', right_on=\"txId\", left_on=0)\n",
    "    # Merge features with classes\n",
    "    df_merge = df_merge.sort_values(1).reset_index(drop=True)\n",
    "    # Setup trans ID to node ID mapping\n",
    "    nodes = df_merge[0].values\n",
    "\n",
    "    map_id = {j:i for i,j in enumerate(nodes)} # mapping nodes to indexes\n",
    "\n",
    "    # Create edge df that has transID mapped to nodeIDs\n",
    "    edges = df_edges.copy()\n",
    "    edges.txId1 = edges.txId1.map(map_id) #get nodes idx1 from edges list and filtered data\n",
    "    edges.txId2 = edges.txId2.map(map_id)\n",
    "\n",
    "    edges = edges.astype(int)\n",
    "\n",
    "    edge_index = np.array(edges.values).T #convert into an array\n",
    "    edge_index = torch.tensor(edge_index, dtype=torch.long).contiguous() # create a tensor\n",
    "\n",
    "    #create weights tensor with same shape of edge_index\n",
    "    weights = torch.tensor([1]* edge_index.shape[1] , dtype=torch.double) \n",
    "\n",
    "    # Define labels\n",
    "    labels = df_merge['class'].values\n",
    "\n",
    "    # mapping txIds to corresponding indices, to pass node features to the model\n",
    "\n",
    "    node_features = df_merge.drop(['txId'], axis=1).copy()\n",
    "\n",
    "    # Retain known vs unknown IDs\n",
    "    classified_idx = node_features['class'].loc[node_features['class']!=2].index # filter on known labels\n",
    "    unclassified_idx = node_features['class'].loc[node_features['class']==2].index\n",
    "\n",
    "    classified_illicit_idx = node_features['class'].loc[node_features['class']==1].index # filter on illicit labels\n",
    "    classified_licit_idx = node_features['class'].loc[node_features['class']==0].index # filter on licit labels\n",
    "\n",
    "    # Drop unwanted columns, 0 = transID, 1=time period, class = labels\n",
    "    node_features = node_features.drop(columns=[0, 1, 'class'])\n",
    "\n",
    "    # Convert to tensor\n",
    "    node_features_t = torch.tensor(np.array(node_features.values, dtype=np.double), dtype=torch.double)# drop unused columns\n",
    "\n",
    "    # Create a known vs unknown mask\n",
    "    train_idx, valid_idx = train_test_split(classified_idx.values, test_size=0.15)\n",
    "\n",
    "    # Create pyg dataset\n",
    "    data_train = Data(x=node_features_t, edge_index=edge_index, y=torch.tensor(labels, dtype=torch.double))\n",
    "    # Add in the train and valid idx\n",
    "    data_train.train_idx = train_idx\n",
    "    data_train.valid_idx = valid_idx\n",
    "    data_train.test_idx = unclassified_idx.values\n",
    "    data_train.step = torch.tensor(df_merge[1].values)\n",
    "    torch.save(data_train, 'data/EllipticBTC.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda:1' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "dataset = torch.load('./data/EllipticBTC.pt')\n",
    "dataset.edge_index = to_undirected(dataset.edge_index)\n",
    "dataset.y = dataset.y.to(torch.int64)\n",
    "dataset.x = dataset.x.to(torch.float32)\n",
    "dataset.train_mask = (dataset.step<=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch_geometric.loader import NeighborLoader\n",
    "\n",
    "dataset_train, dataset_val, dataset_test = [], [], []\n",
    "train_day = 30\n",
    "train_num = 0\n",
    "\n",
    "for i in dataset.step.unique():\n",
    "    idx = torch.where(dataset.step==i)[0]\n",
    "    x = dataset.x[idx]\n",
    "    y = dataset.y[idx]\n",
    "    edge_index, _= subgraph(idx, dataset.edge_index, relabel_nodes=True)\n",
    "    daily_graph = Data(x=x, edge_index=edge_index, y=y)\n",
    "    daily_graph.labeled_mask = (y<2) \n",
    "    if i<=train_day:\n",
    "        dataset_train.append(daily_graph)\n",
    "        train_num += daily_graph.num_nodes\n",
    "\n",
    "    elif i>train_day and i<=train_day+4:\n",
    "        dataset_val.append(daily_graph)\n",
    "    elif i>train_day+4:\n",
    "        dataset_test.append(daily_graph)\n",
    "\n",
    "kwargs = {'batch_size': 1024}\n",
    "train_loader = NeighborLoader(dataset, input_nodes=dataset.train_mask, num_neighbors=[25, 10], shuffle=True, **kwargs)\n",
    "val_loader = DataLoader(dataset_val, batch_size=1, shuffle=False)\n",
    "test_loader = DataLoader(dataset_test, batch_size=1, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score,f1_score,roc_auc_score\n",
    "\n",
    "def store_metrics(pred_scores, target_labels):\n",
    "\n",
    "  # calculate metrics\n",
    "  pred_labels = pred_scores.argmax(axis=1)\n",
    "  accuracy = accuracy_score(target_labels, pred_labels)\n",
    "  f1macro = f1_score(target_labels, pred_labels,average='macro')\n",
    "  aucroc = roc_auc_score(target_labels, pred_scores[:,-1])\n",
    "  \n",
    "  return accuracy, f1macro, aucroc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch_geometric.nn import GINConv, Linear\n",
    "\n",
    "class GINNet(torch.nn.Module):\n",
    "    def __init__(self, input_dim, output_dim, hidden_dim=256):\n",
    "        super().__init__()\n",
    "        nn1 = nn.Sequential(Linear(input_dim, 4*hidden_dim),\n",
    "                            nn.ReLU(),Linear(4*hidden_dim, 4*hidden_dim))\n",
    "        self.conv1 = GINConv(nn1, train_eps=True)\n",
    "        self.batchnorm1 = nn.BatchNorm1d(4*hidden_dim)\n",
    "        \n",
    "        nn2 = nn.Sequential(Linear(4*hidden_dim,  4*hidden_dim),\n",
    "                            nn.ReLU(),Linear(4*hidden_dim, 4*hidden_dim))\n",
    "        self.conv2 = GINConv(nn2, train_eps=True)\n",
    "        self.batchnorm2 = nn.BatchNorm1d(4*hidden_dim)\n",
    "        \n",
    "        nn3 = nn.Sequential(Linear(4*hidden_dim,  hidden_dim),\n",
    "                            nn.ReLU(),Linear(hidden_dim, hidden_dim))\n",
    "        self.conv3 = GINConv(nn3, train_eps=True)\n",
    "        self.batchnorm3 = nn.BatchNorm1d(hidden_dim)\n",
    "\n",
    "        self.fc1 = Linear(hidden_dim, hidden_dim)\n",
    "        self.batchnorm4 = nn.BatchNorm1d(hidden_dim)\n",
    "\n",
    "        self.fc2 = Linear(hidden_dim, output_dim)\n",
    "\n",
    "        self.dropout = nn.Dropout(p=0.5)\n",
    "\n",
    "    def forward(self, x, edge_index):\n",
    "        x = self.batchnorm1(self.conv1(x, edge_index)).relu()\n",
    "        x = self.batchnorm2(self.conv2(x, edge_index)).relu()\n",
    "        x = self.batchnorm3(self.conv3(x, edge_index)).relu()\n",
    "        x = self.batchnorm4(self.fc1(x)).relu()\n",
    "        x = self.fc2(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pred day 0 test acc: 0.9664 - test f1macro: 0.9220 - test roc: 0.9661\n",
      "pred day 1 test acc: 0.9795 - test f1macro: 0.7550 - test roc: 0.8699\n",
      "pred day 2 test acc: 0.9478 - test f1macro: 0.7540 - test roc: 0.8163\n",
      "pred day 3 test acc: 0.8915 - test f1macro: 0.7165 - test roc: 0.8438\n",
      "pred day 4 test acc: 0.9467 - test f1macro: 0.6861 - test roc: 0.8842\n",
      "pred day 5 test acc: 0.9323 - test f1macro: 0.6973 - test roc: 0.8395\n",
      "pred day 6 test acc: 0.9276 - test f1macro: 0.7335 - test roc: 0.9118\n",
      "pred day 7 test acc: 0.9271 - test f1macro: 0.7487 - test roc: 0.8923\n",
      "pred day 8 test acc: 0.9547 - test f1macro: 0.4884 - test roc: 0.5950\n",
      "pred day 9 test acc: 0.9818 - test f1macro: 0.4954 - test roc: 0.5395\n",
      "pred day 10 test acc: 0.9918 - test f1macro: 0.4979 - test roc: 0.7010\n",
      "pred day 11 test acc: 0.9958 - test f1macro: 0.4989 - test roc: 0.4972\n",
      "pred day 12 test acc: 0.9740 - test f1macro: 0.4934 - test roc: 0.5996\n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.7253\n",
      "pred day 14 test acc: 0.8803 - test f1macro: 0.4682 - test roc: 0.2008\n",
      "final test acc: 0.9525 - test f1macro: 0.7393 - test roc: 0.8259\n",
      "pred day 0 test acc: 0.9553 - test f1macro: 0.9042 - test roc: 0.9742\n",
      "pred day 1 test acc: 0.9660 - test f1macro: 0.6541 - test roc: 0.9275\n",
      "pred day 2 test acc: 0.9116 - test f1macro: 0.6866 - test roc: 0.8766\n",
      "pred day 3 test acc: 0.8743 - test f1macro: 0.7319 - test roc: 0.9010\n",
      "pred day 4 test acc: 0.9290 - test f1macro: 0.6677 - test roc: 0.9256\n",
      "pred day 5 test acc: 0.9273 - test f1macro: 0.7304 - test roc: 0.8724\n",
      "pred day 6 test acc: 0.9002 - test f1macro: 0.6829 - test roc: 0.8749\n",
      "pred day 7 test acc: 0.9383 - test f1macro: 0.8300 - test roc: 0.9191\n",
      "pred day 8 test acc: 0.9029 - test f1macro: 0.4819 - test roc: 0.6199\n",
      "pred day 9 test acc: 0.9554 - test f1macro: 0.4886 - test roc: 0.5496\n",
      "pred day 10 test acc: 0.9672 - test f1macro: 0.4917 - test roc: 0.6951\n",
      "pred day 11 test acc: 0.9775 - test f1macro: 0.4943 - test roc: 0.9472\n",
      "pred day 12 test acc: 0.9598 - test f1macro: 0.4897 - test roc: 0.5749\n",
      "pred day 13 test acc: 0.9087 - test f1macro: 0.4761 - test roc: 0.7503\n",
      "pred day 14 test acc: 0.8403 - test f1macro: 0.4566 - test roc: 0.3139\n",
      "final test acc: 0.9347 - test f1macro: 0.7216 - test roc: 0.8485\n",
      "pred day 0 test acc: 0.9538 - test f1macro: 0.8918 - test roc: 0.9394\n",
      "pred day 1 test acc: 0.9748 - test f1macro: 0.7143 - test roc: 0.8167\n",
      "pred day 2 test acc: 0.9398 - test f1macro: 0.7566 - test roc: 0.8588\n",
      "pred day 3 test acc: 0.8942 - test f1macro: 0.7204 - test roc: 0.8291\n",
      "pred day 4 test acc: 0.9408 - test f1macro: 0.6480 - test roc: 0.7904\n",
      "pred day 5 test acc: 0.9339 - test f1macro: 0.7192 - test roc: 0.7887\n",
      "pred day 6 test acc: 0.9240 - test f1macro: 0.7108 - test roc: 0.7741\n",
      "pred day 7 test acc: 0.9355 - test f1macro: 0.7899 - test roc: 0.9051\n",
      "pred day 8 test acc: 0.9380 - test f1macro: 0.4840 - test roc: 0.5810\n",
      "pred day 9 test acc: 0.9818 - test f1macro: 0.5277 - test roc: 0.4718\n",
      "pred day 10 test acc: 0.9934 - test f1macro: 0.4984 - test roc: 0.3543\n",
      "pred day 11 test acc: 0.9930 - test f1macro: 0.4982 - test roc: 0.3634\n",
      "pred day 12 test acc: 0.9740 - test f1macro: 0.4934 - test roc: 0.5970\n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.4609\n",
      "pred day 14 test acc: 0.8824 - test f1macro: 0.4688 - test roc: 0.7023\n",
      "final test acc: 0.9501 - test f1macro: 0.7349 - test roc: 0.8002\n",
      "pred day 0 test acc: 0.9493 - test f1macro: 0.8837 - test roc: 0.9608\n",
      "pred day 1 test acc: 0.9614 - test f1macro: 0.6727 - test roc: 0.8981\n",
      "pred day 2 test acc: 0.9297 - test f1macro: 0.7346 - test roc: 0.9034\n",
      "pred day 3 test acc: 0.9272 - test f1macro: 0.8509 - test roc: 0.9441\n",
      "pred day 4 test acc: 0.9484 - test f1macro: 0.7604 - test roc: 0.9053\n",
      "pred day 5 test acc: 0.9356 - test f1macro: 0.7610 - test roc: 0.8662\n",
      "pred day 6 test acc: 0.9081 - test f1macro: 0.7068 - test roc: 0.8422\n",
      "pred day 7 test acc: 0.9345 - test f1macro: 0.8044 - test roc: 0.9278\n",
      "pred day 8 test acc: 0.9175 - test f1macro: 0.4785 - test roc: 0.5855\n",
      "pred day 9 test acc: 0.9761 - test f1macro: 0.5190 - test roc: 0.5157\n",
      "pred day 10 test acc: 0.9894 - test f1macro: 0.4973 - test roc: 0.5891\n",
      "pred day 11 test acc: 0.9579 - test f1macro: 0.4892 - test roc: 0.3711\n",
      "pred day 12 test acc: 0.9716 - test f1macro: 0.4928 - test roc: 0.7678\n",
      "pred day 13 test acc: 0.9193 - test f1macro: 0.4790 - test roc: 0.5839\n",
      "pred day 14 test acc: 0.8782 - test f1macro: 0.4842 - test roc: 0.5715\n",
      "final test acc: 0.9447 - test f1macro: 0.7482 - test roc: 0.8559\n",
      "pred day 0 test acc: 0.9508 - test f1macro: 0.8769 - test roc: 0.9626\n",
      "pred day 1 test acc: 0.9760 - test f1macro: 0.6342 - test roc: 0.8285\n",
      "pred day 2 test acc: 0.9438 - test f1macro: 0.7351 - test roc: 0.8538\n",
      "pred day 3 test acc: 0.8889 - test f1macro: 0.6928 - test roc: 0.8949\n",
      "pred day 4 test acc: 0.9408 - test f1macro: 0.6200 - test roc: 0.8194\n",
      "pred day 5 test acc: 0.9199 - test f1macro: 0.6195 - test roc: 0.6940\n",
      "pred day 6 test acc: 0.9108 - test f1macro: 0.6372 - test roc: 0.7597\n",
      "pred day 7 test acc: 0.9183 - test f1macro: 0.7112 - test roc: 0.8605\n",
      "pred day 8 test acc: 0.9774 - test f1macro: 0.4943 - test roc: 0.4545\n",
      "pred day 9 test acc: 0.9824 - test f1macro: 0.4956 - test roc: 0.3845\n",
      "pred day 10 test acc: 0.9869 - test f1macro: 0.4967 - test roc: 0.3416\n",
      "pred day 11 test acc: 0.9888 - test f1macro: 0.4972 - test roc: 0.3944\n",
      "pred day 12 test acc: 0.9740 - test f1macro: 0.4934 - test roc: 0.4879\n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.4334\n",
      "pred day 14 test acc: 0.8803 - test f1macro: 0.4682 - test roc: 0.8015\n",
      "final test acc: 0.9484 - test f1macro: 0.6946 - test roc: 0.8009\n",
      "acc: 0.9461 - f1macro: 0.7277 with var 0.0003 -aucroc: 0.8263 with var 0.0005\n"
     ]
    }
   ],
   "source": [
    "import torch.nn.functional as F\n",
    "from sklearn.metrics import f1_score\n",
    "from torch_geometric.loader import NeighborLoader\n",
    "\n",
    "num_features = 165\n",
    "num_classes = 2\n",
    "REPEAT = 5\n",
    "\n",
    "def train(train_loader, model, optimizer):    \n",
    "    model.train()\n",
    "    total_loss = 0\n",
    "    ys, preds = [], []\n",
    "    for batch in train_loader:\n",
    "        batch = batch.to(device)\n",
    "        labels = batch.y[:batch.batch_size]\n",
    "        labels = labels[(labels<2)]\n",
    "        weight = (labels==0).sum().item() / labels.sum().item()\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        out = model(batch.x, batch.edge_index)[:batch.batch_size]       \n",
    "        loss = F.cross_entropy(out[(batch.y[:batch.batch_size]<2)], labels, weight=torch.tensor([1., weight], device=device))\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        ys.append(labels)\n",
    "        preds.append(torch.softmax(out[(batch.y[:batch.batch_size]<2)], dim=1))\n",
    "        total_loss += loss.item() * batch.batch_size\n",
    "\n",
    "    target_labels = torch.cat(ys, dim=0).detach().cpu().numpy()\n",
    "    pred_scores = torch.cat(preds, dim=0).detach().cpu().numpy()\n",
    "    train_acc, train_f1macro, train_aucroc = store_metrics(pred_scores, target_labels)\n",
    "\n",
    "    return total_loss / len(train_loader.dataset), train_acc, train_f1macro, train_aucroc\n",
    "\n",
    "\n",
    "@torch.no_grad()\n",
    "def test(loader, model):\n",
    "    model.eval()\n",
    "    ys, preds = [], []\n",
    "    daily_acc, daily_f1macro, daily_aucroc = [], [], []\n",
    "    for day, data in enumerate(loader):\n",
    "        out = model(data.x.to(device), data.edge_index.to(device))\n",
    "        target_labels = data.y[data.labeled_mask].detach().cpu().numpy()\n",
    "        pred_scores = torch.softmax(out[data.labeled_mask].squeeze(), dim=1).detach().cpu().numpy()\n",
    "\n",
    "        ys.append(target_labels)\n",
    "        preds.append(pred_scores)\n",
    "        test_acc, test_f1macro, test_aucroc= store_metrics(pred_scores, target_labels)\n",
    "\n",
    "        print(\"pred day {:} test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f}\".format(\n",
    "            day, test_acc, test_f1macro, test_aucroc))\n",
    "        daily_acc.append(test_acc)\n",
    "        daily_f1macro.append(test_f1macro)\n",
    "        daily_aucroc.append(test_aucroc)\n",
    "\n",
    "\n",
    "    target_labels = np.concatenate(ys, axis=0) \n",
    "    pred_scores = np.concatenate(preds, axis=0) \n",
    "    test_acc, test_f1macro, test_aucroc= store_metrics(pred_scores, target_labels)\n",
    "    print(\"final test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f}\".format(test_acc, test_f1macro, test_aucroc))\n",
    "\n",
    "    return test_acc, test_f1macro, test_aucroc, daily_acc, daily_f1macro, daily_aucroc\n",
    "\n",
    "results_full = {\n",
    "    'acc': {},\n",
    "    'f1macro': {},\n",
    "    'f1micro': {},\n",
    "    'aucroc': {}\n",
    "}\n",
    "\n",
    "results_full = {\n",
    "    'acc' : np.zeros((16, REPEAT)),\n",
    "    'f1macro' : np.zeros((16, REPEAT)),\n",
    "    'aucroc' : np.zeros((16, REPEAT))\n",
    "}\n",
    "\n",
    "for repeat in range(REPEAT):\n",
    "    model = GINNet(num_features, num_classes).to(device)\n",
    "    optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=1e-5)\n",
    "\n",
    "    for epoch in range(1, 201):\n",
    "        loss, train_acc, train_f1macro, train_aucroc = train(train_loader, model, optimizer)\n",
    "\n",
    "    test_acc, test_f1macro, test_aucroc, daily_acc, daily_f1macro, daily_aucroc= test(test_loader, model)\n",
    "\n",
    "    results_full['acc'][:15,repeat] = np.array(daily_acc)\n",
    "    results_full['f1macro'][:15,repeat] = np.array(daily_f1macro)\n",
    "    results_full['aucroc'][:15,repeat] = np.array(daily_aucroc)\n",
    "    results_full['acc'][15,repeat] = test_acc\n",
    "    results_full['f1macro'][15,repeat] = test_f1macro\n",
    "    results_full['aucroc'][15,repeat] = test_aucroc\n",
    "\n",
    "print(\"acc: {:.4f} - f1macro: {:.4f} with var {:.4f} -aucroc: {:.4f} with var {:.4f}\".format(\n",
    "                                            np.asarray(results_full['acc'][15,:]).mean(), \n",
    "                                            np.asarray(results_full['f1macro'][15,:]).mean(), \n",
    "                                            np.asarray(results_full['f1macro'][15,:]).var(), \n",
    "                                            np.asarray(results_full['aucroc'][15,:]).mean(), \n",
    "                                            np.asarray(results_full['aucroc'][15,:]).var()\n",
    "                                                ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "method:Fast, running time:2.6492013931274414, balance:-0.022588832487309644, fairness:-0.02784466404446429\n",
      "method:Fast, running time:0.7876403331756592, balance:-0.030281691484048332, fairness:-0.03441527466679974\n",
      "method:Fast, running time:1.626281499862671, balance:-0.0221265669838393, fairness:-0.026063615053586627\n",
      "method:Fast, running time:1.3226237297058105, balance:-0.029834886018534276, fairness:-0.025382574670866086\n",
      "method:Fast, running time:1.6924829483032227, balance:-0.02491547846538292, fairness:-0.027027371620224984\n",
      "method:Fast, running time:0.7259957790374756, balance:-0.030499075785582256, fairness:-0.02294045189340277\n",
      "method:Fast, running time:1.4213600158691406, balance:-0.03356481481481482, fairness:-0.02428206754243715\n",
      "method:Fast, running time:0.7750153541564941, balance:-0.027294142673655906, fairness:-0.039143345045793344\n",
      "method:Fast, running time:0.9760699272155762, balance:-0.028983187160079627, fairness:-0.027722462180385897\n",
      "method:Fast, running time:1.7558667659759521, balance:-0.02391110541155279, fairness:-0.018182937776974793\n",
      "method:Fast, running time:0.7200288772583008, balance:-0.03505586734253157, fairness:-0.03366009139035499\n",
      "method:Fast, running time:0.22252750396728516, balance:-0.053492916463116755, fairness:-0.03835194205705762\n",
      "method:Fast, running time:0.800213098526001, balance:-0.042181978798586574, fairness:-0.027348116601505112\n",
      "method:Fast, running time:0.2133619785308838, balance:-0.05489614243323442, fairness:-0.04042803732553192\n",
      "method:Fast, running time:0.5485410690307617, balance:-0.041371256678141746, fairness:-0.037359596201509575\n",
      "method:Fast, running time:0.3988509178161621, balance:-0.046638655462184875, fairness:-0.03321815856195302\n",
      "method:Fast, running time:0.4859902858734131, balance:-0.04290989660265879, fairness:-0.03341005551660804\n",
      "method:Fast, running time:0.20663166046142578, balance:-0.06578947368421052, fairness:-0.039842672948085786\n",
      "method:Fast, running time:0.509427547454834, balance:-0.02330290842790707, fairness:-0.02956223250040585\n",
      "method:Fast, running time:0.7261567115783691, balance:-0.026322534828301606, fairness:-0.030594443973432372\n",
      "method:Fast, running time:0.5199496746063232, balance:-0.046268022023320435, fairness:-0.032455925305568144\n",
      "method:Fast, running time:1.3481640815734863, balance:-0.042297249371050005, fairness:-0.0339372461091786\n",
      "method:Fast, running time:0.6035714149475098, balance:-0.030612244897959183, fairness:-0.0366187496705425\n",
      "method:Fast, running time:0.8921942710876465, balance:-0.03314460196146151, fairness:-0.029850925331031945\n",
      "method:Fast, running time:0.265047550201416, balance:-0.05799481944988386, fairness:-0.0445431565112828\n",
      "method:Fast, running time:0.2942066192626953, balance:-0.042489099067872446, fairness:-0.04956197400813264\n",
      "method:Fast, running time:0.12097787857055664, balance:-0.09361799536110393, fairness:-0.06618673435637903\n",
      "method:Fast, running time:0.17197036743164062, balance:-0.073502722323049, fairness:-0.058696516938639906\n",
      "method:Fast, running time:0.7105128765106201, balance:-0.030058479532163743, fairness:-0.04186449112930842\n",
      "method:Fast, running time:0.2877960205078125, balance:-0.031453888885829895, fairness:-0.03842825638227012\n"
     ]
    }
   ],
   "source": [
    "from GUIDE.guide import GUIDE\n",
    "args = {}\n",
    "args['dataname'] = 'BTC'\n",
    "args['shards_number'] = 20\n",
    "args['device'] = device\n",
    "partations_fast = []\n",
    "for gi in range(len(dataset_train)):\n",
    "    args['gi'] = gi    \n",
    "    GUIDE_Fast = GUIDE(edge_indexs=dataset_train[gi].edge_index, labels=dataset_train[gi].y, k=args['shards_number'])\n",
    "    p1 = GUIDE_Fast.fit(method='Fast', alpha_=1e-3)\n",
    "    p1_saved = GUIDE_Fast.subgraph_repair(x = dataset_train[gi].x, REPAIR_METHOD='MixUp', PATH='checkpoints/',DATA_NAME='{}'.format(args['dataname']), MULTI_GRAPH = args['gi'])\n",
    "    savename = p1_saved.DPATH.split('partid')[0]+p1_saved.method+'_saved.pt'\n",
    "    torch.save(p1_saved, savename)\n",
    "    partations_fast.append(p1_saved)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from GUIDE.kernel_vector import PyramidMatchVector\n",
    "\n",
    "gis_keys_graph = {}\n",
    "pm_kernel = PyramidMatchVector()\n",
    "\n",
    "for gi in range(len(dataset_train)):\n",
    "    gis_keys_graph[gi] = {}\n",
    "    for part_id in range(args['shards_number']):\n",
    "        loadname = partations_fast[gi].DPATH.replace('partid', 'part'+str(part_id))\n",
    "        sub_graph = torch.load(loadname)\n",
    "        gis_keys_graph[gi][part_id] = sub_graph\n",
    "        # kernel feature extraction for further use\n",
    "        subg_kfea = pm_kernel.parse_input([sub_graph.edge_index.cpu()])\n",
    "        savekfean = loadname.replace('subgraphs', 'subgkfeas').replace('.pt', '_PM'+'.pt')\n",
    "        os.makedirs(os.path.dirname(savekfean), exist_ok=True)\n",
    "        torch.save(subg_kfea, savekfean)\n",
    "        partations_fast[gi].FPATH = savekfean.replace('part'+str(part_id), 'partid')\n",
    "        # model save path\n",
    "        savemodeln = loadname.replace('subgraphs', 'submodels').replace('.pt', '_'+ 'GINNet' +'.pt').replace('part'+str(part_id), 'partid')\n",
    "        savemodeln = savemodeln.split('graph')[0] + 'model_part' + savemodeln.split('_part')[1]\n",
    "        partations_fast[gi].MPATH = savemodeln.replace('part'+str(part_id), 'partid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pred day 0 test acc: 0.9493 - test f1macro: 0.8866 - test roc: 0.9588 \n",
      "pred day 1 test acc: 0.9690 - test f1macro: 0.6649 - test roc: 0.8888 \n",
      "pred day 2 test acc: 0.9438 - test f1macro: 0.7904 - test roc: 0.8826 \n",
      "pred day 3 test acc: 0.9034 - test f1macro: 0.8148 - test roc: 0.9019 \n",
      "pred day 4 test acc: 0.9163 - test f1macro: 0.6846 - test roc: 0.9147 \n",
      "pred day 5 test acc: 0.9207 - test f1macro: 0.7119 - test roc: 0.8345 \n",
      "pred day 6 test acc: 0.9072 - test f1macro: 0.7367 - test roc: 0.8185 \n",
      "pred day 7 test acc: 0.9401 - test f1macro: 0.8445 - test roc: 0.9447 \n",
      "pred day 8 test acc: 0.9642 - test f1macro: 0.4909 - test roc: 0.6000 \n",
      "pred day 9 test acc: 0.8605 - test f1macro: 0.4625 - test roc: 0.5510 \n",
      "pred day 10 test acc: 0.9525 - test f1macro: 0.4878 - test roc: 0.6219 \n",
      "pred day 11 test acc: 0.9621 - test f1macro: 0.4903 - test roc: 0.8577 \n",
      "pred day 12 test acc: 0.9622 - test f1macro: 0.4904 - test roc: 0.6551 \n",
      "pred day 13 test acc: 0.9214 - test f1macro: 0.4796 - test roc: 0.7888 \n",
      "pred day 14 test acc: 0.8508 - test f1macro: 0.4597 - test roc: 0.7677 \n",
      "final test acc: 0.9312 - test f1macro: 0.7279 - test roc: 0.8660\n",
      "pred day 0 test acc: 0.9530 - test f1macro: 0.8885 - test roc: 0.9547 \n",
      "pred day 1 test acc: 0.9713 - test f1macro: 0.6571 - test roc: 0.8663 \n",
      "pred day 2 test acc: 0.9177 - test f1macro: 0.6891 - test roc: 0.8505 \n",
      "pred day 3 test acc: 0.8955 - test f1macro: 0.7875 - test roc: 0.9015 \n",
      "pred day 4 test acc: 0.9205 - test f1macro: 0.6479 - test roc: 0.9069 \n",
      "pred day 5 test acc: 0.9265 - test f1macro: 0.7200 - test roc: 0.7983 \n",
      "pred day 6 test acc: 0.9231 - test f1macro: 0.7687 - test roc: 0.8327 \n",
      "pred day 7 test acc: 0.9457 - test f1macro: 0.8574 - test roc: 0.9457 \n",
      "pred day 8 test acc: 0.9693 - test f1macro: 0.4922 - test roc: 0.5928 \n",
      "pred day 9 test acc: 0.8636 - test f1macro: 0.4634 - test roc: 0.5565 \n",
      "pred day 10 test acc: 0.9730 - test f1macro: 0.4932 - test roc: 0.6456 \n",
      "pred day 11 test acc: 0.9551 - test f1macro: 0.4885 - test roc: 0.8232 \n",
      "pred day 12 test acc: 0.9681 - test f1macro: 0.4919 - test roc: 0.6226 \n",
      "pred day 13 test acc: 0.9214 - test f1macro: 0.4796 - test roc: 0.8440 \n",
      "pred day 14 test acc: 0.8571 - test f1macro: 0.4615 - test roc: 0.6972 \n",
      "final test acc: 0.9355 - test f1macro: 0.7289 - test roc: 0.8589\n",
      "pred day 0 test acc: 0.9553 - test f1macro: 0.8990 - test roc: 0.9598 \n",
      "pred day 1 test acc: 0.9614 - test f1macro: 0.6234 - test roc: 0.8810 \n",
      "pred day 2 test acc: 0.9398 - test f1macro: 0.7495 - test roc: 0.8758 \n",
      "pred day 3 test acc: 0.9048 - test f1macro: 0.8205 - test roc: 0.9096 \n",
      "pred day 4 test acc: 0.9087 - test f1macro: 0.6462 - test roc: 0.9059 \n",
      "pred day 5 test acc: 0.9158 - test f1macro: 0.7000 - test roc: 0.8326 \n",
      "pred day 6 test acc: 0.9178 - test f1macro: 0.7610 - test roc: 0.8227 \n",
      "pred day 7 test acc: 0.9457 - test f1macro: 0.8546 - test roc: 0.9487 \n",
      "pred day 8 test acc: 0.9664 - test f1macro: 0.4915 - test roc: 0.6353 \n",
      "pred day 9 test acc: 0.8605 - test f1macro: 0.4625 - test roc: 0.5200 \n",
      "pred day 10 test acc: 0.9509 - test f1macro: 0.4874 - test roc: 0.5502 \n",
      "pred day 11 test acc: 0.9593 - test f1macro: 0.4896 - test roc: 0.7549 \n",
      "pred day 12 test acc: 0.9645 - test f1macro: 0.4910 - test roc: 0.7084 \n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.8273 \n",
      "pred day 14 test acc: 0.8592 - test f1macro: 0.4621 - test roc: 0.7517 \n",
      "final test acc: 0.9317 - test f1macro: 0.7270 - test roc: 0.8673\n",
      "pred day 0 test acc: 0.9545 - test f1macro: 0.8903 - test roc: 0.9561 \n",
      "pred day 1 test acc: 0.9701 - test f1macro: 0.6332 - test roc: 0.8769 \n",
      "pred day 2 test acc: 0.9418 - test f1macro: 0.7468 - test roc: 0.8557 \n",
      "pred day 3 test acc: 0.8783 - test f1macro: 0.7071 - test roc: 0.8587 \n",
      "pred day 4 test acc: 0.9307 - test f1macro: 0.6663 - test roc: 0.9124 \n",
      "pred day 5 test acc: 0.9249 - test f1macro: 0.6826 - test roc: 0.8308 \n",
      "pred day 6 test acc: 0.9284 - test f1macro: 0.7465 - test roc: 0.8421 \n",
      "pred day 7 test acc: 0.9540 - test f1macro: 0.8657 - test roc: 0.9423 \n",
      "pred day 8 test acc: 0.9745 - test f1macro: 0.4935 - test roc: 0.6364 \n",
      "pred day 9 test acc: 0.9434 - test f1macro: 0.4854 - test roc: 0.5442 \n",
      "pred day 10 test acc: 0.9787 - test f1macro: 0.4946 - test roc: 0.6400 \n",
      "pred day 11 test acc: 0.9817 - test f1macro: 0.4954 - test roc: 0.7725 \n",
      "pred day 12 test acc: 0.9704 - test f1macro: 0.4925 - test roc: 0.6650 \n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.8666 \n",
      "pred day 14 test acc: 0.8739 - test f1macro: 0.4664 - test roc: 0.6864 \n",
      "final test acc: 0.9478 - test f1macro: 0.7406 - test roc: 0.8629\n",
      "pred day 0 test acc: 0.9575 - test f1macro: 0.8975 - test roc: 0.9553 \n",
      "pred day 1 test acc: 0.9789 - test f1macro: 0.6946 - test roc: 0.9195 \n",
      "pred day 2 test acc: 0.9438 - test f1macro: 0.7517 - test roc: 0.8596 \n",
      "pred day 3 test acc: 0.9220 - test f1macro: 0.8307 - test roc: 0.9134 \n",
      "pred day 4 test acc: 0.9290 - test f1macro: 0.6677 - test roc: 0.9084 \n",
      "pred day 5 test acc: 0.9265 - test f1macro: 0.7005 - test roc: 0.8333 \n",
      "pred day 6 test acc: 0.9231 - test f1macro: 0.7278 - test roc: 0.7993 \n",
      "pred day 7 test acc: 0.9503 - test f1macro: 0.8586 - test roc: 0.9351 \n",
      "pred day 8 test acc: 0.9745 - test f1macro: 0.4935 - test roc: 0.5729 \n",
      "pred day 9 test acc: 0.9202 - test f1macro: 0.4792 - test roc: 0.5327 \n",
      "pred day 10 test acc: 0.9836 - test f1macro: 0.4959 - test roc: 0.6604 \n",
      "pred day 11 test acc: 0.9831 - test f1macro: 0.4958 - test roc: 0.7035 \n",
      "pred day 12 test acc: 0.9669 - test f1macro: 0.4916 - test roc: 0.6403 \n",
      "pred day 13 test acc: 0.9214 - test f1macro: 0.4796 - test roc: 0.7961 \n",
      "pred day 14 test acc: 0.8676 - test f1macro: 0.4646 - test roc: 0.6386 \n",
      "final test acc: 0.9479 - test f1macro: 0.7504 - test roc: 0.8532\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import torch.nn.functional as F\n",
    "\n",
    "args['num_features'] = num_features\n",
    "args['num_classes'] = num_classes\n",
    "results_Fast = {\n",
    "    'acc' : np.zeros((16, REPEAT)),\n",
    "    'f1macro' : np.zeros((16, REPEAT)),\n",
    "    'aucroc' : np.zeros((16, REPEAT))\n",
    "}\n",
    "\n",
    "for repeat in range(REPEAT):\n",
    "    for part_id in p1_saved.shards_ids.keys():\n",
    "        # submodel training\n",
    "        start = time.time()\n",
    "        submodel = GINNet(input_dim=args['num_features'],\n",
    "                    output_dim=args['num_classes']).to(args['device'])\n",
    "\n",
    "        optimizer = torch.optim.Adam(\n",
    "            submodel.parameters(), lr=0.01, weight_decay=1e-5)\n",
    "        for epoch in range(1, 201):\n",
    "            submodel.train()\n",
    "            for gi in range(len(dataset_train)):\n",
    "                sub_graph = gis_keys_graph[gi][part_id]\n",
    "                sub_graph = sub_graph.to(device)\n",
    "                labels = sub_graph.y[sub_graph.y < 2]\n",
    "                if labels.sum() > 0:\n",
    "                    weight = (labels == 0).sum().item() / labels.sum().item()\n",
    "                else:\n",
    "                    weight = 1.0\n",
    "                optimizer.zero_grad()\n",
    "                out = submodel(sub_graph.x, sub_graph.edge_index)[\n",
    "                    sub_graph.train_mask]\n",
    "                loss = F.cross_entropy(out[torch.where(sub_graph.y < 2)[\n",
    "                    0]], labels, weight=torch.tensor([1., weight], device=device))\n",
    "\n",
    "                loss.backward()\n",
    "                optimizer.step()\n",
    "\n",
    "        submodel.to('cpu')\n",
    "        sub_graph = sub_graph.to('cpu')\n",
    "        savemodeln = partations_fast[gi].MPATH.replace('partid', 'part'+str(part_id))\n",
    "        os.makedirs(os.path.dirname(savemodeln), exist_ok=True)\n",
    "        torch.save(submodel.state_dict(), savemodeln)\n",
    "        del submodel, sub_graph, optimizer, out, loss\n",
    "\n",
    "    preds, targets = [], []\n",
    "    for day, data in enumerate(test_loader):\n",
    "        pm_kernel = PyramidMatchVector()\n",
    "        subgt_kfea = pm_kernel.parse_input([data.edge_index.cpu()])\n",
    "        subg_kfea = []\n",
    "        subouts = []\n",
    "        for part_id in p1_saved.shards_ids.keys():\n",
    "            for gi in range(len(dataset_train)):\n",
    "                loadname = partations_fast[gi].FPATH.replace('partid', 'part'+str(part_id))\n",
    "                subg_kfea += torch.load(loadname)\n",
    "\n",
    "            submodel = GINNet(input_dim=args['num_features'],\n",
    "                        output_dim=args['num_classes']).to(args['device'])\n",
    "            savemodeln = partations_fast[gi].MPATH.replace('partid', 'part'+str(part_id))\n",
    "            submodel.load_state_dict(torch.load(savemodeln))\n",
    "            submodel.eval()\n",
    "            subout = torch.softmax(submodel(data.x.to(args['device']), data.edge_index.to(args['device'])), dim=1)\n",
    "            if 'labeled_mask' in data:\n",
    "                subout = subout[data.labeled_mask]\n",
    "\n",
    "            subouts.append(subout.detach().to('cpu'))\n",
    "            submodel.to('cpu')\n",
    "            del submodel, subout\n",
    "\n",
    "        suboutTensor = torch.stack(subouts)\n",
    "\n",
    "        daily_weights = pm_kernel.kernel_similarity(subgt_kfea, subg_kfea)\n",
    "        daily_weights = torch.tensor(daily_weights).to(torch.float32).reshape(len(p1_saved.shards_ids.keys()), -1).mean(dim=-1)\n",
    "        saveweightn = partations_fast[gi].MPATH.replace('partid/submodels/','').replace('model_partid', f'day{day}_weight')\n",
    "        os.makedirs(os.path.dirname(saveweightn), exist_ok=True)\n",
    "        torch.save(daily_weights, saveweightn)\n",
    "\n",
    "        weights = F.softmax(daily_weights, dim=0)\n",
    "        weighted_pred = torch.tensordot(suboutTensor, weights, dims=([0], [0]))\n",
    "\n",
    "        if 'labeled_mask' in data:\n",
    "            target_labels = data.y[data.labeled_mask].detach().cpu().numpy()\n",
    "        else:\n",
    "            target_labels = data.y.detach().cpu().numpy()\n",
    "        \n",
    "        test_acc, test_f1macro, test_aucroc= store_metrics(weighted_pred.detach().cpu().numpy(), target_labels)\n",
    "        print(\"pred day {} test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f} \".format(day,\n",
    "            test_acc, test_f1macro, test_aucroc))\n",
    "        results_Fast['acc'][day, repeat] = test_acc\n",
    "        results_Fast['f1macro'][day, repeat] = test_f1macro\n",
    "        results_Fast['aucroc'][day, repeat] = test_aucroc\n",
    "        preds.append(weighted_pred.detach().cpu().numpy())\n",
    "        targets.append(target_labels)\n",
    "\n",
    "    pred_all = np.concatenate(preds, axis=0)\n",
    "    targets_all = np.concatenate(targets, axis=0)\n",
    "    test_acc, test_f1macro, test_aucroc= store_metrics(pred_all, targets_all)\n",
    "    print(\"final test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f}\".format(test_acc, test_f1macro, test_aucroc))\n",
    "    results_Fast['acc'][-1, repeat] = test_acc\n",
    "    results_Fast['f1macro'][-1, repeat] = test_f1macro\n",
    "    results_Fast['aucroc'][-1, repeat] = test_aucroc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.04010152284263959 0.03331892601423409 -0.05\n",
      "method:SR, running time:6.143684387207031, balance:-0.04010152284263959, fairness:-0.03331892601423409\n",
      "0.053036971830985914 0.03348032911508182 -0.05\n",
      "method:SR, running time:2.7043659687042236, balance:-0.053036971830985914, fairness:-0.03348032911508182\n",
      "0.05354176106328349 0.03564643422207327 -0.05\n",
      "method:SR, running time:4.2195470333099365, balance:-0.05354176106328349, fairness:-0.03564643422207327\n",
      "0.05289829508106172 0.031452222527289624 -0.05\n",
      "method:SR, running time:3.734086513519287, balance:-0.05289829508106172, fairness:-0.031452222527289624\n",
      "0.061245038056222896 0.02726945598790409 -0.05\n",
      "method:SR, running time:4.745819807052612, balance:-0.061245038056222896, fairness:-0.02726945598790409\n",
      "0.044454712083291215 0.02646041691012944 -0.05\n",
      "method:SR, running time:2.6471478939056396, balance:-0.044454712083291215, fairness:-0.02646041691012944\n",
      "0.05406746031746032 0.03189369192743313 -0.05\n",
      "method:SR, running time:3.806093692779541, balance:-0.05406746031746032, fairness:-0.03189369192743313\n",
      "0.04495176264382222 0.02906529790750474 -0.05\n",
      "method:SR, running time:2.707611322402954, balance:-0.04495176264382222, fairness:-0.02906529790750474\n",
      "0.06044835868694956 0.02797354565787923 -0.05\n",
      "method:SR, running time:3.2162246704101562, balance:-0.06044835868694956, fairness:-0.02797354565787923\n",
      "0.05338932568700526 0.028283278456222243 -0.05\n",
      "method:SR, running time:4.75530219078064, balance:-0.05338932568700526, fairness:-0.028283278456222243\n",
      "0.04771880819366853 0.02041189659276436 -0.05\n",
      "method:SR, running time:2.3707048892974854, balance:-0.04771880819366853, fairness:-0.02041189659276436\n",
      "0.04550561872295013 0.04042613103948078 -0.05\n",
      "method:SR, running time:0.8710448741912842, balance:-0.04550561872295013, fairness:-0.04042613103948078\n",
      "0.03727914924756377 0.025351141152698597 -0.05\n",
      "method:SR, running time:3.064105272293091, balance:-0.03727914924756377, fairness:-0.025351141152698597\n",
      "0.05336300767847622 0.04349917968617312 -0.05\n",
      "method:SR, running time:0.9874231815338135, balance:-0.05336300767847622, fairness:-0.04349917968617312\n",
      "0.04381698184892613 0.027492389874567962 -0.05\n",
      "method:SR, running time:2.140104293823242, balance:-0.04381698184892613, fairness:-0.027492389874567962\n",
      "0.04630252100840336 0.03332176749350232 -0.05\n",
      "method:SR, running time:1.3128516674041748, balance:-0.04630252100840336, fairness:-0.03332176749350232\n",
      "0.048005908419497784 0.035236008163709656 -0.05\n",
      "method:SR, running time:1.7671051025390625, balance:-0.048005908419497784, fairness:-0.035236008163709656\n",
      "0.041396762678015085 0.043545501326195904 -0.05\n",
      "method:SR, running time:0.8356344699859619, balance:-0.041396762678015085, fairness:-0.043545501326195904\n",
      "0.0398174575309514 0.02690429577296314 -0.05\n",
      "method:SR, running time:1.638930082321167, balance:-0.0398174575309514, fairness:-0.02690429577296314\n",
      "0.043486368236661764 0.036416922619090644 -0.05\n",
      "method:SR, running time:2.636406183242798, balance:-0.043486368236661764, fairness:-0.036416922619090644\n",
      "0.050989540883097435 0.02792793519023809 -0.05\n",
      "method:SR, running time:1.8842430114746094, balance:-0.050989540883097435, fairness:-0.02792793519023809\n",
      "0.03965049409688348 0.03626229073580221 -0.05\n",
      "method:SR, running time:3.871070146560669, balance:-0.03965049409688348, fairness:-0.03626229073580221\n",
      "0.040096038415366145 0.04363067590902814 -0.05\n",
      "method:SR, running time:2.263397216796875, balance:-0.040096038415366145, fairness:-0.04363067590902814\n",
      "0.03876306620209059 0.031142665404397817 -0.05\n",
      "method:SR, running time:2.841122627258301, balance:-0.03876306620209059, fairness:-0.031142665404397817\n",
      "0.03128781067263802 0.04240402256458938 -0.05\n",
      "method:SR, running time:1.4145927429199219, balance:-0.03128781067263802, fairness:-0.04240402256458938\n",
      "0.04623464190482715 0.04253469419122125 -0.05\n",
      "method:SR, running time:1.245065450668335, balance:-0.04623464190482715, fairness:-0.04253469419122125\n",
      "0.05463728191000918 0.05565683334271968 -0.05\n",
      "method:SR, running time:0.45393872261047363, balance:-0.05463728191000918, fairness:-0.05565683334271968\n",
      "0.04748941318814277 0.04498814947789695 -0.05\n",
      "method:SR, running time:0.5908341407775879, balance:-0.04748941318814277, fairness:-0.04498814947789695\n",
      "0.055321637426900584 0.0570874895884423 -0.05\n",
      "method:SR, running time:2.3649306297302246, balance:-0.055321637426900584, fairness:-0.0570874895884423\n",
      "0.045308095046314945 0.04332583281755857 -0.05\n",
      "method:SR, running time:1.2342772483825684, balance:-0.045308095046314945, fairness:-0.04332583281755857\n"
     ]
    }
   ],
   "source": [
    "from GUIDE.guide import GUIDE\n",
    "args = {}\n",
    "args['dataname'] = 'BTC'\n",
    "args['shards_number'] = 20\n",
    "args['device'] = device\n",
    "partations_sr = []\n",
    "for gi in range(len(dataset_train)):\n",
    "    args['gi'] = gi    \n",
    "    GUIDE_SR = GUIDE(edge_indexs=dataset_train[gi].edge_index, labels=dataset_train[gi].y, k=args['shards_number'])\n",
    "    p1 = GUIDE_SR.fit(method='SR', alpha_=1e-3, beta_=1e-2)\n",
    "    p1_saved = GUIDE_SR.subgraph_repair(x = dataset_train[gi].x, REPAIR_METHOD='MixUp', PATH='checkpoints/',DATA_NAME='{}'.format(args['dataname']), MULTI_GRAPH = args['gi'])\n",
    "    savename = p1_saved.DPATH.split('partid')[0]+p1_saved.method+'_saved.pt'\n",
    "    torch.save(p1_saved, savename)\n",
    "    partations_sr.append(p1_saved)\n",
    "\n",
    "from GUIDE.kernel_vector import PyramidMatchVector\n",
    "\n",
    "gis_keys_graph = {}\n",
    "pm_kernel = PyramidMatchVector()\n",
    "\n",
    "for gi in range(len(dataset_train)):\n",
    "    gis_keys_graph[gi] = {}\n",
    "    for part_id in range(args['shards_number']):\n",
    "        loadname = partations_sr[gi].DPATH.replace('partid', 'part'+str(part_id))\n",
    "        sub_graph = torch.load(loadname)\n",
    "        gis_keys_graph[gi][part_id] = sub_graph\n",
    "        # kernel feature extraction for further use\n",
    "        subg_kfea = pm_kernel.parse_input([sub_graph.edge_index.cpu()])\n",
    "        savekfean = loadname.replace('subgraphs', 'subgkfeas').replace('.pt', '_PM'+'.pt')\n",
    "        os.makedirs(os.path.dirname(savekfean), exist_ok=True)\n",
    "        torch.save(subg_kfea, savekfean)\n",
    "        partations_sr[gi].FPATH = savekfean.replace('part'+str(part_id), 'partid')\n",
    "        # model save path\n",
    "        savemodeln = loadname.replace('subgraphs', 'submodels').replace('.pt', '_'+ 'GINNet' +'.pt').replace('part'+str(part_id), 'partid')\n",
    "        savemodeln = savemodeln.split('graph')[0] + 'model_part' + savemodeln.split('_part')[1]\n",
    "        partations_sr[gi].MPATH = savemodeln.replace('part'+str(part_id), 'partid')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pred day 0 test acc: 0.9545 - test f1macro: 0.8943 - test roc: 0.9477 \n",
      "pred day 1 test acc: 0.9754 - test f1macro: 0.6755 - test roc: 0.8349 \n",
      "pred day 2 test acc: 0.9378 - test f1macro: 0.7449 - test roc: 0.8676 \n",
      "pred day 3 test acc: 0.8942 - test f1macro: 0.7695 - test roc: 0.9108 \n",
      "pred day 4 test acc: 0.9239 - test f1macro: 0.6489 - test roc: 0.8701 \n",
      "pred day 5 test acc: 0.9273 - test f1macro: 0.7089 - test roc: 0.8213 \n",
      "pred day 6 test acc: 0.9311 - test f1macro: 0.7622 - test roc: 0.8089 \n",
      "pred day 7 test acc: 0.9540 - test f1macro: 0.8669 - test roc: 0.9456 \n",
      "pred day 8 test acc: 0.9657 - test f1macro: 0.4913 - test roc: 0.5456 \n",
      "pred day 9 test acc: 0.9208 - test f1macro: 0.4794 - test roc: 0.5058 \n",
      "pred day 10 test acc: 0.9730 - test f1macro: 0.4932 - test roc: 0.5883 \n",
      "pred day 11 test acc: 0.9831 - test f1macro: 0.4958 - test roc: 0.8430 \n",
      "pred day 12 test acc: 0.9704 - test f1macro: 0.4925 - test roc: 0.7226 \n",
      "pred day 13 test acc: 0.9236 - test f1macro: 0.4801 - test roc: 0.8709 \n",
      "pred day 14 test acc: 0.8782 - test f1macro: 0.4676 - test roc: 0.6351 \n",
      "final test acc: 0.9457 - test f1macro: 0.7452 - test roc: 0.8479\n",
      "pred day 0 test acc: 0.9575 - test f1macro: 0.9038 - test roc: 0.9528 \n",
      "pred day 1 test acc: 0.9684 - test f1macro: 0.6544 - test roc: 0.8287 \n",
      "pred day 2 test acc: 0.9337 - test f1macro: 0.7498 - test roc: 0.8631 \n",
      "pred day 3 test acc: 0.8981 - test f1macro: 0.7944 - test roc: 0.9087 \n",
      "pred day 4 test acc: 0.9062 - test f1macro: 0.6539 - test roc: 0.8575 \n",
      "pred day 5 test acc: 0.9166 - test f1macro: 0.6954 - test roc: 0.8221 \n",
      "pred day 6 test acc: 0.9161 - test f1macro: 0.7637 - test roc: 0.8450 \n",
      "pred day 7 test acc: 0.9392 - test f1macro: 0.8433 - test roc: 0.9413 \n",
      "pred day 8 test acc: 0.9708 - test f1macro: 0.4926 - test roc: 0.5617 \n",
      "pred day 9 test acc: 0.7976 - test f1macro: 0.4468 - test roc: 0.4811 \n",
      "pred day 10 test acc: 0.9484 - test f1macro: 0.4868 - test roc: 0.5584 \n",
      "pred day 11 test acc: 0.9396 - test f1macro: 0.4844 - test roc: 0.8880 \n",
      "pred day 12 test acc: 0.9610 - test f1macro: 0.4901 - test roc: 0.7566 \n",
      "pred day 13 test acc: 0.9193 - test f1macro: 0.4790 - test roc: 0.7487 \n",
      "pred day 14 test acc: 0.8424 - test f1macro: 0.4572 - test roc: 0.6035 \n",
      "final test acc: 0.9236 - test f1macro: 0.7108 - test roc: 0.8448\n",
      "pred day 0 test acc: 0.9560 - test f1macro: 0.9014 - test roc: 0.9529 \n",
      "pred day 1 test acc: 0.9713 - test f1macro: 0.6571 - test roc: 0.8820 \n",
      "pred day 2 test acc: 0.9297 - test f1macro: 0.7120 - test roc: 0.8701 \n",
      "pred day 3 test acc: 0.8770 - test f1macro: 0.7162 - test roc: 0.8972 \n",
      "pred day 4 test acc: 0.9121 - test f1macro: 0.6388 - test roc: 0.8776 \n",
      "pred day 5 test acc: 0.9166 - test f1macro: 0.6985 - test roc: 0.8132 \n",
      "pred day 6 test acc: 0.9134 - test f1macro: 0.7428 - test roc: 0.8458 \n",
      "pred day 7 test acc: 0.9485 - test f1macro: 0.8620 - test roc: 0.9478 \n",
      "pred day 8 test acc: 0.9693 - test f1macro: 0.4922 - test roc: 0.6505 \n",
      "pred day 9 test acc: 0.8133 - test f1macro: 0.4485 - test roc: 0.5115 \n",
      "pred day 10 test acc: 0.9533 - test f1macro: 0.4881 - test roc: 0.4849 \n",
      "pred day 11 test acc: 0.9551 - test f1macro: 0.4885 - test roc: 0.8930 \n",
      "pred day 12 test acc: 0.9574 - test f1macro: 0.4891 - test roc: 0.6475 \n",
      "pred day 13 test acc: 0.9151 - test f1macro: 0.4778 - test roc: 0.8653 \n",
      "pred day 14 test acc: 0.8487 - test f1macro: 0.4591 - test roc: 0.6687 \n",
      "final test acc: 0.9265 - test f1macro: 0.7064 - test roc: 0.8581\n",
      "pred day 0 test acc: 0.9538 - test f1macro: 0.8923 - test roc: 0.9297 \n",
      "pred day 1 test acc: 0.9707 - test f1macro: 0.6354 - test roc: 0.8312 \n",
      "pred day 2 test acc: 0.9297 - test f1macro: 0.7035 - test roc: 0.8758 \n",
      "pred day 3 test acc: 0.8823 - test f1macro: 0.7207 - test roc: 0.8793 \n",
      "pred day 4 test acc: 0.9138 - test f1macro: 0.6370 - test roc: 0.8505 \n",
      "pred day 5 test acc: 0.9216 - test f1macro: 0.6765 - test roc: 0.8038 \n",
      "pred day 6 test acc: 0.9205 - test f1macro: 0.7364 - test roc: 0.8310 \n",
      "pred day 7 test acc: 0.9526 - test f1macro: 0.8677 - test roc: 0.9446 \n",
      "pred day 8 test acc: 0.9774 - test f1macro: 0.4943 - test roc: 0.5509 \n",
      "pred day 9 test acc: 0.8580 - test f1macro: 0.4618 - test roc: 0.4719 \n",
      "pred day 10 test acc: 0.9615 - test f1macro: 0.4902 - test roc: 0.2957 \n",
      "pred day 11 test acc: 0.9621 - test f1macro: 0.4903 - test roc: 0.8507 \n",
      "pred day 12 test acc: 0.9693 - test f1macro: 0.4922 - test roc: 0.6729 \n",
      "pred day 13 test acc: 0.9214 - test f1macro: 0.4796 - test roc: 0.8797 \n",
      "pred day 14 test acc: 0.8634 - test f1macro: 0.4634 - test roc: 0.7014 \n",
      "final test acc: 0.9350 - test f1macro: 0.7146 - test roc: 0.8425\n",
      "pred day 0 test acc: 0.9523 - test f1macro: 0.8978 - test roc: 0.9542 \n",
      "pred day 1 test acc: 0.9655 - test f1macro: 0.6597 - test roc: 0.8406 \n",
      "pred day 2 test acc: 0.9277 - test f1macro: 0.7159 - test roc: 0.8843 \n",
      "pred day 3 test acc: 0.8929 - test f1macro: 0.7959 - test roc: 0.8940 \n",
      "pred day 4 test acc: 0.9121 - test f1macro: 0.6704 - test roc: 0.8802 \n",
      "pred day 5 test acc: 0.9158 - test f1macro: 0.7000 - test roc: 0.8367 \n",
      "pred day 6 test acc: 0.9019 - test f1macro: 0.7261 - test roc: 0.8326 \n",
      "pred day 7 test acc: 0.9369 - test f1macro: 0.8406 - test roc: 0.9435 \n",
      "pred day 8 test acc: 0.9635 - test f1macro: 0.5099 - test roc: 0.6401 \n",
      "pred day 9 test acc: 0.7586 - test f1macro: 0.4389 - test roc: 0.5053 \n",
      "pred day 10 test acc: 0.9402 - test f1macro: 0.4846 - test roc: 0.6094 \n",
      "pred day 11 test acc: 0.9256 - test f1macro: 0.4807 - test roc: 0.8754 \n",
      "pred day 12 test acc: 0.9551 - test f1macro: 0.4885 - test roc: 0.7506 \n",
      "pred day 13 test acc: 0.9214 - test f1macro: 0.4796 - test roc: 0.8649 \n",
      "pred day 14 test acc: 0.8403 - test f1macro: 0.4566 - test roc: 0.7049 \n",
      "final test acc: 0.9158 - test f1macro: 0.7000 - test roc: 0.8576\n"
     ]
    }
   ],
   "source": [
    "\n",
    "import os\n",
    "import torch.nn.functional as F\n",
    "\n",
    "args['num_features'] = num_features\n",
    "args['num_classes'] = num_classes\n",
    "results_SR = {\n",
    "    'acc' : np.zeros((16, REPEAT)),\n",
    "    'f1macro' : np.zeros((16, REPEAT)),\n",
    "    'aucroc' : np.zeros((16, REPEAT))\n",
    "}\n",
    "\n",
    "for repeat in range(REPEAT):\n",
    "    for part_id in p1_saved.shards_ids.keys():\n",
    "        # submodel training\n",
    "        start = time.time()\n",
    "        submodel = GINNet(input_dim=args['num_features'],\n",
    "                    output_dim=args['num_classes']).to(args['device'])\n",
    "\n",
    "        optimizer = torch.optim.Adam(\n",
    "            submodel.parameters(), lr=0.01, weight_decay=1e-5)\n",
    "        for epoch in range(1, 201):\n",
    "            submodel.train()\n",
    "            for gi in range(len(dataset_train)):\n",
    "                sub_graph = gis_keys_graph[gi][part_id]\n",
    "                sub_graph = sub_graph.to(device)\n",
    "                labels = sub_graph.y[sub_graph.y < 2]\n",
    "                if labels.sum() > 0:\n",
    "                    weight = (labels == 0).sum().item() / labels.sum().item()\n",
    "                else:\n",
    "                    weight = 1.0\n",
    "                optimizer.zero_grad()\n",
    "                out = submodel(sub_graph.x, sub_graph.edge_index)[\n",
    "                    sub_graph.train_mask]\n",
    "                loss = F.cross_entropy(out[torch.where(sub_graph.y < 2)[\n",
    "                    0]], labels, weight=torch.tensor([1., weight], device=device))\n",
    "\n",
    "                loss.backward()\n",
    "                optimizer.step()\n",
    "\n",
    "        submodel.to('cpu')\n",
    "        sub_graph = sub_graph.to('cpu')\n",
    "        savemodeln = partations_sr[gi].MPATH.replace('partid', 'part'+str(part_id))\n",
    "        os.makedirs(os.path.dirname(savemodeln), exist_ok=True)\n",
    "        torch.save(submodel.state_dict(), savemodeln)\n",
    "        del submodel, sub_graph, optimizer, out, loss\n",
    "\n",
    "    preds, targets = [], []\n",
    "    for day, data in enumerate(test_loader):\n",
    "        pm_kernel = PyramidMatchVector()\n",
    "        subgt_kfea = pm_kernel.parse_input([data.edge_index.cpu()])\n",
    "        subg_kfea = []\n",
    "        subouts = []\n",
    "        for part_id in p1_saved.shards_ids.keys():\n",
    "            for gi in range(len(dataset_train)):\n",
    "                loadname = partations_sr[gi].FPATH.replace('partid', 'part'+str(part_id))\n",
    "                subg_kfea += torch.load(loadname)\n",
    "\n",
    "            submodel = GINNet(input_dim=args['num_features'],\n",
    "                        output_dim=args['num_classes']).to(args['device'])\n",
    "            savemodeln = partations_sr[gi].MPATH.replace('partid', 'part'+str(part_id))\n",
    "            submodel.load_state_dict(torch.load(savemodeln))\n",
    "            submodel.eval()\n",
    "            subout = torch.softmax(submodel(data.x.to(args['device']), data.edge_index.to(args['device'])), dim=1)\n",
    "            if 'labeled_mask' in data:\n",
    "                subout = subout[data.labeled_mask]\n",
    "\n",
    "            subouts.append(subout.detach().to('cpu'))\n",
    "            submodel.to('cpu')\n",
    "            del submodel, subout\n",
    "\n",
    "        suboutTensor = torch.stack(subouts)\n",
    "\n",
    "        daily_weights = pm_kernel.kernel_similarity(subgt_kfea, subg_kfea)\n",
    "        daily_weights = torch.tensor(daily_weights).to(torch.float32).reshape(len(p1_saved.shards_ids.keys()), -1).mean(dim=-1)\n",
    "        saveweightn = partations_sr[gi].MPATH.replace('partid/submodels/','').replace('model_partid', f'day{day}_weight')\n",
    "        os.makedirs(os.path.dirname(saveweightn), exist_ok=True)\n",
    "        torch.save(daily_weights, saveweightn)\n",
    "\n",
    "        weights = F.softmax(daily_weights, dim=0)\n",
    "        weighted_pred = torch.tensordot(suboutTensor, weights, dims=([0], [0]))\n",
    "\n",
    "        if 'labeled_mask' in data:\n",
    "            target_labels = data.y[data.labeled_mask].detach().cpu().numpy()\n",
    "        else:\n",
    "            target_labels = data.y.detach().cpu().numpy()\n",
    "        \n",
    "        test_acc, test_f1macro, test_aucroc= store_metrics(weighted_pred.detach().cpu().numpy(), target_labels)\n",
    "        print(\"pred day {} test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f} \".format(day,\n",
    "            test_acc, test_f1macro, test_aucroc))\n",
    "        results_SR['acc'][day, repeat] = test_acc\n",
    "        results_SR['f1macro'][day, repeat] = test_f1macro\n",
    "        results_SR['aucroc'][day, repeat] = test_aucroc\n",
    "        preds.append(weighted_pred.detach().cpu().numpy())\n",
    "        targets.append(target_labels)\n",
    "\n",
    "    pred_all = np.concatenate(preds, axis=0)\n",
    "    targets_all = np.concatenate(targets, axis=0)\n",
    "    test_acc, test_f1macro, test_aucroc= store_metrics(pred_all, targets_all)\n",
    "    print(\"final test acc: {:.4f} - test f1macro: {:.4f} - test roc: {:.4f}\".format(test_acc, test_f1macro, test_aucroc))\n",
    "    results_SR['acc'][-1, repeat] = test_acc\n",
    "    results_SR['f1macro'][-1, repeat] = test_f1macro\n",
    "    results_SR['aucroc'][-1, repeat] = test_aucroc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_87122/15375388.py:15: MatplotlibDeprecationWarning: The seaborn styles shipped by Matplotlib are deprecated since 3.6, as they no longer correspond to the styles shipped by seaborn. However, they will remain available as 'seaborn-v0_8-<style>'. Alternatively, directly use the seaborn API instead.\n",
      "  plt.style.use('seaborn')\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import style\n",
    "matplotlib.rcParams.update(matplotlib.rcParamsDefault)\n",
    "plot_daily = {}\n",
    "plot_daily['licit'] = []\n",
    "plot_daily['illicit'] = []\n",
    "plot_daily['unknown'] = []\n",
    "for t in range(len(dataset_test)):\n",
    "    plot_daily['licit'].append((dataset_test[t].y==0).sum().item())\n",
    "    plot_daily['illicit'].append((dataset_test[t].y==1).sum().item())\n",
    "    plot_daily['unknown'].append((dataset_test[t].y==2).sum().item())\n",
    "#     print(plot_daily['illicit'][-1], (dataset_test[t].y==1).sum().item()/(dataset_test[t].y==0).sum().item())\n",
    "    \n",
    "plt.style.use('seaborn')\n",
    "fig, axright =plt.subplots(figsize=(12, 6))\n",
    "x = range(35,50)\n",
    "axleft = axright.twinx()\n",
    "axright.bar(x,list(np.asarray(plot_daily['unknown'] )), bottom= np.asarray(plot_daily['illicit'] )+np.asarray(plot_daily['licit'] ), \n",
    "        label='# Unknown',color=(150/256, 150/256, 150/256, 0.4))\n",
    "axright.bar(x,list(np.asarray(plot_daily['illicit'] )), label='# Illicit',bottom= np.asarray(plot_daily['licit'] ), \n",
    "         color=(246/256, 78/256, 139/256, 0.4), linewidth=1, edgecolor = (246/256, 78/256, 139/256, 1.0))\n",
    "\n",
    "axright.bar(x,list(np.asarray(plot_daily['licit'] )), \n",
    "        label='# licit',color=(58/256, 190/256, 120/256, 0.4), linewidth=1, edgecolor=(58/256, 190/256, 120/256, 1))\n",
    "\n",
    "axright.set_ylim([0,8000])\n",
    "# axright.legend(loc=(0.79, 0.75), prop={'size':17})\n",
    "axright.yaxis.tick_right()\n",
    "axright.set_ylabel('Number of Samples', fontsize=24, labelpad=10)\n",
    "axright.yaxis.set_label_position(\"right\")\n",
    "axright.yaxis.set_tick_params(labelsize=20)\n",
    "axright.set_xlabel('Time Step', fontsize=24)\n",
    "axright.set_xticks(range(35,50))\n",
    "axright.xaxis.set_tick_params(labelsize=20)\n",
    "\n",
    "axleft.plot(x,list(results_full['f1macro'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8, color=\"royalblue\", linestyle=\"-\", marker = 'o', label= 'Scratch')\n",
    "axleft.plot(x,list(results_Fast['f1macro'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8,  color=\"teal\", linestyle=\"-\", marker = 'D',label= 'GUIDE-Fast')\n",
    "axleft.plot(x,list(results_SR['f1macro'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8,  color=\"indigo\", linestyle=\"-\", marker = 's',label= 'GUIDE-SR')\n",
    "\n",
    "axleft.legend(prop={'size':17})\n",
    "# axleft.legend(loc='upper center', bbox_to_anchor=(0.5, -0.13), ncol=4, prop={'size':17})\n",
    "axleft.set_ylim([0.2, 1])\n",
    "axleft.yaxis.tick_left()\n",
    "axleft.set_ylabel('Macro F1', fontsize=24)\n",
    "axleft.yaxis.set_label_position(\"left\")\n",
    "axleft.yaxis.set_tick_params(labelsize=20)\n",
    "axleft.grid()\n",
    "axleft.set_xlabel('Time Step', fontsize=24)\n",
    "\n",
    "# plt.xlabel('Macro F1 results over test time span', fontsize=20)\n",
    "# plt.grid()\n",
    "dataname = 'EllipticBTC'\n",
    "# plt.savefig(f'logs/BTC_GIN_20_MacroF1_{dataname}.pdf', format=\"pdf\", bbox_inches='tight')\n",
    "plt.show()\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_87122/2370332187.py:5: MatplotlibDeprecationWarning: The seaborn styles shipped by Matplotlib are deprecated since 3.6, as they no longer correspond to the styles shipped by seaborn. However, they will remain available as 'seaborn-v0_8-<style>'. Alternatively, directly use the seaborn API instead.\n",
      "  plt.style.use('seaborn')\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import style\n",
    "matplotlib.rcParams.update(matplotlib.rcParamsDefault)\n",
    "plt.style.use('seaborn')\n",
    "fig, axright =plt.subplots(figsize=(12, 6))\n",
    "x = range(35,50)\n",
    "axleft = axright.twinx()\n",
    "axright.bar(x,list(np.asarray(plot_daily['licit'] )), \n",
    "        label='# licit',color=(58/256, 190/256, 120/256, 0.4), linewidth=1, edgecolor=(58/256, 190/256, 120/256, 1))\n",
    "axright.bar(x,list(np.asarray(plot_daily['illicit'] )), label='# Illicit',bottom= np.asarray(plot_daily['licit'] ), \n",
    "         color=(246/256, 78/256, 139/256, 0.4), linewidth=1, edgecolor = (246/256, 78/256, 139/256, 1.0))\n",
    "\n",
    "axright.bar(x,list(np.asarray(plot_daily['unknown'] )), bottom= np.asarray(plot_daily['illicit'] )+np.asarray(plot_daily['licit'] ), \n",
    "        label='# Unknown',color=(150/256, 150/256, 150/256, 0.4))\n",
    "\n",
    "axright.set_ylim([0,8000])\n",
    "# axright.legend(loc='best', prop={'size':17})\n",
    "axright.legend(loc='upper left', bbox_to_anchor=(0.71, 1.179), ncol=2, columnspacing=0.6, prop={'size':17})\n",
    "\n",
    "axright.yaxis.tick_right()\n",
    "axright.set_ylabel('Number of Samples', fontsize=24, labelpad=10)\n",
    "axright.yaxis.set_label_position(\"right\")\n",
    "axright.yaxis.set_tick_params(labelsize=20)\n",
    "axright.set_xticks(range(35,50))\n",
    "axright.xaxis.set_tick_params(labelsize=20)\n",
    "axright.set_xlabel('Time Step', fontsize=24)\n",
    "\n",
    "axleft.plot(x,list(results_full['aucroc'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8,  color=\"royalblue\", linestyle=\"-\", marker = 'o', label= 'Scratch')\n",
    "axleft.plot(x,list(results_Fast['aucroc'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8,  color=\"teal\", linestyle=\"-\", marker = 'D',label= 'GUIDE-Fast')\n",
    "axleft.plot(x,list(results_SR['aucroc'][:-1].mean(axis=1)), linewidth=3.0, alpha=0.8,  color=\"indigo\", linestyle=\"-\", marker = 's',label= 'GUIDE-SR')\n",
    "\n",
    "axleft.legend(prop={'size':17})\n",
    "axleft.legend(loc='upper center', bbox_to_anchor=(0.36, 1.18), ncol=3, prop={'size':17})\n",
    "\n",
    "axleft.set_ylim([0, 1])\n",
    "axleft.yaxis.tick_left()\n",
    "axleft.set_ylabel('AUC', fontsize=24)\n",
    "axleft.yaxis.set_label_position(\"left\")\n",
    "axleft.yaxis.set_tick_params(labelsize=20)\n",
    "axleft.grid()\n",
    "axleft.set_xlabel('Time Step', fontsize=24)\n",
    "\n",
    "# plt.savefig(f'logs/BTC_GIN_20_AUC_{dataname}.pdf', format=\"pdf\", bbox_inches='tight')\n",
    "plt.show()\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(partations_fast, './checkpoints/BTC/partitions_fast.pt')\n",
    "torch.save(partations_sr, './checkpoints/BTC/partitions_sr.pt')\n",
    "torch.save(results_full, './checkpoints/BTC/results_full.pt')\n",
    "torch.save(results_Fast, './checkpoints/BTC/results_fast.pt')\n",
    "torch.save(results_SR, './checkpoints/BTC/results_sr.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "partations_fast = torch.load('./checkpoints/BTC/partitions_fast.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'checkpoints/BTC_copy'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import shutil\n",
    "\n",
    "src_dir = 'checkpoints/BTC'\n",
    "dst_dir = 'checkpoints/BTC_copy'\n",
    "\n",
    "if os.path.exists(dst_dir):\n",
    "    shutil.rmtree(dst_dir)\n",
    "\n",
    "shutil.copytree(src_dir, dst_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_distance(data0, data1, distance='l2_norm' ):\n",
    "    if distance == 'l2_norm':\n",
    "        return torch.norm(data0-data1,dim=1)\n",
    "    elif distance =='direct_diff':\n",
    "        return data0 - data1\n",
    "    else:\n",
    "        raise Exception(\"Unsupported distance\")\n",
    "\n",
    "def evaluate_attack_with_AUC(data, label):\n",
    "    return roc_auc_score(label, data.reshape(-1, 1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# gis_keys_graph\n",
    "id_gi_part_innerid = []\n",
    "for gi in range(train_day):\n",
    "    ids = torch.where(dataset.step==gi+1)[0].tolist()\n",
    "    gis = [gi] * len(ids)\n",
    "    inday_ids = range(len(ids))\n",
    "    part_ids = partations_fast[gi].ids_shards.values()\n",
    "    id_gi_part_innerid += list(zip(ids, gis, part_ids, inday_ids))\n",
    "\n",
    "torch.save(id_gi_part_innerid, partations_fast[gi].DPATH.split('partid')[0]+ 'id_gi_part_innerid.pt')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unlearning_number:100\n"
     ]
    }
   ],
   "source": [
    "import copy\n",
    "import random\n",
    "import collections\n",
    "from itertools import compress\n",
    "from GUIDE.kernel_vector import PyramidMatchVector\n",
    "num_features = 165\n",
    "num_classes = 2\n",
    "\n",
    "args = {}\n",
    "args['dataname'] = 'BTC'\n",
    "args['shards_number'] = 20\n",
    "args['device'] = device\n",
    "args['num_features'] = num_features\n",
    "args['num_classes'] = num_classes\n",
    "\n",
    "unlearning_number = 100\n",
    "print(f'unlearning_number:{unlearning_number}')\n",
    "G_ny0 = [dataset_test[-1].edge_index.cpu()]\n",
    "pm_kernel = PyramidMatchVector(normalize=True, with_labels=False)\n",
    "G_ny0 = pm_kernel.parse_input(G_ny0)\n",
    "average_countpart = []\n",
    "train_labeled_idx = torch.where(torch.logical_and(dataset.step<=30,dataset.y<2))[0].tolist()\n",
    "unlearning_id = random.sample(train_labeled_idx,k=int(unlearning_number))\n",
    "sub_tuples = [id_gi_part_innerid[i] for i in unlearning_id]\n",
    "\n",
    "gi_tuples = collections.defaultdict(list)\n",
    "for tup in sub_tuples:\n",
    "    gi_tuples[tup[1]].append(tup[2:])\n",
    "\n",
    "gi_part_uid = {}\n",
    "average_countpart_ = []\n",
    "for i in gi_tuples.keys():\n",
    "    gi_tuples_in = collections.defaultdict(list)\n",
    "    for j in gi_tuples[i]:\n",
    "        gi_tuples_in[j[0]].append(j[1])\n",
    "\n",
    "    gi_part_uid[i] = gi_tuples_in\n",
    "    average_countpart_ += list(gi_tuples_in.keys())\n",
    "\n",
    "average_countpart.append(len(set(average_countpart_)))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn.functional as F\n",
    "\n",
    "gis_keys_graph = {}\n",
    "pm_kernel = PyramidMatchVector()\n",
    "\n",
    "for gi in range(len(dataset_train)):\n",
    "    gis_keys_graph[gi] = {}\n",
    "    for part_id in range(args['shards_number']):\n",
    "        loadname = partations_fast[gi].DPATH.replace('partid', 'part'+str(part_id))\n",
    "        sub_graph = torch.load(loadname)\n",
    "        gis_keys_graph[gi][part_id] = sub_graph\n",
    "\n",
    "gis_keys_graph_update = copy.deepcopy(gis_keys_graph)\n",
    "part_set = []\n",
    "for gi in gi_part_uid.keys():\n",
    "    for part in gi_part_uid[gi].keys():\n",
    "        part_set.append(part)\n",
    "        reserve_mask = torch.tensor([False if id in gi_part_uid[gi][part] else True for id in gis_keys_graph_update[gi][part].uids])\n",
    "        sub_graph = gis_keys_graph_update[gi][part].subgraph(reserve_mask)\n",
    "        sub_graph.uids = list(compress(sub_graph.uids, reserve_mask.tolist()))\n",
    "        sub_graph.labeled_mask =  (sub_graph.y <2)                  \n",
    "        gis_keys_graph_update[gi][part] = sub_graph\n",
    "\n",
    "G_nx0 = []\n",
    "part_set = set(part_set)\n",
    "# the following part can be parallel\n",
    "for part in part_set:\n",
    "    for gi in gis_keys_graph_update.keys():\n",
    "        G_nx0.append(gis_keys_graph_update[gi][part].edge_index.to('cpu'))\n",
    "\n",
    "    submodel = GINNet(args['num_features'],\n",
    "                        args['num_classes']).to(args['device'])\n",
    "    optimizer = torch.optim.Adam(\n",
    "        submodel.parameters(), lr=0.01, weight_decay=1e-5)\n",
    "\n",
    "    for epoch in range(1, 201):\n",
    "        submodel.train()\n",
    "        for gi in range(len(dataset_train)):\n",
    "            sub_graph = gis_keys_graph_update[gi][part]\n",
    "            sub_graph = sub_graph.to(args['device'])\n",
    "            labels = sub_graph.y[sub_graph.y<2]\n",
    "            if labels.sum() > 0:\n",
    "                weight = (labels == 0).sum().item() / labels.sum().item()\n",
    "            else:\n",
    "                weight = 1.0\n",
    "            optimizer.zero_grad()\n",
    "            out = submodel(sub_graph.x, sub_graph.edge_index)[\n",
    "                sub_graph.train_mask]\n",
    "            loss = F.cross_entropy(out[torch.where(sub_graph.y<2)[0]], labels,  weight=torch.tensor([1., weight], device=device))\n",
    "\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "\n",
    "    submodel.to('cpu')\n",
    "    sub_graph.to('cpu')\n",
    "    savemodeln = partations_fast[gi].MPATH.replace('/partid', '_copy/partid').replace('partid', 'part'+str(part_id))\n",
    "    os.makedirs(os.path.dirname(savemodeln), exist_ok=True)\n",
    "    torch.save(submodel.state_dict(), savemodeln)\n",
    "\n",
    "G_nx0 = pm_kernel.parse_input(G_nx0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Fast'\n",
    "positive0, negative0 = {}, {}\n",
    "positive0[method], negative0[method]  = [], []\n",
    "\n",
    "submodellist = []\n",
    "part_set_all = range(args['shards_number'])\n",
    "for part in part_set_all:\n",
    "    submodel = GINNet(args['num_features'],\n",
    "                        args['num_classes'])\n",
    "\n",
    "    loadname = partations_fast[gi].MPATH.replace('partid', 'part'+str(part_id))\n",
    "\n",
    "    submodel.load_state_dict(torch.load(loadname))\n",
    "    submodel.eval()\n",
    "    submodellist.append(submodel)\n",
    "\n",
    "subouts_train = []\n",
    "subout_tests = []\n",
    "part_set_all = range(args['shards_number'])\n",
    "for part in part_set_all:\n",
    "    subout_train = []\n",
    "    for data in dataset_train:\n",
    "        subout_train_daily= torch.softmax(submodellist[part](data.x, data.edge_index),dim=1)\n",
    "        subout_train.append(subout_train_daily)\n",
    "\n",
    "    subout_train = torch.concat(subout_train)\n",
    "    subout_train_ = subout_train[unlearning_id]\n",
    "    subouts_train.append(subout_train_.detach())\n",
    "\n",
    "    subout_test = []\n",
    "    for data in dataset_test:\n",
    "        subout_test_daily= torch.softmax(submodellist[part](data.x, data.edge_index),dim=1)\n",
    "        subout_test.append(subout_test_daily)\n",
    "\n",
    "    subout_test = torch.concat(subout_test)\n",
    "    subout_tests.append(subout_test.detach())\n",
    "\n",
    "suboutTensor = torch.stack(subouts_train)\n",
    "suboutTensorn = torch.stack(subout_tests)\n",
    "\n",
    "# weights_gi = torch.tensor([1.0/args['shards_number']\n",
    "#                         for _ in range(args['shards_number'])])\n",
    "weights_gi = torch.load(partations_fast[gi].MPATH.replace('partid/submodels/','').replace('model_partid', f'day{14}_weight'))\n",
    "\n",
    "summed = torch.tensordot(\n",
    "    suboutTensor, torch.softmax(weights_gi, dim=0), dims=([0], [0]))\n",
    "positive0[method].append(summed.cpu())\n",
    "\n",
    "summed = torch.tensordot(\n",
    "    suboutTensorn, torch.softmax(weights_gi, dim=0), dims=([0], [0]))\n",
    "negative0[method].append(summed.cpu())\n",
    "del suboutTensor, suboutTensorn, summed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp = time.strftime(\"%Y%m%d-%H%M%S\")\n",
    "savename = partations_fast[gi].MPATH.split('partid')[0]+f'evaluation_attack_partition_{method}_BTC_unlearning_idrnd'\n",
    "savename += ''.join(tmp)\n",
    "savename += ''.join('.pt')\n",
    "torch.save(unlearning_id,savename)\n",
    "\n",
    "savename = partations_fast[gi].MPATH.split('partid')[0]+f'evaluation_attack_partition_{method}_BTC_positive0'\n",
    "savename += ''.join(tmp)\n",
    "savename += ''.join('.pt')\n",
    "torch.save(positive0,savename)\n",
    "savename = partations_fast[gi].MPATH.split('partid')[0]+f'evaluation_attack_partition_{method}_BTC_negative0'\n",
    "savename += ''.join(tmp)\n",
    "savename += ''.join('.pt')\n",
    "torch.save(negative0,savename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Fast'\n",
    "positive1, negative1 = {}, {}\n",
    "positive1[method], negative1[method]  = [], []\n",
    "\n",
    "submodellist = []\n",
    "part_set_all = range(args['shards_number'])\n",
    "for part in part_set_all:\n",
    "    submodel = GINNet(args['num_features'],\n",
    "                        args['num_classes'])\n",
    "\n",
    "    loadname = partations_fast[gi].MPATH.replace('partid', 'part'+str(part_id)).replace('BTC', 'BTC_copy')\n",
    "\n",
    "    submodel.load_state_dict(torch.load(loadname))\n",
    "    submodel.eval()\n",
    "    submodellist.append(submodel)\n",
    "\n",
    "subouts_train = []\n",
    "subout_tests = []\n",
    "part_set_all = range(args['shards_number'])\n",
    "for part in part_set_all:\n",
    "    subout_train = []\n",
    "    for data in dataset_train:\n",
    "        subout_train_daily= torch.softmax(submodellist[part](data.x, data.edge_index),dim=1)\n",
    "        subout_train.append(subout_train_daily)\n",
    "\n",
    "    subout_train = torch.concat(subout_train)\n",
    "    subout_train_ = subout_train[unlearning_id]\n",
    "    subouts_train.append(subout_train_.detach())\n",
    "\n",
    "    subout_test = []\n",
    "    for data in dataset_test:\n",
    "        subout_test_daily= torch.softmax(submodellist[part](data.x, data.edge_index),dim=1)\n",
    "        subout_test.append(subout_test_daily)\n",
    "\n",
    "    subout_test = torch.concat(subout_test)\n",
    "    subout_tests.append(subout_test.detach())\n",
    "\n",
    "suboutTensor = torch.stack(subouts_train)\n",
    "suboutTensorn = torch.stack(subout_tests)\n",
    "\n",
    "# weights_gi = torch.tensor([1.0/args['shards_number']\n",
    "#                         for _ in range(args['shards_number'])])\n",
    "weights_gi = torch.load(partations_fast[gi].MPATH.replace('partid/submodels/','').replace('model_partid', f'day{14}_weight'))\n",
    "\n",
    "summed = torch.tensordot(\n",
    "    suboutTensor, torch.softmax(weights_gi, dim=0), dims=([0], [0]))\n",
    "positive1[method].append(summed.cpu())\n",
    "\n",
    "summed = torch.tensordot(\n",
    "    suboutTensorn, torch.softmax(weights_gi, dim=0), dims=([0], [0]))\n",
    "negative1[method].append(summed.cpu())\n",
    "del suboutTensor, suboutTensorn, summed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "savename = partations_fast[gi].MPATH.split('partid')[0]+f'evaluation_attack_partition_{method}_BTC_positive1'\n",
    "savename += ''.join(tmp)\n",
    "savename += ''.join('.pt')\n",
    "torch.save(positive1,savename)\n",
    "savename = partations_fast[gi].MPATH.split('partid')[0]+f'evaluation_attack_partition_{method}_BTC_negative1'\n",
    "savename += ''.join(tmp)\n",
    "savename += ''.join('.pt')\n",
    "torch.save(negative1,savename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BTC Pos:Neg Average Attack_Model_B  AUC: Fast 0.4781488   0.0007089196385600001\n"
     ]
    }
   ],
   "source": [
    "positive_posteriors, negative_posteriors = [], []\n",
    "positive_posteriors.append(positive0[method])\n",
    "positive_posteriors.append(positive1[method])\n",
    "negative_posteriors.append(negative0[method])\n",
    "negative_posteriors.append(negative1[method])\n",
    "\n",
    "\n",
    "size_number = 100\n",
    "# break\n",
    "\n",
    "positive_posteriors[0] = torch.cat(positive_posteriors[0], dim=0)\n",
    "positive_posteriors[1] = torch.cat(positive_posteriors[1], dim=0)\n",
    "negative_posteriors[0] = torch.cat(negative_posteriors[0], dim=0)\n",
    "negative_posteriors[1] = torch.cat(negative_posteriors[1], dim=0)\n",
    "\n",
    "attack_auc_b = []\n",
    "for _ in range(1000):\n",
    "    batch_test = random.choices(range(negative_posteriors[0].shape[0]),k=size_number)\n",
    "    label = torch.cat((torch.ones(size_number), torch.zeros(size_number)))\n",
    "    data={}\n",
    "    for i in range(2):\n",
    "        data[i] = torch.cat((positive_posteriors[i], negative_posteriors[i][batch_test,:]),0)\n",
    "\n",
    "    # calculate l2 distance\n",
    "    model_b_distance = calculate_distance(data[0], data[1])\n",
    "    # directly calculate AUC with feature and labels\n",
    "    attack_auc_b_ = evaluate_attack_with_AUC(model_b_distance.cpu().numpy(), label)\n",
    "    # print(\"Attack_Model_B AUC: %s \" % (attack_auc_b_))\n",
    "    attack_auc_b.append(attack_auc_b_)\n",
    "\n",
    "\n",
    "print(f'BTC Pos:Neg Average Attack_Model_B  AUC: {method} {(np.array(attack_auc_b).mean())}   {(np.array(attack_auc_b).var())}')\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "GUIDE",
   "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": "3.9.16"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "297047daa5b79515bffddd111d4dbac359b4d66c6c54e395340ce1927d9897c3"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}