freq_graph_gcn_train_on_camda_tiny.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import networkx as nx\n",
    "import numpy as np\n",
    "import torch\n",
    "from torch_geometric.utils.convert import from_networkx\n",
    "import random\n",
    "import networkx as nx\n",
    "from collections import Counter, defaultdict\n",
    "from torch_geometric.loader import DataLoader\n",
    "from torch.nn import Linear\n",
    "import torch.nn.functional as F\n",
    "from torch_geometric.nn import GCNConv\n",
    "from torch_geometric.nn import global_mean_pool\n",
    "import src.utils.utils as ut\n",
    "import os\n",
    "import pickle\n",
    "import datetime\n",
    "from multiprocessing import Process, Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "skipping SCL.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_BCN_017_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_ILR_022_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD17_HKG_010_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "skipping HKG.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_BCN_012_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_FAI_020_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_SCL_009_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_DOH_044_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_FAI_009_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_BCN_026_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_FAI_024_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_DOH_013_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "reading CAMDA20_MetaSUB_CSD16_BCN_038_1_kneaddata_subsampled_20_percent.labeled_graph_max\n",
      "skipping BCN.labeled_graph_max\n",
      "skipping ILR.labeled_graph_max\n",
      "skipping DOH.labeled_graph_max\n",
      "skipping FAI.labeled_graph_max\n"
     ]
    }
   ],
   "source": [
    "graphs_dir = '/home/sysbio/DBG-GNN/data/camda2020_tiny_max'\n",
    "graphs = []\n",
    "for file in os.listdir(graphs_dir):\n",
    "    if not file.startswith('CAMDA'):\n",
    "        print(f'skipping {file}')\n",
    "        continue\n",
    "    print(f'reading {file}')\n",
    "    with open(os.path.join(graphs_dir, file), 'rb') as f:\n",
    "        graph = pickle.load(f)\n",
    "    graphs.append(graph)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Data(x=[512, 25], edge_index=[2, 2161], weight=[2161], y=[1])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graphs[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Data(x=[561, 25], edge_index=[2, 2343], weight=[2343], y=[1])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graphs[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "graphs_n = len(graphs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "random.shuffle(graphs)\n",
    "split_value = 0.8\n",
    "train_dataset = graphs[:int(graphs_n * split_value)]\n",
    "val_dataset = graphs[int(graphs_n * split_value):]\n",
    "\n",
    "train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)\n",
    "val_loader = DataLoader(val_dataset, batch_size=1, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<torch_geometric.loader.dataloader.DataLoader at 0x7536f25bbb90>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "class GCN(torch.nn.Module):\n",
    "    def __init__(self, num_features, hidden_channels, num_classes):\n",
    "        super(GCN, self).__init__()\n",
    "        torch.manual_seed(12345)\n",
    "        self.conv1 = GCNConv(num_features, hidden_channels)\n",
    "        self.conv2 = GCNConv(hidden_channels, hidden_channels)\n",
    "        self.conv3 = GCNConv(hidden_channels, hidden_channels)\n",
    "        self.lin = Linear(hidden_channels, num_classes)\n",
    "\n",
    "    def forward(self, x, edge_index, edge_weight, batch):\n",
    "        # 1. Obtain node embeddings \n",
    "        x = self.conv1(x, edge_index, edge_weight)\n",
    "        x = x.relu()\n",
    "        x = self.conv2(x, edge_index, edge_weight)\n",
    "        x = x.relu()\n",
    "        x = self.conv3(x, edge_index, edge_weight)\n",
    "\n",
    "        # 2. Readout layer\n",
    "        x = global_mean_pool(x, batch)  # [batch_size, hidden_channels]\n",
    "\n",
    "        # 3. Apply a final classifier\n",
    "        x = F.dropout(x, p=0.5, training=self.training)\n",
    "        x = self.lin(x)\n",
    "        \n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmer_len = 4\n",
    "subkmer_len = 2\n",
    "num_features = 5**subkmer_len\n",
    "num_classes = 6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1, Train Loss: 1.7988, Train Acc: 0.3333, Val Acc: 0.3333\n",
      "Epoch 201, Train Loss: 0.9024, Train Acc: 0.6667, Val Acc: 0.6667\n",
      "Epoch 401, Train Loss: 0.4798, Train Acc: 0.7778, Val Acc: 0.6667\n",
      "Epoch 601, Train Loss: 0.3656, Train Acc: 0.8889, Val Acc: 0.6667\n",
      "Epoch 801, Train Loss: 0.0921, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 1001, Train Loss: 0.2507, Train Acc: 0.8889, Val Acc: 0.6667\n",
      "Epoch 1201, Train Loss: 0.1015, Train Acc: 0.8889, Val Acc: 0.6667\n",
      "Epoch 1401, Train Loss: 0.1052, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 1601, Train Loss: 0.0098, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 1801, Train Loss: 0.0121, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 2001, Train Loss: 0.0009, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 2201, Train Loss: 0.0073, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 2401, Train Loss: 0.0101, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 2601, Train Loss: 0.0196, Train Acc: 1.0000, Val Acc: 0.6667\n",
      "Epoch 2801, Train Loss: 0.1605, Train Acc: 0.8889, Val Acc: 0.6667\n"
     ]
    }
   ],
   "source": [
    "def train():\n",
    "    model.train()\n",
    "    total_loss = 0\n",
    "    correct = 0\n",
    "    for data in train_loader:\n",
    "        optimizer.zero_grad()\n",
    "        out = model(data.x, data.edge_index, data.weight, data.batch)\n",
    "        loss = criterion(out, data.y)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        total_loss += loss.item()\n",
    "        pred = out.argmax(dim=1)\n",
    "        correct += (pred == data.y).sum().item()\n",
    "    return total_loss / len(train_loader), correct / len(train_dataset)\n",
    "\n",
    "# Validation function\n",
    "def validate(loader):\n",
    "    model.eval()\n",
    "    correct = 0\n",
    "    for data in loader:\n",
    "        with torch.no_grad():\n",
    "            out = model(data.x, data.edge_index, data.weight, data.batch)\n",
    "            pred = out.argmax(dim=1)\n",
    "            correct += (pred == data.y).sum().item()\n",
    "    return correct / len(loader.dataset)\n",
    "\n",
    "# Initialize the model, optimizer, and loss function\n",
    "model = GCN(num_features=num_features, hidden_channels=32, num_classes=num_classes)\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=0.01)\n",
    "criterion = torch.nn.CrossEntropyLoss()\n",
    "\n",
    "# Training loop\n",
    "train_loss_for_plot = []\n",
    "train_acc_for_plot = []\n",
    "val_acc_for_plot = []\n",
    "num_epochs = 3000\n",
    "for epoch in range(num_epochs):\n",
    "    train_loss, train_acc = train()\n",
    "    val_acc = validate(val_loader)\n",
    "    train_loss_for_plot.append(train_loss)\n",
    "    train_acc_for_plot.append(train_acc)\n",
    "    val_acc_for_plot.append(val_acc)\n",
    "    if epoch % 200 == 0:\n",
    "        print(f'Epoch {epoch+1}, Train Loss: {train_loss:.4f}, Train Acc: {train_acc:.4f}, Val Acc: {val_acc:.4f}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'train loss')"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_loss_for_plot)\n",
    "plt.title('train loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'train acc')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_acc_for_plot)\n",
    "plt.title('train acc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'val acc')"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(val_acc_for_plot)\n",
    "plt.title('val acc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "models_outdir = '/home/sysbio/DBG-GNN/trained_models'\n",
    "model_outfile = os.path.join(models_outdir, 'camda2020_tiny_max_gcn.model')\n",
    "with open(model_outfile, 'wb') as f:\n",
    "    pickle.dump(model, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "GNNs",
   "language": "python",
   "name": "gnns"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}