{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Elliptic tutorial \n",
    "\n",
    "<a target=\"_blank\" href=\"https://colab.research.google.com/github/AI4FinTech/FinTorch/blob/main/docs/tutorials/elliptic/Elliptic.ipynb\">\n",
    "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n",
    "</a>\n",
    "\n",
    "In this hands-on Python tutorial, we'll delve into an intriguing dataset that captures the dynamics of transactions within a blockchain network. Our goal will be to prepare this data for training machine learning models using the powerful PyTorch and PyTorch Lightning frameworks. By the end of this tutorial, you'll have a solid foundation for training machine learning models on this dataset. This foundation sets the stage for exciting applications such as: Fraud detection, Transaction pattern analysis, Blockchain network behavior prediction. Let's get started!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tutorial Outline\n",
    "\n",
    "1. Loading the Data\n",
    "2. Exploratory Data Analysis (EDA)\n",
    "3. Graph Visualization\n",
    "4. Machine Learning Models\n",
    "5. Simple Graph Neural Network (GNN) Model\n",
    "6. GNN for Heterogeneous Graphs (Elliptic++)\n",
    "7. Conclusion and Further Explorations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Install FinTorch package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install fintorch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "# Installation of PyTorch Geometric and dependencies based on detected versions\n",
    "def install_pyg_and_dependencies():\n",
    "  !pip install pyg-lib -f https://data.pyg.org/whl/torch-{torch.__version__}.html\n",
    "  !pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-{torch.__version__}.html\n",
    "\n",
    "# Detect PyTorch version\n",
    "if torch.__version__ >= \"1.13.0\":\n",
    "  print(\"PyTorch version 1.13.0 or newer detected. Installing PyG and dependencies...\")\n",
    "  install_pyg_and_dependencies()\n",
    "else:\n",
    "  print(\"PyTorch version is older than 1.13.0. PyG might not work correctly. Please upgrade PyTorch or use the pip install torch_geometric method.\")\n",
    "  \n",
    "\n",
    "# Verify installation\n",
    "try:\n",
    "  import torch_geometric\n",
    "  print(f\"PyTorch Geometric successfully installed. Version: {torch_geometric.__version__}\")\n",
    "except ImportError:\n",
    "  print(\"PyTorch Geometric not found. Installation might have failed.\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optional: colab Kaggle setups"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To download the dataset from Kaggle in Colab, you need to set your kaggle username and kaggle secret."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First configure the `KAGGLE_USERNAME` and `KAGGLE_SECRET` in colab.\n",
    "![Colab secrets](colabsecrets.png)\n",
    "\n",
    "Next, make the secrets available as environment variables as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from google.colab import userdata\n",
    "import os\n",
    "\n",
    "os.environ[\"KAGGLE_KEY\"] = userdata.get('KAGGLE_KEY')\n",
    "os.environ[\"KAGGLE_USERNAME\"] = userdata.get('KAGGLE_USERNAME')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset background\n",
    "The dataset description originates from [Kaggle Elliptic Dataset](https://www.kaggle.com/datasets/ellipticco/elliptic-data-set) and [GitHub Elliptic++ Dataset](https://github.com/git-disl/EllipticPlusPlus/tree/main?tab=readme-ov-file). The dataset consist of two parts:\n",
    "\n",
    "* **Actor dataset**: this dataset consists of 822k wallet addresses to enable the detection of illicit addresses (actors) in the Bitcoin network by leveraging graph data.\n",
    "* **Transaction dataset**: This anonymized dataset - with over 200k transactions - is a transaction graph collected from the Bitcoin blockchain. A node in the graph represents a transaction, an edge can be viewed as a flow of Bitcoins between one transaction and the other. Each node has 166 features and has been labeled as being created by a \"licit\", \"illicit\" or \"unknown\" entity.\n",
    "\n",
    "\n",
    "The following figure shows the schematic of the dataset as a graph ([source](https://github.com/git-disl/EllipticPlusPlus/blob/main/images/actorvizaddrtx.jpg)).\n",
    "Note that this is the graph with both wallets and transactions.\n",
    "\n",
    "![Schematic visualization of the Address-Transaction graph](actorvizaddrtx.jpg)\n",
    "\n",
    "We describe how the dataset is structured which enables us to develop models that detect, for example, fraudulent transactions. \n",
    "\n",
    "## Transaction graph\n",
    "We first describe the transaction graph from [Kaggle Elliptic Dataset](https://www.kaggle.com/datasets/ellipticco/elliptic-data-set). \n",
    "\n",
    "### Nodes and edges\n",
    "\n",
    "The transaction graph is made of 203,769 nodes and 234,355 edges. Two percent (4,545) of the nodes are labelled class 1 (illicit). Twenty-one percent (42,019) are labelled class 2 (licit). The remaining transactions are not labelled with regard to licit versus illicit.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Features\n",
    "\n",
    "Each node in our dataset has 166 associated features.  Unfortunately, we can't disclose the full feature list but we can show categories of features.\n",
    "\n",
    "One key feature is a time step (1 to 49) representing when a Bitcoin transaction was broadcast. These time steps are spaced approximately two weeks apart. Within each time step, a single connected component captures transactions appearing on the blockchain within three hours of one another. Note that no connections exist between different time steps.\n",
    "\n",
    "The first 94 features provide localized transaction data. This includes the time step, input/output counts, fees, output volume, and averages like BTC received/spent by inputs/outputs, and average incoming/outgoing transaction counts for inputs/outputs.\n",
    "\n",
    "The final 72 features are aggregates derived from transactions one hop away (backward/forward) from the central node. These include maximums, minimums, standard deviations, and correlation coefficients for the same localized data (input/output counts, fees, etc.).\n",
    "\n",
    "\n",
    "For detailed statistics, please visit the Kaggle Data Explorer of the [Elliptic Data Set](https://www.kaggle.com/datasets/ellipticco/elliptic-data-set). "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading the data\n",
    "We use the FinTorch.datasets library to load the [Elliptic Data Set](https://www.kaggle.com/datasets/ellipticco/elliptic-data-set). The following code downloads the dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Processing...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "# from fintorch.datasets import elliptic\n",
    "from fintorch.datasets import elliptic\n",
    "\n",
    "# Load the elliptic dataset\n",
    "elliptic_dataset = elliptic.TransactionDataset('~/.fintorch_data', force_reload=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's discuss the code line by line:\n",
    "\n",
    "1. **Importing:** We import the elliptic module from the fintorch.datasets package. This module provides convenient access to the Elliptic Bitcoin Dataset.\n",
    "\n",
    "2. **Loading the Dataset:** We create an instance of the elliptic.TransactionDataset class and store it in the dataset variable. This loads the dataset from Kaggle and places it in the .fintorch_data/ directory. The fintorch framework uses with the [Kaggle API](https://github.com/Kaggle/kaggle-api) to download datasets. Make sure you've followed the instructions in the fintorch documentation to set up your Kaggle API credentials for seamless data access.\n",
    "\n",
    "\n",
    "\n",
    "With the dataset ready, let's examine its structure. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will render the transaction graph first. We will sample from the complete graph and display a sub-graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start-node BFS:118214\n",
      "Sub-graph nodes:244 \n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import networkx as nx\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from torch_geometric.utils.convert import to_networkx\n",
    "\n",
    "# Create a networkx dataset for plotting\n",
    "graph = to_networkx(elliptic_dataset[0], to_undirected=True)\n",
    "\n",
    "# Perform a breadth-first random walk from an illicit node\n",
    "start_node = np.random.choice(np.where(elliptic_dataset[0].y.numpy() == 1)[0]) \n",
    "print(f\"Start-node BFS:{start_node}\")\n",
    "subgraph_edges = nx.bfs_edges(graph, start_node, depth_limit=5)\n",
    "\n",
    "# Create the sub-graph\n",
    "subgraph = graph.edge_subgraph(subgraph_edges)\n",
    "\n",
    "print(f\"Sub-graph nodes:{len(subgraph.nodes())} \")\n",
    "\n",
    "# Plot the network using seaborn\n",
    "sns.set(style=\"white\")\n",
    "plt.figure(figsize=(10, 8))\n",
    "pos = nx.spring_layout(subgraph)\n",
    "nx.draw_networkx(subgraph, pos, with_labels=False, node_color=elliptic_dataset[0].y.numpy()[subgraph.nodes()], cmap='viridis', node_size=20)\n",
    "plt.title('Network Visualization')\n",
    "plt.axis('off')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here a subgraph is shown. We sampled a random illicit node from the complete undirected graph. Then, we perform a breadth-first search to select nodes in the sub-graph. The colors represent the different types of nodes: licit, illicit, and unknown."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploration\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We convert the PyTorch DataSet into a Polars DataSet and perform basic exploratory data analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fintorch.datasets.elliptic.TransactionDataset"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(elliptic_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([203769, 166])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "elliptic_dataset.x.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have a single graph thus we access element 0 in the data list:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Data(x=[203769, 166], edge_index=[2, 234355], y=[203769], train_mask=[203769], val_mask=[203769], test_mask=[203769])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "elliptic_dataset[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have the following elements in the dataset:\n",
    "\n",
    "* **x:** 203.769 nodes with 166 feature values\n",
    "* **edge_index:** 234.355 pairs of nodes representing the edges between nodes. Note that we transformed the node names into indices. The mapping is stored in *elliptic_dataset.map_id*\n",
    "* **train_mask:** a mask to indicate which nodes are used to train the model\n",
    "* **val_mask:** a mask to indicate which nodes are used as validation set\n",
    "* **test_mask:** a mask to indicate which nodes are used as a test set\n",
    "\n",
    "In addition, we can query some properties of the dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of node features: 166\n",
      "Number of edge features: 0\n",
      "Number of classes: 3\n",
      "Feature input matrix shape:torch.Size([203769, 166])\n",
      "Edge index feature matrix shape:torch.Size([2, 234355])\n",
      "Label feature matrix shape:torch.Size([203769])\n"
     ]
    }
   ],
   "source": [
    "print(f'Number of node features: {elliptic_dataset.num_features}')\n",
    "print(f'Number of edge features: {elliptic_dataset.num_edge_features}')\n",
    "print(f'Number of classes: {elliptic_dataset.num_classes}')\n",
    "print(f'Feature input matrix shape:{elliptic_dataset.x.shape}')\n",
    "print(f'Edge index feature matrix shape:{elliptic_dataset.edge_index.shape}')\n",
    "print(f'Label feature matrix shape:{elliptic_dataset.y.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape: (3, 3)\n",
      "┌─────┬────────┬──────────────────┐\n",
      "│     ┆ count  ┆ count_normalized │\n",
      "│ --- ┆ ---    ┆ ---              │\n",
      "│ i64 ┆ u32    ┆ f64              │\n",
      "╞═════╪════════╪══════════════════╡\n",
      "│ 2   ┆ 157205 ┆ 0.771486         │\n",
      "│ 0   ┆ 42019  ┆ 0.206209         │\n",
      "│ 1   ┆ 4545   ┆ 0.022305         │\n",
      "└─────┴────────┴──────────────────┘\n"
     ]
    }
   ],
   "source": [
    "import polars as pol\n",
    "\n",
    "# Convert elliptic_dataset.y to a numpy array and then to a polars Series\n",
    "y_series = pol.Series(elliptic_dataset.y.numpy())\n",
    "\n",
    "# Calculate the fraction of each value in the distribution\n",
    "fraction = y_series.value_counts() \n",
    "# Normalize the count column in fraction\n",
    "fraction = fraction.with_columns(count_normalized = fraction['count'] / y_series.shape[0])\n",
    "\n",
    "\n",
    "# Print the fraction of the value distribution\n",
    "print(fraction)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The class distribution reveals a severe imbalance, with the \"unknown\" class dominating 80% of the dataset. This means that a naive model could achieve a misleadingly high accuracy of 80% simply by always predicting the majority class.  It's crucial to be aware of this imbalance when evaluating model performance on this dataset. \n",
    "Relying solely on accuracy could lead to the false impression that a model is performing well, when in reality it's merely taking advantage of the skewed distribution. \n",
    "To get a true understanding of model performance, consider metrics like precision, recall, and F1-score. \n",
    "Additionally, it's important to explore techniques for addressing class imbalance, such as resampling, cost-sensitive learning, or specialized loss functions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple model\n",
    "While we'll demonstrate loading the Elliptic dataset into a simple graph neural network model and use accuracy as the evaluation metric for illustrative purposes, it's important to remember the class imbalance we just discussed. In this case, accuracy alone won't be a reliable measure of performance. Our main focus here is to showcase the loading process, not achieve optimal performance. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code defines a Graph Neural Network (GNN) model for processing data arranged in graphs. It takes node features and information about how the nodes connect as inputs. The model then stacks several layers that combine features from neighboring nodes, similar to how information spreads in a network. Finally, it outputs a new set of features for each node, potentially useful for classification tasks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch import Tensor\n",
    "import torch.nn as nn\n",
    "import torch_geometric.nn as geom_nn\n",
    "\n",
    "import torch.optim as optim\n",
    "\n",
    "\n",
    "class GNNModel(nn.Module):\n",
    "\n",
    "    def __init__(self, c_in: int, c_hidden: int, c_out: int, num_layers: int = 5, dp_rate: float = 0.1, **kwargs):\n",
    "        \"\"\"\n",
    "        Initialize the Elliptical class.\n",
    "\n",
    "        Args:\n",
    "            c_in (int): Number of input channels.\n",
    "            c_hidden (int): Number of hidden channels.\n",
    "            c_out (int): Number of output channels.\n",
    "            num_layers (int, optional): Number of GNN layers. Defaults to 5.\n",
    "            dp_rate (float, optional): Dropout rate. Defaults to 0.1.\n",
    "            **kwargs: Additional keyword arguments to be passed to the GNN layers.\n",
    "\n",
    "        Returns:\n",
    "            None\n",
    "        \"\"\"\n",
    "\n",
    "        super().__init__()\n",
    "        gnn_layer = geom_nn.GCNConv\n",
    "\n",
    "        layers = []\n",
    "        in_channels, out_channels = c_in, c_hidden\n",
    "        for _ in range(num_layers-1):\n",
    "            layers += [\n",
    "                gnn_layer(in_channels=in_channels,\n",
    "                          out_channels=out_channels,\n",
    "                          **kwargs),\n",
    "                nn.ReLU(inplace=True),\n",
    "                nn.Dropout(dp_rate)\n",
    "            ]\n",
    "            in_channels = c_hidden\n",
    "        layers += [gnn_layer(in_channels=in_channels,\n",
    "                             out_channels=c_out,\n",
    "                             **kwargs)]\n",
    "        self.layers = nn.ModuleList(layers)\n",
    "\n",
    "    def forward(self, x: Tensor, edge_index: Tensor) -> Tensor:\n",
    "        \"\"\"\n",
    "        Forward pass of the model.\n",
    "\n",
    "        Args:\n",
    "            x (Tensor): Input tensor.\n",
    "            edge_index (Tensor): Edge index tensor.\n",
    "\n",
    "        Returns:\n",
    "            Tensor: Output tensor.\n",
    "        \"\"\"\n",
    "\n",
    "        for l in self.layers:\n",
    "            if isinstance(l, geom_nn.MessagePassing):\n",
    "                # In case of a geom layer, also pass the edge_index list\n",
    "                x = l(x, edge_index)\n",
    "            else:\n",
    "                x = l(x)\n",
    "\n",
    "        return x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following code builds on the GNN model by turning it into a PyTorch Lightning module for training and evaluation. We define a GNN class that trains the model to predict the class of individual nodes in a graph. During training, we calculate accuracy and loss on a subset of nodes (masks) and update the model weights to minimize the loss. We also track the accuracy on validation and test sets to monitor the model's performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torchmetrics\n",
    "import lightning.pytorch as pl\n",
    "\n",
    "class GNN(pl.LightningModule):\n",
    "\n",
    "    def __init__(self, **model_kwargs):\n",
    "        super().__init__()\n",
    "        self.save_hyperparameters()\n",
    "\n",
    "        self.accuracy = torchmetrics.classification.Accuracy(task=\"multiclass\", num_classes=3)\n",
    "\n",
    "        self.model = GNNModel(**model_kwargs)\n",
    "        self.loss_module = nn.CrossEntropyLoss()\n",
    "\n",
    "    def forward(self, data, mode=\"train\"):\n",
    "        x, edge_index = data.x, data.edge_index\n",
    "        x = self.model(x, edge_index)\n",
    "\n",
    "        # Get the mask \n",
    "        if mode == \"train\":\n",
    "            mask = data.train_mask\n",
    "        elif mode == \"val\":\n",
    "            mask = data.val_mask\n",
    "        elif mode == \"test\":\n",
    "            mask = data.test_mask\n",
    "        else:\n",
    "            assert False, f\"Unknown forward mode: {mode}\"\n",
    "\n",
    "        # Calculate the loss for the mask\n",
    "        loss = self.loss_module(x[mask], data.y[mask].long())\n",
    "        pred = x[mask].argmax(dim=-1)\n",
    "        \n",
    "        return loss, pred, data.y[mask]\n",
    "\n",
    "    def configure_optimizers(self):\n",
    "        optimizer = optim.Adam(self.parameters(), lr=0.05)\n",
    "        return optimizer\n",
    "\n",
    "    def training_step(self, batch, batch_idx):\n",
    "        loss, preds, y = self.forward(batch, mode=\"train\")\n",
    "\n",
    "        # log step metric\n",
    "        self.accuracy(preds, y)\n",
    "        self.log('train_acc_step', self.accuracy)\n",
    "\n",
    "        self.log('train_loss', loss)\n",
    "        return loss\n",
    "\n",
    "    def validation_step(self, batch, batch_idx):\n",
    "        loss, preds, y = self.forward(batch, mode=\"val\")\n",
    "\n",
    "        # log step metric\n",
    "        self.accuracy(preds, y)\n",
    "        self.log('val_acc_step', self.accuracy)\n",
    "\n",
    "    def test_step(self, batch, batch_idx):\n",
    "        loss, preds, y = self.forward(batch, mode=\"test\")\n",
    "\n",
    "        # log step metric\n",
    "        self.accuracy(preds, y)\n",
    "        self.log('test_acc_step', self.accuracy)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following code sets up the training process for our model. It prepares the graph dataset for training using a DataLoader, creates the GNN model, configures a PyTorch Lightning trainer to manage the training process on a GPU or CPU, trains the model, tests its performance on a test set, and finally returns the trained GNN model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch_geometric.loader import DataLoader\n",
    "\n",
    "def train_node_classifier(dataset, **model_kwargs):\n",
    "    node_data_loader = DataLoader(dataset, batch_size = 1)\n",
    "\n",
    "    # Create a PyTorch Lightning trainer with the generation callback\n",
    "    trainer = pl.Trainer(accelerator=\"cpu\",\n",
    "                         devices=1,\n",
    "                         max_epochs=10,\n",
    "                         enable_progress_bar=False) # False because epoch size is 1\n",
    "    \n",
    "    # Note: the dimensions are specific for the Elliptic dataset\n",
    "    model = GNN(c_in=166, c_out=3, **model_kwargs)\n",
    "    trainer.fit(model, train_dataloaders=node_data_loader, val_dataloaders=node_data_loader)\n",
    "\n",
    "    # Test best model on the test set\n",
    "    trainer.test(model, node_data_loader, verbose=True)\n",
    "\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we call the code to train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU available: False, used: False\n",
      "TPU available: False, using: 0 TPU cores\n",
      "IPU available: False, using: 0 IPUs\n",
      "HPU available: False, using: 0 HPUs\n",
      "/home/vscode/.local/lib/python3.11/site-packages/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py:75: Starting from v1.9.0, `tensorboardX` has been removed as a dependency of the `pytorch_lightning` package, due to potential conflicts with other packages in the ML ecosystem. For this reason, `logger=True` will use `CSVLogger` as the default logger, unless the `tensorboard` or `tensorboardX` packages are found. Please `pip install lightning[extra]` or one of them to enable TensorBoard support by default\n",
      "\n",
      "  | Name        | Type               | Params\n",
      "---------------------------------------------------\n",
      "0 | accuracy    | MulticlassAccuracy | 0     \n",
      "1 | model       | GNNModel           | 3.5 K \n",
      "2 | loss_module | CrossEntropyLoss   | 0     \n",
      "---------------------------------------------------\n",
      "3.5 K     Trainable params\n",
      "0         Non-trainable params\n",
      "3.5 K     Total params\n",
      "0.014     Total estimated model params size (MB)\n",
      "/home/vscode/.local/lib/python3.11/site-packages/pytorch_lightning/trainer/connectors/data_connector.py:441: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=9` in the `DataLoader` to improve performance.\n",
      "/home/vscode/.local/lib/python3.11/site-packages/pytorch_lightning/trainer/connectors/data_connector.py:441: The 'train_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=9` in the `DataLoader` to improve performance.\n",
      "/home/vscode/.local/lib/python3.11/site-packages/pytorch_lightning/loops/fit_loop.py:298: The number of training batches (1) is smaller than the logging interval Trainer(log_every_n_steps=50). Set a lower value for log_every_n_steps if you want to see logs for the training epoch.\n",
      "`Trainer.fit` stopped: `max_epochs=10` reached.\n",
      "/home/vscode/.local/lib/python3.11/site-packages/pytorch_lightning/trainer/connectors/data_connector.py:441: The 'test_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=9` in the `DataLoader` to improve performance.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────\n",
      "       Test metric             DataLoader 0\n",
      "────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────\n",
      "      test_acc_step         0.7706350088119507\n",
      "────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────\n"
     ]
    }
   ],
   "source": [
    "node_gnn_model = train_node_classifier(dataset=elliptic_dataset,\n",
    "                                                        c_hidden=16,\n",
    "                                                        num_layers=5,\n",
    "                                                        dp_rate=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The training ran on a GPU or CPU and successfully trained your GNN model with 3.5k parameters. It reached the maximum of 10 epochs. On the test data, the model achieved an accuracy of around 80%. Note that this accuracy level is misleading! \n",
    "Let's check what it actually predicts with the following code:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape: (2, 3)\n",
      "┌─────┬────────┬──────────────────┐\n",
      "│     ┆ count  ┆ count_normalized │\n",
      "│ --- ┆ ---    ┆ ---              │\n",
      "│ i64 ┆ u32    ┆ f64              │\n",
      "╞═════╪════════╪══════════════════╡\n",
      "│ 2   ┆ 174605 ┆ 0.856877         │\n",
      "│ 0   ┆ 29164  ┆ 0.143123         │\n",
      "└─────┴────────┴──────────────────┘\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import polars as pol\n",
    "\n",
    "output = node_gnn_model.model(elliptic_dataset.x, elliptic_dataset.edge_index)\n",
    "# Assuming your tensor is named 'tensor'\n",
    "argmax_tensor = torch.argmax(output, dim=1)\n",
    "\n",
    "# Convert elliptic_dataset.y to a numpy array and then to a polars Series\n",
    "y_series = pol.Series(argmax_tensor.numpy())\n",
    "\n",
    "# Calculate the fraction of each value in the distribution\n",
    "fraction = y_series.value_counts() \n",
    "# Normalize the count column in fraction\n",
    "fraction = fraction.with_columns(count_normalized = fraction['count'] / y_series.shape[0])\n",
    "\n",
    "# Print the fraction of the value distribution\n",
    "print(fraction)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model almost only predicts a single class and achieves high levels of accuracy!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The Actor-Transaction graph\n",
    "Here we discuss the Elliptic++ dataset [GitHub Elliptic++ Dataset](https://github.com/git-disl/EllipticPlusPlus/tree/main?tab=readme-ov-file). This includes transactions and wallets, where the wallet/actor can be illicit, licit, or unknown status. Each actor sends transactions in the networks, some of which are illicit, licit, or unknown status.\n",
    "Compared to the previous graph, we know have the actors/wallets as a new set of nodes in the graph structure. This results in a Bipartite graph which we want to analyze using Pytorch Geometric. Therefore, the dataset returned is of type HeteroData, and here we show how to build a simple Graph Neural Network for HeteroData. Let's first discuss the features of the actor/wallet dataset:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Actor features:\n",
    "\n",
    "| Feature | Description |\n",
    "|---|---|\n",
    "| **Transaction related** | |\n",
    "| BTCtransacted | Total BTC transacted (sent+received) |\n",
    "| BTCsent | Total BTC sent |\n",
    "| BTCreceived | Total BTC received |\n",
    "| Fees | Total fees in BTC |\n",
    "| Feesshare | Total fees as share of BTC transacted |\n",
    "| **Time related** | |\n",
    "| Blockstxs | Number of blocks between transactions |\n",
    "| Blocksinput | Number of blocks between being an input address |\n",
    "| Blocksoutput | Number of blocks between being an output address |\n",
    "| Addr interactions | Number of interactions among addresses<br>5 values: total, min, max, mean, median |\n",
    "| Class | Class label: {illicit, licit, unknown} |\n",
    "| **Transaction related** | |\n",
    "| Txstotal | Total number of blockchain transactions |\n",
    "| TxSinput | Total number of dataset transactions as input address |\n",
    "| TxSoutput | Total number of dataset transactions as output address |\n",
    "| **Time related** | |\n",
    "| Timesteps | Number of time steps transacting in |\n",
    "| Lifetime | Lifetime in blocks |\n",
    "| Block first | Block height first transacted in |\n",
    "| Blocklast | Block height last transacted in |\n",
    "| Block first sent | Block height first sent in |\n",
    "| Block first receive | Block height first received in |\n",
    "| Repeat interactions | Number of addresses transacted with multiple times<br>single value |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/vscode/.local/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start download from HuggingFace...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 7/7 [00:01<00:00,  6.72it/s]\n",
      "Processing...\n",
      "Done!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "HeteroData(\n",
       "  wallets={\n",
       "    x=[1268260, 55],\n",
       "    y=[1268260],\n",
       "    train_mask=[1268260],\n",
       "    val_mask=[1268260],\n",
       "    test_mask=[1268260],\n",
       "  },\n",
       "  transactions={\n",
       "    x=[203769, 182],\n",
       "    y=[203769],\n",
       "    train_mask=[203769],\n",
       "    val_mask=[203769],\n",
       "    test_mask=[203769],\n",
       "  },\n",
       "  (transactions, to, transactions)={ edge_index=[2, 234355] },\n",
       "  (transactions, to, wallets)={ edge_index=[2, 837124] },\n",
       "  (wallets, to, transactions)={ edge_index=[2, 477117] },\n",
       "  (wallets, to, wallets)={ edge_index=[2, 2868964] }\n",
       ")"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from fintorch.datasets import ellipticpp\n",
    "\n",
    "# Load the elliptic dataset\n",
    "actor_transaction_graph = ellipticpp.TransactionActorDataset('~/.fintorch_data', force_reload=True)\n",
    "actor_transaction_graph[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple hetero model\n",
    "Let's create a GGN as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(['wallets', 'transactions'], [('transactions', 'to', 'transactions'), ('transactions', 'to', 'wallets'), ('wallets', 'to', 'transactions'), ('wallets', 'to', 'wallets')])\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GraphModule(\n",
       "  (conv1): ModuleDict(\n",
       "    (transactions__to__transactions): SAGEConv((-1, -1), 64, aggr=mean)\n",
       "    (transactions__to__wallets): SAGEConv((-1, -1), 64, aggr=mean)\n",
       "    (wallets__to__transactions): SAGEConv((-1, -1), 64, aggr=mean)\n",
       "    (wallets__to__wallets): SAGEConv((-1, -1), 64, aggr=mean)\n",
       "  )\n",
       "  (conv2): ModuleDict(\n",
       "    (transactions__to__transactions): SAGEConv((-1, -1), 3, aggr=mean)\n",
       "    (transactions__to__wallets): SAGEConv((-1, -1), 3, aggr=mean)\n",
       "    (wallets__to__transactions): SAGEConv((-1, -1), 3, aggr=mean)\n",
       "    (wallets__to__wallets): SAGEConv((-1, -1), 3, aggr=mean)\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch_geometric.transforms as T\n",
    "from torch_geometric.datasets import OGB_MAG\n",
    "from torch_geometric.nn import SAGEConv, to_hetero\n",
    "\n",
    "\n",
    "data = actor_transaction_graph[0]\n",
    "\n",
    "class GNN(torch.nn.Module):\n",
    "    def __init__(self, hidden_channels, out_channels):\n",
    "        super().__init__()\n",
    "        self.conv1 = SAGEConv((-1, -1), hidden_channels)\n",
    "        self.conv2 = SAGEConv((-1, -1), out_channels)\n",
    "\n",
    "    def forward(self, x, edge_index):\n",
    "        x = self.conv1(x, edge_index)\n",
    "        x = self.conv2(x, edge_index)\n",
    "        return x\n",
    "\n",
    "\n",
    "model = GNN(hidden_channels=64, out_channels=3)\n",
    "model = to_hetero(model, data.metadata(), aggr='sum')\n",
    "print(data.metadata())\n",
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We define a Graph Neural Network (GNN) model capable of processing graphs. The GNN class we construct takes hidden and output feature dimensions as input and sets up a two-layer GNN architecture using SAGEConv layers. The forward method outlines how node features and graph connectivity information are processed through these layers to learn new feature representations for the nodes.\n",
    "\n",
    "Next, we instantiate this GNN and convert it to work with a heterogeneous graph using the to_hetero function. A heterogeneous graph contains different types of nodes (like wallets and transactions) and edges (representing relationships between them). This conversion likely creates multiple specialized SAGEConv layers within our model, one for each type of edge relationship in the graph. Finally, we inspect the graph structure (node and edge types) and the structure of our resulting GNN module, including its specialized convolutions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next we show a simple forward pass through the model using the HeteroData dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'wallets': tensor([[  -9480.3672,  -68641.8281,  -84252.0000],\n",
      "        [   1002.8320,  -82406.6406,  -88378.7812],\n",
      "        [   1002.8320,  -82406.6406,  -88378.7812],\n",
      "        ...,\n",
      "        [ -24063.7500,    7173.9922, -130682.9297],\n",
      "        [ -10134.7188,  -66146.9062,  -87580.1641],\n",
      "        [  61478.8516,  -96403.6094,   -3982.5977]]), 'transactions': tensor([[-6.8772e+04,  1.1524e+05, -1.4699e+04],\n",
      "        [-6.6411e+04,  1.1910e+05, -1.5014e+04],\n",
      "        [-7.8126e+04,  1.3920e+05, -8.8606e+03],\n",
      "        ...,\n",
      "        [-5.6223e-01,  1.8661e-01,  2.4042e-01],\n",
      "        [-3.8677e-02,  3.6173e-01,  1.0702e-01],\n",
      "        [ 1.1682e-01,  1.0893e-01,  1.8528e-01]])}\n"
     ]
    }
   ],
   "source": [
    "with torch.no_grad(): \n",
    "     out = model(data.x_dict, data.edge_index_dict)\n",
    "     print(out)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We start by using with torch.no_grad() to temporarily disable gradient calculations within the block of code that follows. This is a common technique for improving efficiency and reducing memory usage during inference when we don't need to update the model's parameters. Next, we pass our input data (data.x_dict and data.edge_index_dict) to the GNN model and store the result in the out variable. Finally, we print the out variable to inspect the output generated by the model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion\n",
    "In this hands-on Python tutorial, we've dived into an intriguing blockchain transaction dataset. We explored how to prepare this data for machine learning models using the powerful PyTorch Geometric and PyTorch Lightning frameworks.  By following these steps, you've built a solid foundation for training various machine learning models on this dataset. This opens doors to exciting applications like fraud detection, transaction pattern analysis, and even predicting blockchain network behavior. Remember, practice and exploration are key to mastering these techniques.  Feel free to revisit this dataset and try your hand at fraud detection or other applications that interest you!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References\n",
    "- Youssef Elmougy and Ling Liu. 2023. Demystifying Fraudulent Transactions and Illicit Nodes in the Bitcoin Network for Financial Forensics. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD ’23), August 6–10, 2023, Long Beach, CA, USA. ACM, New York, NY, USA, 12 pages. https://doi.org/10.1145/3580305.3599803"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}