modules/transformers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a href=\"https://colab.research.google.com/github/ekrell/ams_ai_shortcourse_2024/blob/main/modules/transformers.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Transformers Explained Step by Step:\n",
    "##### Notebook by Hamid Kamangir (University of California, Davis)\n",
    "\n",
    "\n",
    "- [Vaswani et al, 2017](https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf) [Attention Is All You Need]\n",
    "\n",
    "### I. Introduction\n",
    "    - Tokenization\n",
    "    - Embedding\n",
    "    - Self-Attention Mechanism\n",
    "    - Keys, Queries, and Values\n",
    "    - Multi-head Attention\n",
    "    - Self-Attention vs CNNs\n",
    "### II. Implementation\n",
    "### III. Examples\n",
    "    - Synthetic Data\n",
    "    - Coastal Fog Forecasting\n",
    "----------------------------------------------------------------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## I. Introduction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section, we will: \n",
    "- learn the theoretical background on how transformers work and which are the basic components.\n",
    "- code all the different blocks of transformers architecture from scratch using PyTorch. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Transformers explained step by step\n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://nlp.seas.harvard.edu/images/the-annotated-transformer_0_0.png\" alt=\"Alt text for the image\" width = \"700px\"/>\n",
    "        <figcaption>Attention is All You Need. Figure Credit: <a href=\"https://nlp.seas.harvard.edu/2018/04/03/attention.html\">Vaswani et al, 2017</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n",
    "\n",
    "<b><font size=\"4\">What is Attention?</font></b>  \n",
    "The Attention mechanism describes a weighted averages of (sequence) elements with the <b>weights dynamically</b> computed based on an input query and elements' keys.\n",
    "\n",
    "\n",
    "\n",
    "This is the diagram of the Transformers network presented in the `Attention is All You Need` paper. We will go through all the differet pieces of this network throughout this notebook. \n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://www.researchgate.net/publication/344197785/figure/fig2/AS:934416989843456@1599793779015/Transformer-model-architecture-described-in-Attention-Is-All-You-Need-6.ppm\" alt=\"Alt text for the image\" width = \"400px\"/>\n",
    "        <figcaption>Transformers Architecture. Figure Credit: <a href=\"https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf\">Vaswani et al, 2017</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tokenization \n",
    "Tokenization is the the first step in processing input data, including text, audio, image, video or 4D environmental data. <br> In processing text each token correspond roughly to words and punctuation, although a word can also be split into several tokens. <br> In processing image it can small patch of image, etc. \n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/kCoLZuG.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Word Tokenizer. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">link</a></figcaption>\n",
    "    </p>\n",
    "</figure>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Embedding \n",
    "The Second step is to associate each token with an <b> embedding </b>, which is nothing more than a vector of real numbers. \n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/NulRCFU.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Embedding. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n",
    "- The values inside an embedding carry information about the meaning of the token, but they are also arranged in such a way that one can perform mathematical operations on them, which correspond to semantic changes.\n",
    "- Embedding are associated with tokens, which means that the same token always gets the same embedding, regardless of its context. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Image Embedding\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://media.giphy.com/media/ATsWtUsuuFRfq8OhZ7/source.gif\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Vision Transformers [image embedding]. Figure Credit: <a href=\"https://theaisummer.com/hugging-face-vit/\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Self-Attention \n",
    "Self-attention is the center of Attention mechanism, which in particular it has four parts: \n",
    "- <b>Query</b>: The quary is a feature vector describes what are looking for in the sequence, i,e. what would we maybe to <b>pay attention to</b>.\n",
    "- <b>Keys</b>: For each each input element, we have a key which is roughly describes what the elements is <b>offering</b>. \n",
    "- <b>Values</b>: This feature vector is the one we want to <b>average over</b>.\n",
    "- <b>Score function</b>: To rate input elements we want to pay attention to, we need to specify a <b>score function</b>. The score function takes the query and a key as input, output the score/attention weight of the query-key pair. It is usually implemented by simple similarity matrices like a <b> dot product</b>. \n",
    "\n",
    "\n",
    "Query, keys and values are weights matrices, which are adjusted as a model parameters during training. \n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/lIhueb8.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Self-Attention. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n",
    "##### Scaled dot-product\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/qfFLyND.gif\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>scaled dot-product. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multi-Head Attention\n",
    "In addition, the same process can be repeated many times with different Key, Query, and Value projections, forming what is called a multi-head attention.\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/pH4NcnC.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Multi-Head Attention. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Self-Attention vs CNNs\n",
    "\n",
    "There are couple of drawbacks related to CNNs for environmental modeling: \n",
    "- CNNs are using pooling, e.g., they do not encode the position and orientation of object, Lack of location sensitivity (invariance of activities!)\n",
    "- CNNs are using fixed kernels for all input variables. \n",
    "- CNNs are limited to learn 4D/5D data structure. \n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://miro.medium.com/v2/resize:fit:1400/format:webp/1*pTu8CbnA_MzRbTh6Ia87hA.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>CNNs limitations. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On the other hand, Self-Attention:\n",
    "\n",
    "- Utilizes dynamic filters that are learned during the training process.\n",
    "- Self-attention mechanisms are sample-specific and are based on learning models.\n",
    "- They are capable of handling higher-dimensional data structures without limitations.\n",
    "- Enable learning processes to occur in parallel.\n",
    "\n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://ar5iv.labs.arxiv.org/html/1911.03584/assets/x1.png\" alt=\"Alt text for the image\" width = \"700px\"/>\n",
    "        <figcaption>Dynamic Filters and Self-attention Heads. Figure Credit: <a href=\"https://ar5iv.labs.arxiv.org/html/1911.03584\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://ar5iv.labs.arxiv.org/html/1911.03584/assets/plots/epochs_iso_layer_4_small.png\" alt=\"Alt text for the image\" width = \"700px\"/>\n",
    "        <figcaption>Lerning Local Features by Multi-Heads. Figure Credit: <a href=\"https://ar5iv.labs.arxiv.org/html/1911.03584\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## II. Implementation (Pytorch)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Importing Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import time \n",
    "import torch.optim as optim\n",
    "import math\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. MultiHead Attention"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class MultiHeadAttention(nn.Module):\n",
    "    def __init__(self, embed_size, heads):\n",
    "        \"\"\"\n",
    "        Initializes the MultiHeadAttention module. This module is a key component in Transformer architectures,\n",
    "        allowing the model to jointly attend to information from different representation subspaces.\n",
    "\n",
    "        Args:\n",
    "        embed_size (int): Size of the input embeddings.\n",
    "        heads (int): Number of attention heads.\n",
    "\n",
    "        The embeddings are split into 'heads' number of chunks, and each chunk is processed by a different attention mechanism.\n",
    "        \"\"\"\n",
    "        super(MultiHeadAttention, self).__init__()\n",
    "        self.embed_size = embed_size # Typically 512 in a Transformer model.\n",
    "        self.heads = heads # Number of attention heads, usually 8.\n",
    "        self.head_dim = embed_size // heads # Dimension of each head, e.g., 64.\n",
    "\n",
    "        # Ensure that the embedding size is divisible by the number of heads.\n",
    "        assert (\n",
    "            self.head_dim * self.heads == embed_size\n",
    "        ), \"Embedding size must be divisible by the number of heads.\"\n",
    "\n",
    "        # Linear layers to project the input embeddings into Query, Key, and Value spaces.\n",
    "        # We project embeddings to the same size and then split them into different heads.\n",
    "        self.values = nn.Linear(embed_size, embed_size)\n",
    "        self.keys = nn.Linear(embed_size, embed_size)\n",
    "        self.queries = nn.Linear(embed_size, embed_size)\n",
    "        self.fc_out = nn.Linear(embed_size, embed_size) # Final linear layer.\n",
    "\n",
    "    def forward(self, values, keys, query, mask):\n",
    "        \"\"\"\n",
    "        Forward pass of the MultiHeadAttention layer.\n",
    "\n",
    "        Args:\n",
    "        values (Tensor): Value tensor.\n",
    "        keys (Tensor): Key tensor.\n",
    "        query (Tensor): Query tensor.\n",
    "        mask (Tensor): Tensor to mask certain positions (e.g., for padding).\n",
    "\n",
    "        Each of these tensors has the shape (batch_size, sequence_length, embedding_size).\n",
    "        \"\"\"\n",
    "        batch_size = query.shape[0]\n",
    "        value_len, key_len, query_len = values.shape[1], keys.shape[1], query.shape[1]\n",
    "\n",
    "        # Pass through the linear layers.\n",
    "        values = self.values(values)\n",
    "        keys = self.keys(keys)\n",
    "        queries = self.queries(query)\n",
    "\n",
    "        # Split the embeddings into 'heads' number of chunks.\n",
    "        values = values.reshape(batch_size, value_len, self.heads, self.head_dim)\n",
    "        keys = keys.reshape(batch_size, key_len, self.heads, self.head_dim)\n",
    "        queries = queries.reshape(batch_size, query_len, self.heads, self.head_dim)\n",
    "\n",
    "        # Calculate attention scores using matrix multiplication.\n",
    "        energy = torch.einsum(\"nqhd,nkhd->nhqk\", [queries, keys])\n",
    "\n",
    "        # Apply mask to the attention scores, if provided.\n",
    "        if mask is not None:\n",
    "            energy = energy.masked_fill(mask == 0, float(\"-1e20\"))\n",
    "\n",
    "        # Apply softmax to normalize the attention scores.\n",
    "        attention = torch.softmax(energy / (self.embed_size ** (1 / 2)), dim=3)\n",
    "\n",
    "        # Multiply attention scores with values and combine heads.\n",
    "        out = torch.einsum(\"nhql,nlhd->nqhd\", [attention, values]).reshape(\n",
    "            batch_size, query_len, self.heads * self.head_dim\n",
    "        )\n",
    "\n",
    "        # Final linear layer to produce the output.\n",
    "        out = self.fc_out(out)\n",
    "        return out\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Attention(nn.Module):\n",
    "\n",
    "    def __init__(\n",
    "            self,\n",
    "            embed_size, heads\n",
    "    ):\n",
    "        super().__init__()\n",
    "        self.embed_size = embed_size\n",
    "        self.heads = heads\n",
    "        qkv_bias = False,\n",
    "        qk_norm  = False,\n",
    "\n",
    "        assert self.embed_size % self.heads == 0, 'dim should be divisible by num_heads'\n",
    "   \n",
    "        self.head_dim = self.embed_size // self.heads\n",
    "        self.scale = self.head_dim ** -0.5\n",
    "\n",
    "\n",
    "        self.qkv = nn.Linear(self.embed_size, self.embed_size * 3, bias=qkv_bias)\n",
    "        self.q_norm = nn.LayerNorm(self.head_dim) if qk_norm else nn.Identity()\n",
    "        self.k_norm = nn.LayerNorm(self.head_dim) if qk_norm else nn.Identity()\n",
    "        self.attn_drop = nn.Dropout(0.1)\n",
    "        self.proj = nn.Linear(self.embed_size, self.embed_size)\n",
    "        self.proj_drop = nn.Dropout(0.1)\n",
    "\n",
    "\n",
    "    #     self._init_weights()\n",
    "\n",
    "    # def _init_weights(self):\n",
    "    #     nn.init.xavier_uniform_(self.proj.weight)\n",
    "    #     nn.init.xavier_uniform_(self.qkv.weight)\n",
    "    #     nn.init.normal_(self.proj.bias, std=1)\n",
    "    #     nn.init.normal_(self.qkv.bias, std=1)\n",
    "\n",
    "\n",
    "    def forward(self, x):\n",
    "        B, N, C = x.shape\n",
    "        qkv = self.qkv(x).reshape(B, N, 3, self.heads, self.head_dim).permute(2, 0, 3, 1, 4)\n",
    "        q, k, v = qkv.unbind(0)\n",
    "        q, k = self.q_norm(q), self.k_norm(k)\n",
    "\n",
    "\n",
    "        q = q * self.scale\n",
    "        attn = q @ k.transpose(-2, -1)\n",
    "        weights = attn.softmax(dim=-1)\n",
    "        attn = self.attn_drop(weights)\n",
    "        x = attn @ v\n",
    "\n",
    "        x = x.transpose(1, 2).reshape(B, N, C)\n",
    "        x = self.proj(x)\n",
    "        x = self.proj_drop(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encoder Layer\n",
    "\n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/vUOhpoC.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Encoder Layer. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TransformerLayer(nn.Module):\n",
    "    def __init__(self, embed_size=512, heads=8, dropout=0.0, forward_expansion=4):\n",
    "        \"\"\"\n",
    "        Initializes a Transformer Layer used in the Transformer model. This layer consists of a MultiHeadAttention mechanism,\n",
    "        followed by layer normalization, a feed-forward network, and another layer normalization.\n",
    "\n",
    "        Args:\n",
    "        embed_size (int): Size of the embedding.\n",
    "        heads (int): Number of attention heads.\n",
    "        dropout (float): Dropout rate.\n",
    "        forward_expansion (int): Expansion factor for the feed-forward network.\n",
    "        \"\"\"\n",
    "        super(TransformerLayer, self).__init__()\n",
    "        self.attention = MultiHeadAttention(embed_size, heads) # Multi-head attention mechanism.\n",
    "        self.norm1 = nn.LayerNorm(embed_size) # Layer normalization.\n",
    "        self.norm2 = nn.LayerNorm(embed_size) # Another layer normalization.\n",
    "        \n",
    "        # Feed-forward network.\n",
    "        self.feed_forward = nn.Sequential(\n",
    "            nn.Linear(embed_size, forward_expansion * embed_size), # Expands the embedding.\n",
    "            nn.ReLU(), # Activation function.\n",
    "            nn.Linear(forward_expansion * embed_size, embed_size), # Projects back to embedding size.\n",
    "        )\n",
    "        self.dropout = nn.Dropout(dropout) # Dropout layer.\n",
    "\n",
    "    def forward(self, value, key, query, mask):\n",
    "        \"\"\"\n",
    "        Forward pass of the Transformer Layer.\n",
    "\n",
    "        Args:\n",
    "        value (Tensor): Value tensor.\n",
    "        key (Tensor): Key tensor.\n",
    "        query (Tensor): Query tensor.\n",
    "        mask (Tensor): Mask tensor for attention.\n",
    "\n",
    "        Each tensor has the shape (batch_size, sequence_length, embedding_size).\n",
    "        \"\"\"\n",
    "        # Apply the MultiHeadAttention mechanism.\n",
    "        attention = self.attention(value, key, query, mask)\n",
    "\n",
    "        # Apply the first layer normalization and dropout with skip connection.\n",
    "        x = self.dropout(self.norm1(attention + query))\n",
    "\n",
    "        # Pass through the feed-forward network.\n",
    "        forward = self.feed_forward(x)\n",
    "\n",
    "        # Apply the second layer normalization and dropout with skip connection.\n",
    "        out = self.dropout(self.norm2(forward + x))\n",
    "\n",
    "        return out\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. Encoder \n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/rbEe0lW.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Encoder. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Encoder(nn.Module):\n",
    "    def __init__(\n",
    "        self, \n",
    "        embed_size: int, \n",
    "        num_layers: int, \n",
    "        heads: int,\n",
    "        device: str, \n",
    "        forward_expansion: int, \n",
    "        dropout: float,\n",
    "        src_vocab_size: int, \n",
    "        max_length: int\n",
    "    ) -> None:\n",
    "        \"\"\"\n",
    "        Initializes the Encoder part of the Transformer model. This encoder processes the input sequence\n",
    "        using a series of Transformer layers.\n",
    "\n",
    "        Args:\n",
    "        embed_size (int): Size of the input embeddings.\n",
    "        num_layers (int): Number of Transformer layers in the encoder.\n",
    "        heads (int): Number of attention heads.\n",
    "        device (str): Device to run the model on ('cuda' or 'cpu').\n",
    "        forward_expansion (int): Expansion factor for the feed-forward network in Transformer layers.\n",
    "        dropout (float): Dropout rate.\n",
    "        src_vocab_size (int): Size of the source vocabulary.\n",
    "        max_length (int): Maximum length of input sequences.\n",
    "        \"\"\"\n",
    "        super(Encoder, self).__init__()\n",
    "        self.embed_size = embed_size\n",
    "        self.device = device\n",
    "\n",
    "        # Embedding layer for input words.\n",
    "        self.word_embedding = nn.Embedding(src_vocab_size, embed_size)\n",
    "\n",
    "        # Positional embedding layer for word positions.\n",
    "        self.position_embedding = nn.Embedding(max_length, embed_size)\n",
    "\n",
    "        # Stack of Transformer layers.\n",
    "        self.layers = nn.ModuleList(\n",
    "            [TransformerLayer(embed_size, heads, dropout, forward_expansion) for _ in range(num_layers)]\n",
    "        )\n",
    "        self.dropout = nn.Dropout(dropout)\n",
    "\n",
    "    def forward(self, x, mask):\n",
    "        \"\"\"\n",
    "        Forward pass through the encoder.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Input sequence of word indices. Shape: (batch_size, sequence_length).\n",
    "        mask (Tensor): Mask tensor to ignore certain positions (e.g., padding).\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Output tensor after processing with Transformer layers. Shape: (batch_size, sequence_length, embed_size).\n",
    "        \"\"\"\n",
    "        batch_size, seq_length = x.shape\n",
    "\n",
    "        # Generate position indices and send to the same device as input.\n",
    "        positions = torch.arange(0, seq_length).expand(batch_size, seq_length).to(self.device)\n",
    "\n",
    "        # Combine word and position embeddings with dropout.\n",
    "        out = self.dropout(self.word_embedding(x) + self.position_embedding(positions))\n",
    "\n",
    "        # Pass through each Transformer layer.\n",
    "        for layer in self.layers:\n",
    "            out = layer(out, out, out, mask)  # In the encoder, Q, K, V are all the same.\n",
    "\n",
    "        return out\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Decoder Layer \n",
    "<figure>\n",
    "    <p align=\"center\">\n",
    "        <img src=\"https://i.imgur.com/yV18zvn.png\" alt=\"Alt text for the image\"/>\n",
    "        <figcaption>Decoder Layer. Figure Credit: <a href=\"https://www.kaggle.com/code/alejopaullier/introduction-to-transformers\">Link</a></figcaption>\n",
    "    </p>\n",
    "</figure>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DecoderLayer(nn.Module):\n",
    "    def __init__(self, embed_size, heads, forward_expansion, dropout, device):\n",
    "        \"\"\"\n",
    "        Initializes a Decoder Layer used in the Transformer model. This layer consists of a MultiHeadAttention mechanism \n",
    "        focused on the target sequence, followed by another MultiHeadAttention mechanism that takes into account \n",
    "        the output of the encoder. Finally, it applies a feed-forward network.\n",
    "\n",
    "        Args:\n",
    "        embed_size (int): Size of the input embeddings.\n",
    "        heads (int): Number of attention heads.\n",
    "        forward_expansion (int): Expansion factor for the feed-forward network.\n",
    "        dropout (float): Dropout rate.\n",
    "        device (str): Device to run the model on ('cuda' or 'cpu').\n",
    "        \"\"\"\n",
    "        super(DecoderLayer, self).__init__()\n",
    "        self.norm = nn.LayerNorm(embed_size) # Layer normalization.\n",
    "        self.attention = MultiHeadAttention(embed_size, heads=heads) # Self-attention for the decoder.\n",
    "        \n",
    "        # Transformer block (Self-attention + feed-forward).\n",
    "        self.transformer_block = TransformerLayer(\n",
    "            embed_size, heads, dropout, forward_expansion\n",
    "        )\n",
    "        self.dropout = nn.Dropout(dropout) # Dropout layer.\n",
    "\n",
    "    def forward(self, x, value, key, src_mask, trg_mask):\n",
    "        \"\"\"\n",
    "        Forward pass through the decoder layer.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Target sequence input. Shape: (batch_size, target_sequence_length, embedding_size).\n",
    "        value (Tensor): Value tensor from the encoder output.\n",
    "        key (Tensor): Key tensor from the encoder output.\n",
    "        src_mask (Tensor): Mask for source sequence to ignore padding.\n",
    "        trg_mask (Tensor): Mask for target sequence for causal attention.\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Output after processing with the decoder layer. Shape: (batch_size, target_sequence_length, embedding_size).\n",
    "        \"\"\"\n",
    "        # Apply self-attention to the target sequence.\n",
    "        attention = self.attention(x, x, x, trg_mask)\n",
    "\n",
    "        # Combine the attention output with the original input and apply dropout and normalization.\n",
    "        query = self.dropout(self.norm(attention + x))\n",
    "\n",
    "        # Pass through the Transformer block with encoder's key and value.\n",
    "        out = self.transformer_block(value, key, query, src_mask)\n",
    "        \n",
    "        return out\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5. Decoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Decoder(nn.Module):\n",
    "    def __init__(self, embed_size, num_layers, heads, forward_expansion,\n",
    "                 dropout, device, trg_vocab_size, max_length):\n",
    "        \"\"\"\n",
    "        Initializes the Decoder part of the Transformer model. This decoder takes the output from the Transformer's encoder\n",
    "        and generates the target sequence.\n",
    "\n",
    "        Args:\n",
    "        trg_vocab_size (int): Size of the target vocabulary.\n",
    "        embed_size (int): Size of the embeddings.\n",
    "        num_layers (int): Number of DecoderLayer modules in the Decoder.\n",
    "        heads (int): Number of attention heads in each DecoderLayer.\n",
    "        forward_expansion (int): Expansion factor for the feed-forward network in each DecoderLayer.\n",
    "        dropout (float): Dropout rate.\n",
    "        device (str): Device the model is running on ('cuda' or 'cpu').\n",
    "        max_length (int): Maximum length of the target sequence.\n",
    "        \"\"\"\n",
    "        super(Decoder, self).__init__()\n",
    "        self.device = device\n",
    "\n",
    "        # Embedding layers for the target words and their positions.\n",
    "        self.word_embedding = nn.Embedding(trg_vocab_size, embed_size)\n",
    "        self.position_embedding = nn.Embedding(max_length, embed_size)\n",
    "\n",
    "        # Stack of DecoderLayer modules.\n",
    "        self.layers = nn.ModuleList(\n",
    "            [DecoderLayer(embed_size, heads, forward_expansion, dropout, device) for _ in range(num_layers)]\n",
    "        )\n",
    "\n",
    "        # Final linear layer to output probabilities over the target vocabulary.\n",
    "        self.fc_out = nn.Linear(embed_size, trg_vocab_size)\n",
    "        self.dropout = nn.Dropout(dropout)\n",
    "\n",
    "    def forward(self, x, enc_out, src_mask, trg_mask):\n",
    "        \"\"\"\n",
    "        Forward pass through the decoder.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Input sequence of target word indices. Shape: (batch_size, target_sequence_length).\n",
    "        enc_out (Tensor): Output from the encoder. Shape: (batch_size, source_sequence_length, embedding_size).\n",
    "        src_mask (Tensor): Mask for source sequence to ignore padding.\n",
    "        trg_mask (Tensor): Mask for target sequence to allow causal (autoregressive) decoding.\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Output tensor with logits for the next word prediction. Shape: (batch_size, target_sequence_length, trg_vocab_size).\n",
    "        \"\"\"\n",
    "        batch_size, seq_length = x.shape\n",
    "\n",
    "        # Generate position indices and send to the same device as input.\n",
    "        positions = torch.arange(0, seq_length).expand(batch_size, seq_length).to(self.device)\n",
    "\n",
    "        # Combine word and position embeddings with dropout.\n",
    "        x = self.dropout(self.word_embedding(x) + self.position_embedding(positions))\n",
    "\n",
    "        # Pass through each DecoderLayer, using encoder output as key and value.\n",
    "        for layer in self.layers:\n",
    "            x = layer(x, enc_out, enc_out, src_mask, trg_mask)\n",
    "\n",
    "        # Final linear layer to produce output logits.\n",
    "        out = self.fc_out(x)\n",
    "\n",
    "        return out\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## III. Examples "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A. Synthetic Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Text_Transformer(nn.Module):\n",
    "    def __init__(self, \n",
    "                 src_vocab_size: int = 0, \n",
    "                 trg_vocab_size: int = 0, \n",
    "                 src_pad_idx: int = 10, \n",
    "                 trg_pad_idx: int = 10, \n",
    "                 embed_size: int = 512,\n",
    "                 num_layers: int = 6, \n",
    "                 forward_expansion: int = 4, \n",
    "                 heads: int = 8, \n",
    "                 dropout: float = 0., \n",
    "                 device = \"cpu\", \n",
    "                 max_length: int =100):\n",
    "\n",
    "        super(Text_Transformer, self).__init__()\n",
    "        # === Encoder ===\n",
    "\n",
    "        self.encoder = Encoder(embed_size = embed_size, \n",
    "                               num_layers = num_layers, \n",
    "                               heads = heads, \n",
    "                               device = device, \n",
    "                               forward_expansion = forward_expansion, \n",
    "                               dropout = dropout, \n",
    "                               src_vocab_size = src_vocab_size, \n",
    "                               max_length = max_length)\n",
    "        # === Decoder ===\n",
    "        self.decoder = Decoder(embed_size = embed_size, \n",
    "                               num_layers = num_layers, \n",
    "                               heads = heads, \n",
    "                               forward_expansion = forward_expansion, \n",
    "                               dropout = dropout, \n",
    "                               device = device, \n",
    "                               trg_vocab_size = trg_vocab_size, \n",
    "                               max_length = max_length)\n",
    "        \n",
    "        self.src_pad_idx = src_pad_idx\n",
    "        self.trg_pad_idx = trg_pad_idx\n",
    "        self.device = device\n",
    "\n",
    "    def make_src_mask(self, src):\n",
    "        src_mask = (src != self.src_pad_idx).unsqueeze(1).unsqueeze(2)\n",
    "        # (N, 1, 1, src_len)\n",
    "        return src_mask.to(self.device)\n",
    "\n",
    "    def make_trg_mask(self, trg):\n",
    "        \"\"\"\n",
    "        Returns a lower triangular matrix filled with 1s. The shape of the mask is (target_size, target_size).\n",
    "        Example: for a target of shape (batch_size=1, target_size=4)\n",
    "        tensor([[[[1., 0., 0., 0.],\n",
    "                  [1., 1., 0., 0.],\n",
    "                  [1., 1., 1., 0.],\n",
    "                  [1., 1., 1., 1.]]]])\n",
    "        \"\"\"\n",
    "        N, trg_len = trg.shape\n",
    "        trg_mask = torch.tril(torch.ones((trg_len, trg_len))).expand(\n",
    "            N, 1, trg_len, trg_len\n",
    "        )\n",
    "        return trg_mask.to(self.device)\n",
    "\n",
    "    def forward(self, src, trg):\n",
    "        src_mask = self.make_src_mask(src) # src_mask shape: \n",
    "        trg_mask = self.make_trg_mask(trg) # trg_mask shape: \n",
    "        enc_src = self.encoder(src, src_mask) # enc_src shape:\n",
    "        out = self.decoder(trg, enc_src, src_mask, trg_mask) # out shape: \n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input shape: torch.Size([2, 9])\n",
      "Target shape: torch.Size([2, 8])\n",
      "Device available: cuda\n",
      "Output shape: torch.Size([2, 7, 10])\n",
      "Output: tensor([[[-0.0965,  0.2323, -0.1462,  0.2858, -0.2389, -0.8906, -0.2154,\n",
      "           0.3922, -0.0356, -0.6345],\n",
      "         [-0.3655,  0.0582,  0.0070, -0.0417, -0.1682, -0.7169,  0.1575,\n",
      "           0.5706,  0.8792, -0.2640],\n",
      "         [-0.0842,  0.0204,  0.5319, -0.4663, -0.6955,  0.2332, -0.0203,\n",
      "           1.2905, -0.0321, -0.4428],\n",
      "         [ 0.0527, -0.5582,  0.4036,  0.7214, -0.8287,  0.1015, -0.2411,\n",
      "           0.1389,  0.0895, -0.0966],\n",
      "         [ 0.2539,  0.3697,  0.1253,  0.7705, -0.8360, -0.2558, -0.4588,\n",
      "           0.6847,  0.2984, -0.5582],\n",
      "         [-0.0207,  0.4457,  0.9303,  0.1525, -0.4096, -0.7391, -0.4372,\n",
      "           0.7414,  0.5331,  0.1490],\n",
      "         [-0.4979, -0.3434,  0.8171,  0.1524, -0.5741,  0.6096, -0.3287,\n",
      "           0.6273, -0.0580, -1.2756]],\n",
      "\n",
      "        [[-0.1739,  0.2163, -0.0080,  0.1986, -0.3770, -0.8058, -0.0052,\n",
      "           0.5131,  0.0220, -0.6471],\n",
      "         [-0.2803,  0.4590, -0.1591, -0.0094, -0.9798, -0.6434, -0.3213,\n",
      "           0.8218,  1.2239, -0.5008],\n",
      "         [-0.3325,  0.5382,  0.3921, -0.4780,  0.3088,  0.5930, -0.0493,\n",
      "           0.7064,  0.0219, -0.4654],\n",
      "         [-0.4596, -0.2853,  0.4759,  0.5552, -0.6080,  0.6893,  0.6566,\n",
      "           0.6829,  0.1955, -0.8531],\n",
      "         [-0.0933, -0.2556,  0.7507,  0.7493, -0.8912, -0.4852,  0.1738,\n",
      "           1.0737, -0.1447, -0.2067],\n",
      "         [-0.2402,  0.0886,  0.2896, -0.3820,  0.3409, -0.3556,  0.2424,\n",
      "           0.8473,  0.8887,  0.1043],\n",
      "         [-0.2919,  0.1887,  0.4014,  0.0094,  0.3186,  0.1154, -0.2378,\n",
      "           0.3146, -0.3507, -0.4726]]], device='cuda:0',\n",
      "       grad_fn=<AddBackward0>)\n"
     ]
    }
   ],
   "source": [
    "x = torch.tensor([[1, 5, 6, 4, 3, 9, 5, 2, 0], [1, 8, 7, 3, 4, 5, 6, 7, 2]]).to(device)\n",
    "trg = torch.tensor([[1, 7, 4, 3, 5, 9, 2, 0], [1, 5, 6, 2, 4, 7, 6, 2]]).to(device)\n",
    "src_pad_idx = 0 # index of the padding token in source vocabulary\n",
    "trg_pad_idx = 0 # index of the padding token in target vocabulary\n",
    "src_vocab_size = 10 # number of unique tokens in source vocabulary\n",
    "trg_vocab_size = 10 # number of unique tokens in target vocabulary\n",
    "\n",
    "print(f\"Input shape: {x.shape}\")\n",
    "print(f\"Target shape: {trg.shape}\")\n",
    "print(f\"Device available: {device}\")\n",
    "\n",
    "model = Text_Transformer(src_vocab_size = 10, \n",
    "                         trg_vocab_size = 10, \n",
    "                         src_pad_idx = 0, \n",
    "                         trg_pad_idx = 0, \n",
    "                         embed_size=512,\n",
    "                         num_layers = 8, \n",
    "                         forward_expansion = 4, \n",
    "                         heads = 8, \n",
    "                         dropout = 0,\n",
    "                         device=device).to(device)\n",
    "\n",
    "\n",
    "out = model(x, trg[:, :-1])\n",
    "print(f\"Output shape: {out.shape}\")\n",
    "print(f\"Output: {out}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Practice: B. Coastal Fog Forecasting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Reading Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2024-01-28 07:25:41--  https://raw.github.com/ekrell/ams_ai_shortcourse_2024/main/data/nam_fog_data_2009_2020.csv\n",
      "Resolving raw.github.com (raw.github.com)... 185.199.110.133, 185.199.111.133, 185.199.108.133, ...\n",
      "Connecting to raw.github.com (raw.github.com)|185.199.110.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 301 Moved Permanently\n",
      "Location: https://raw.githubusercontent.com/ekrell/ams_ai_shortcourse_2024/main/data/nam_fog_data_2009_2020.csv [following]\n",
      "--2024-01-28 07:25:41--  https://raw.githubusercontent.com/ekrell/ams_ai_shortcourse_2024/main/data/nam_fog_data_2009_2020.csv\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.109.133, 185.199.110.133, 185.199.111.133, ...\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.109.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 18923812 (18M) [text/plain]\n",
      "Saving to: ‘nam_fog_data_2009_2020.csv’\n",
      "\n",
      "100%[======================================>] 18,923,812  98.6MB/s   in 0.2s   \n",
      "\n",
      "2024-01-28 07:25:41 (98.6 MB/s) - ‘nam_fog_data_2009_2020.csv’ saved [18923812/18923812]\n",
      "\n",
      "--2024-01-28 07:25:42--  https://raw.github.com/ekrell/ams_ai_shortcourse_2024/main/data/kdfw_processed_sub_data.csv\n",
      "Resolving raw.github.com (raw.github.com)... 185.199.111.133, 185.199.108.133, 185.199.109.133, ...\n",
      "Connecting to raw.github.com (raw.github.com)|185.199.111.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 301 Moved Permanently\n",
      "Location: https://raw.githubusercontent.com/ekrell/ams_ai_shortcourse_2024/main/data/kdfw_processed_sub_data.csv [following]\n",
      "--2024-01-28 07:25:42--  https://raw.githubusercontent.com/ekrell/ams_ai_shortcourse_2024/main/data/kdfw_processed_sub_data.csv\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.110.133, 185.199.111.133, 185.199.108.133, ...\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.110.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 4400996 (4.2M) [text/plain]\n",
      "Saving to: ‘kdfw_processed_sub_data.csv.1’\n",
      "\n",
      "100%[======================================>] 4,400,996   --.-K/s   in 0.1s    \n",
      "\n",
      "2024-01-28 07:25:42 (38.0 MB/s) - ‘kdfw_processed_sub_data.csv.1’ saved [4400996/4400996]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget https://raw.github.com/ekrell/ams_ai_shortcourse_2024/main/data/nam_fog_data_2009_2020.csv\n",
    "!wget https://raw.github.com/ekrell/ams_ai_shortcourse_2024/main/data/kdfw_processed_sub_data.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date_time</th>\n",
       "      <th>vis</th>\n",
       "      <th>vis_class</th>\n",
       "      <th>TMP_2maboveground_000</th>\n",
       "      <th>TMP_975mb_000</th>\n",
       "      <th>TMP_700mb_000</th>\n",
       "      <th>UGRD_10maboveground_000</th>\n",
       "      <th>VGRD_10maboveground_000</th>\n",
       "      <th>UGRD_975mb_000</th>\n",
       "      <th>VGRD_975mb_000</th>\n",
       "      <th>...</th>\n",
       "      <th>Q_700mb_024</th>\n",
       "      <th>RH_975mb_024</th>\n",
       "      <th>RH_700mb_024</th>\n",
       "      <th>DPT_2maboveground_024</th>\n",
       "      <th>Q_surface_024</th>\n",
       "      <th>RH_2maboveground_024</th>\n",
       "      <th>LCLT_024</th>\n",
       "      <th>VIS_surface_024</th>\n",
       "      <th>VVEL_975mb_024</th>\n",
       "      <th>VVEL_700mb_024</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20090102 06:05:00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.211253</td>\n",
       "      <td>-0.638208</td>\n",
       "      <td>-0.922310</td>\n",
       "      <td>0.327913</td>\n",
       "      <td>-0.284829</td>\n",
       "      <td>0.974154</td>\n",
       "      <td>0.444722</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.008090</td>\n",
       "      <td>-0.129521</td>\n",
       "      <td>-0.920972</td>\n",
       "      <td>-0.267073</td>\n",
       "      <td>-0.481775</td>\n",
       "      <td>1.580175</td>\n",
       "      <td>-0.146982</td>\n",
       "      <td>-3.094260</td>\n",
       "      <td>-1.038723</td>\n",
       "      <td>-0.645890</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20090102 12:05:00</td>\n",
       "      <td>0.25</td>\n",
       "      <td>1</td>\n",
       "      <td>-1.271569</td>\n",
       "      <td>-0.185072</td>\n",
       "      <td>-1.213175</td>\n",
       "      <td>0.946982</td>\n",
       "      <td>0.670577</td>\n",
       "      <td>2.525136</td>\n",
       "      <td>1.639543</td>\n",
       "      <td>...</td>\n",
       "      <td>0.843620</td>\n",
       "      <td>0.980444</td>\n",
       "      <td>1.609610</td>\n",
       "      <td>-0.320670</td>\n",
       "      <td>-0.525484</td>\n",
       "      <td>1.676559</td>\n",
       "      <td>-0.178767</td>\n",
       "      <td>0.575341</td>\n",
       "      <td>-2.816515</td>\n",
       "      <td>-7.357260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20090102 18:05:00</td>\n",
       "      <td>0.25</td>\n",
       "      <td>1</td>\n",
       "      <td>-1.130651</td>\n",
       "      <td>-0.411853</td>\n",
       "      <td>-1.283945</td>\n",
       "      <td>0.440011</td>\n",
       "      <td>0.300055</td>\n",
       "      <td>2.165181</td>\n",
       "      <td>1.432275</td>\n",
       "      <td>...</td>\n",
       "      <td>1.301478</td>\n",
       "      <td>0.029046</td>\n",
       "      <td>2.593725</td>\n",
       "      <td>-0.288417</td>\n",
       "      <td>-0.489981</td>\n",
       "      <td>1.676559</td>\n",
       "      <td>-0.145225</td>\n",
       "      <td>0.493484</td>\n",
       "      <td>2.116000</td>\n",
       "      <td>-2.291265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20090103 00:05:00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.201083</td>\n",
       "      <td>-0.410077</td>\n",
       "      <td>-0.973548</td>\n",
       "      <td>0.871125</td>\n",
       "      <td>0.485075</td>\n",
       "      <td>1.733437</td>\n",
       "      <td>1.374883</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.215393</td>\n",
       "      <td>0.663312</td>\n",
       "      <td>0.203731</td>\n",
       "      <td>-0.311617</td>\n",
       "      <td>-0.530781</td>\n",
       "      <td>1.580175</td>\n",
       "      <td>-0.190291</td>\n",
       "      <td>-3.010100</td>\n",
       "      <td>-4.162235</td>\n",
       "      <td>1.148516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20090103 06:05:00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.133033</td>\n",
       "      <td>-0.234904</td>\n",
       "      <td>-0.870015</td>\n",
       "      <td>-0.565579</td>\n",
       "      <td>0.749749</td>\n",
       "      <td>0.789537</td>\n",
       "      <td>2.080852</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.289343</td>\n",
       "      <td>0.425462</td>\n",
       "      <td>-1.249010</td>\n",
       "      <td>-0.226939</td>\n",
       "      <td>-0.414857</td>\n",
       "      <td>1.772943</td>\n",
       "      <td>-0.076921</td>\n",
       "      <td>-3.176563</td>\n",
       "      <td>-0.407135</td>\n",
       "      <td>-0.785856</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20200527 23:55:00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.575668</td>\n",
       "      <td>0.439906</td>\n",
       "      <td>-0.506560</td>\n",
       "      <td>0.400932</td>\n",
       "      <td>0.879720</td>\n",
       "      <td>0.646865</td>\n",
       "      <td>0.516139</td>\n",
       "      <td>...</td>\n",
       "      <td>0.104025</td>\n",
       "      <td>1.012428</td>\n",
       "      <td>-0.030226</td>\n",
       "      <td>0.743894</td>\n",
       "      <td>0.798028</td>\n",
       "      <td>0.568142</td>\n",
       "      <td>0.776823</td>\n",
       "      <td>0.519667</td>\n",
       "      <td>0.917959</td>\n",
       "      <td>2.133400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20200529 05:55:00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.764040</td>\n",
       "      <td>1.032830</td>\n",
       "      <td>0.549119</td>\n",
       "      <td>-0.643590</td>\n",
       "      <td>-0.064092</td>\n",
       "      <td>0.268097</td>\n",
       "      <td>-0.091787</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.526528</td>\n",
       "      <td>-1.213782</td>\n",
       "      <td>-0.545975</td>\n",
       "      <td>0.624009</td>\n",
       "      <td>0.617600</td>\n",
       "      <td>0.308649</td>\n",
       "      <td>0.640879</td>\n",
       "      <td>0.519667</td>\n",
       "      <td>-2.174338</td>\n",
       "      <td>1.302179</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20200529 11:55:00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.612734</td>\n",
       "      <td>0.891709</td>\n",
       "      <td>0.064682</td>\n",
       "      <td>1.030756</td>\n",
       "      <td>-0.637713</td>\n",
       "      <td>-0.165623</td>\n",
       "      <td>-0.604865</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.592482</td>\n",
       "      <td>-1.026576</td>\n",
       "      <td>-0.639495</td>\n",
       "      <td>0.456737</td>\n",
       "      <td>0.382482</td>\n",
       "      <td>-0.688927</td>\n",
       "      <td>0.390877</td>\n",
       "      <td>0.519667</td>\n",
       "      <td>-5.743924</td>\n",
       "      <td>0.370084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20200529 17:55:00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.688505</td>\n",
       "      <td>0.402944</td>\n",
       "      <td>-0.224222</td>\n",
       "      <td>0.568362</td>\n",
       "      <td>-1.231342</td>\n",
       "      <td>0.201695</td>\n",
       "      <td>-1.221644</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.117047</td>\n",
       "      <td>0.640958</td>\n",
       "      <td>-0.209543</td>\n",
       "      <td>0.542961</td>\n",
       "      <td>0.498543</td>\n",
       "      <td>-0.106504</td>\n",
       "      <td>0.524272</td>\n",
       "      <td>0.519667</td>\n",
       "      <td>-0.752231</td>\n",
       "      <td>-0.603256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20200529 23:55:00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.865073</td>\n",
       "      <td>0.554823</td>\n",
       "      <td>-0.723708</td>\n",
       "      <td>0.221239</td>\n",
       "      <td>-1.369482</td>\n",
       "      <td>0.376732</td>\n",
       "      <td>-1.368729</td>\n",
       "      <td>...</td>\n",
       "      <td>1.962647</td>\n",
       "      <td>0.681385</td>\n",
       "      <td>2.360183</td>\n",
       "      <td>0.407376</td>\n",
       "      <td>0.310182</td>\n",
       "      <td>-0.057418</td>\n",
       "      <td>0.395570</td>\n",
       "      <td>-5.650248</td>\n",
       "      <td>3.530415</td>\n",
       "      <td>-2.853806</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>11016 rows × 91 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            date_time    vis  vis_class  TMP_2maboveground_000  TMP_975mb_000  \\\n",
       "0   20090102 06:05:00   7.00          0              -1.211253      -0.638208   \n",
       "0   20090102 12:05:00   0.25          1              -1.271569      -0.185072   \n",
       "0   20090102 18:05:00   0.25          1              -1.130651      -0.411853   \n",
       "0   20090103 00:05:00   4.00          0              -1.201083      -0.410077   \n",
       "0   20090103 06:05:00   4.00          0              -1.133033      -0.234904   \n",
       "..                ...    ...        ...                    ...            ...   \n",
       "0   20200527 23:55:00  10.00          0               0.575668       0.439906   \n",
       "0   20200529 05:55:00  10.00          0               0.764040       1.032830   \n",
       "0   20200529 11:55:00  10.00          0               0.612734       0.891709   \n",
       "0   20200529 17:55:00  10.00          0               0.688505       0.402944   \n",
       "0   20200529 23:55:00  10.00          0               0.865073       0.554823   \n",
       "\n",
       "    TMP_700mb_000  UGRD_10maboveground_000  VGRD_10maboveground_000  \\\n",
       "0       -0.922310                 0.327913                -0.284829   \n",
       "0       -1.213175                 0.946982                 0.670577   \n",
       "0       -1.283945                 0.440011                 0.300055   \n",
       "0       -0.973548                 0.871125                 0.485075   \n",
       "0       -0.870015                -0.565579                 0.749749   \n",
       "..            ...                      ...                      ...   \n",
       "0       -0.506560                 0.400932                 0.879720   \n",
       "0        0.549119                -0.643590                -0.064092   \n",
       "0        0.064682                 1.030756                -0.637713   \n",
       "0       -0.224222                 0.568362                -1.231342   \n",
       "0       -0.723708                 0.221239                -1.369482   \n",
       "\n",
       "    UGRD_975mb_000  VGRD_975mb_000  ...  Q_700mb_024  RH_975mb_024  \\\n",
       "0         0.974154        0.444722  ...    -1.008090     -0.129521   \n",
       "0         2.525136        1.639543  ...     0.843620      0.980444   \n",
       "0         2.165181        1.432275  ...     1.301478      0.029046   \n",
       "0         1.733437        1.374883  ...    -0.215393      0.663312   \n",
       "0         0.789537        2.080852  ...    -1.289343      0.425462   \n",
       "..             ...             ...  ...          ...           ...   \n",
       "0         0.646865        0.516139  ...     0.104025      1.012428   \n",
       "0         0.268097       -0.091787  ...    -0.526528     -1.213782   \n",
       "0        -0.165623       -0.604865  ...    -0.592482     -1.026576   \n",
       "0         0.201695       -1.221644  ...    -0.117047      0.640958   \n",
       "0         0.376732       -1.368729  ...     1.962647      0.681385   \n",
       "\n",
       "    RH_700mb_024  DPT_2maboveground_024  Q_surface_024  RH_2maboveground_024  \\\n",
       "0      -0.920972              -0.267073      -0.481775              1.580175   \n",
       "0       1.609610              -0.320670      -0.525484              1.676559   \n",
       "0       2.593725              -0.288417      -0.489981              1.676559   \n",
       "0       0.203731              -0.311617      -0.530781              1.580175   \n",
       "0      -1.249010              -0.226939      -0.414857              1.772943   \n",
       "..           ...                    ...            ...                   ...   \n",
       "0      -0.030226               0.743894       0.798028              0.568142   \n",
       "0      -0.545975               0.624009       0.617600              0.308649   \n",
       "0      -0.639495               0.456737       0.382482             -0.688927   \n",
       "0      -0.209543               0.542961       0.498543             -0.106504   \n",
       "0       2.360183               0.407376       0.310182             -0.057418   \n",
       "\n",
       "    LCLT_024  VIS_surface_024  VVEL_975mb_024  VVEL_700mb_024  \n",
       "0  -0.146982        -3.094260       -1.038723       -0.645890  \n",
       "0  -0.178767         0.575341       -2.816515       -7.357260  \n",
       "0  -0.145225         0.493484        2.116000       -2.291265  \n",
       "0  -0.190291        -3.010100       -4.162235        1.148516  \n",
       "0  -0.076921        -3.176563       -0.407135       -0.785856  \n",
       "..       ...              ...             ...             ...  \n",
       "0   0.776823         0.519667        0.917959        2.133400  \n",
       "0   0.640879         0.519667       -2.174338        1.302179  \n",
       "0   0.390877         0.519667       -5.743924        0.370084  \n",
       "0   0.524272         0.519667       -0.752231       -0.603256  \n",
       "0   0.395570        -5.650248        3.530415       -2.853806  \n",
       "\n",
       "[11016 rows x 91 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.read_csv('nam_fog_data_2009_2020.csv', index_col=0)\n",
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Select columns from the 3rd to the last\n",
    "selected_columns = dataset.iloc[:, 3:]\n",
    "\n",
    "# Compute the correlation matrix\n",
    "correlation_matrix = selected_columns.corr()\n",
    "\n",
    "# Plotting the heatmap\n",
    "plt.figure(figsize=(12, 10))\n",
    "sns.heatmap(correlation_matrix, annot=False, fmt=\".2f\", cmap='coolwarm')\n",
    "plt.title('Correlation Matrix of Input Variables')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train dataset samples: 6692 | valid dataset samples: 2092 | test dataset samples: 2228\n"
     ]
    }
   ],
   "source": [
    "def split_dataframe_by_date(df, time_col, date_ranges):\n",
    "    \"\"\"\n",
    "    Splits a DataFrame into multiple DataFrames based on specified date ranges.\n",
    "\n",
    "    Args:\n",
    "    df (pd.DataFrame): The original DataFrame.\n",
    "    time_col (str): The name of the column containing time data.\n",
    "    date_ranges (list of tuples): A list of tuples where each tuple contains start and end dates as strings.\n",
    "                                  Example: [('2009-01-01', '2012-12-31'), ('2013-01-01', '2017-12-31')]\n",
    "\n",
    "    Returns:\n",
    "    list: A list containing the split DataFrames.\n",
    "    \"\"\"\n",
    "\n",
    "    # Convert the time column to datetime\n",
    "    df[time_col] = pd.to_datetime(df[time_col])\n",
    "\n",
    "    # Initialize an empty list to store the split DataFrames\n",
    "    split_dfs = []\n",
    "\n",
    "    # Loop through the date ranges and split the DataFrame\n",
    "    for start_date, end_date in date_ranges:\n",
    "        split_df = df[(df[time_col] >= start_date) & (df[time_col] <= end_date)]\n",
    "        split_dfs.append(split_df)\n",
    "\n",
    "    # Handling the remaining data for dates beyond the last specified range\n",
    "    if len(date_ranges) > 1:\n",
    "        last_end_date = date_ranges[-1][1]\n",
    "        remaining_df = df[df[time_col] > last_end_date]\n",
    "        split_dfs.append(remaining_df)\n",
    "\n",
    "    return split_dfs\n",
    "\n",
    "# Example usage\n",
    "date_ranges = [('2009-01-01', '2015-12-31'), ('2016-01-01', '2017-12-31')]\n",
    "train_df, valid_df, test_df = split_dataframe_by_date(dataset, 'date_time', date_ranges)\n",
    "print(f\"train dataset samples: {train_df.shape[0]} | valid dataset samples: {valid_df.shape[0]} | test dataset samples: {test_df.shape[0]}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.reset_index()\n",
    "valid_df = valid_df.reset_index()\n",
    "test_df = test_df.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a figure with two subplots\n",
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(18, 6))\n",
    "\n",
    "# Plotting the bar plot for visibility distribution with 10 bins\n",
    "bins = np.linspace(0, 10, 11)\n",
    "ax1.hist(train_df['vis'], bins=bins, alpha=0.5, label='Train')\n",
    "ax1.hist(valid_df['vis'], bins=bins, alpha=0.5, label='Validation')\n",
    "ax1.hist(test_df['vis'], bins=bins, alpha=0.5, label='Test')\n",
    "ax1.set_title('Visibility Distribution')\n",
    "ax1.set_xlabel('Visibility')\n",
    "ax1.set_ylabel('Count')\n",
    "ax1.legend()\n",
    "\n",
    "# Plotting the bar chart for fog/non-fog cases\n",
    "fog_counts = [\n",
    "    train_df['vis_class'].value_counts(),\n",
    "    valid_df['vis_class'].value_counts(),\n",
    "    test_df['vis_class'].value_counts()\n",
    "]\n",
    "fog_df = pd.DataFrame(fog_counts, index=['Train', 'Validation', 'Test'])\n",
    "fog_df.plot(kind='bar', ax=ax2)\n",
    "\n",
    "# Adding count annotations on the bars\n",
    "for p in ax2.patches:\n",
    "    ax2.annotate(str(p.get_height()), (p.get_x() + p.get_width() / 2., p.get_height()), \n",
    "                 ha='center', va='center', xytext=(0, 10), textcoords='offset points')\n",
    "\n",
    "ax2.set_title('Fog/Non-Fog Cases')\n",
    "ax2.set_xlabel('Dataset')\n",
    "ax2.set_ylabel('Count')\n",
    "\n",
    "# Display the plots\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "class dataloader(): \n",
    "\n",
    "    def __init__(self, dataframe):\n",
    "\n",
    "        self.dataframe = dataframe\n",
    "    def __len__(self):\n",
    "        return len(self.dataframe)\n",
    "\n",
    "    def __getitem__(self, idx): \n",
    "\n",
    "        input = self.dataframe.iloc[idx][3:]\n",
    "        input = torch.as_tensor(input, dtype=torch.float32)\n",
    "\n",
    "        label = self.dataframe.iloc[idx]['vis_class'] \n",
    "        label = torch.as_tensor(label) #, dtype= torch.long\n",
    "\n",
    "        sample = {\n",
    "            \"input\": input, \n",
    "            \"label\": label}\n",
    "\n",
    "        return sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = dataloader(train_df)\n",
    "valid_dataset = dataloader(valid_df)\n",
    "test_dataset = dataloader(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_loader_training = torch.utils.data.DataLoader(train_dataset, batch_size= 32, \n",
    "                                                shuffle = True,  num_workers=0) \n",
    "data_loader_validate = torch.utils.data.DataLoader(valid_dataset, batch_size= 32, \n",
    "                                                shuffle = False, num_workers=0) \n",
    "data_loader_testing  = torch.utils.data.DataLoader(test_dataset, batch_size= 32, \n",
    "                                                shuffle = False,  num_workers=0) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Embedding1D(nn.Module):\n",
    "    def __init__(self, ver_size, embed_size, dropout):\n",
    "        \"\"\"\n",
    "        Initializes a custom embedding module to handle 1D data (e.g., time series or any non-textual numerical data).\n",
    "\n",
    "        Args:\n",
    "        ver_size (int): The size of the variable (number of unique tokens or bins).\n",
    "        embed_size (int): The size of each embedding vector.\n",
    "        dropout (float): Dropout rate for regularization.\n",
    "\n",
    "        The module creates embeddings for discretized 1D data, adds positional embeddings, and includes a classification token ('cls_token').\n",
    "        \"\"\"\n",
    "        super(Embedding1D, self).__init__()\n",
    "\n",
    "        n_tokens = ver_size\n",
    "        # Define an embedding layer. \n",
    "        # This maps the discretized data to an embedding space.\n",
    "        self.embeddings = nn.Embedding(1000, embed_size)  # 1000 bins and 512 embedding size\n",
    "\n",
    "        # Positional embeddings for maintaining the order of the input.\n",
    "        self.position_embeddings = nn.Parameter(torch.zeros(1, n_tokens + 1, embed_size))\n",
    "        \n",
    "        # A special token for classification tasks (similar to BERT's [CLS] token).\n",
    "        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_size))\n",
    "        \n",
    "        # Dropout layer for regularization.\n",
    "        self.dropout = nn.Dropout(dropout)\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward pass of the Embedding1D module.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Input tensor of shape (batch_size, sequence_length).\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Output tensor after applying embeddings and adding positional information.\n",
    "        \"\"\"\n",
    "        # Discretize the float data into integer indices.\n",
    "        min_val = x.min()\n",
    "        max_val = x.max()\n",
    "        bins = np.linspace(min_val.item(), max_val.item(), 1000)\n",
    "\n",
    "        # Convert the values to indices based on the defined bins.\n",
    "        indices = np.digitize(x.cpu().numpy(), bins) - 1\n",
    "        indices = torch.LongTensor(indices).to(device)\n",
    "\n",
    "        # Apply the embedding layer.\n",
    "        x = self.embeddings(indices)\n",
    "\n",
    "        B = x.shape[0]\n",
    "        # Expand the cls_token to match the batch size and concatenate it with the input.\n",
    "        cls_tokens = self.cls_token.expand(B, -1, -1)\n",
    "        x = torch.cat((cls_tokens, x), dim=1)\n",
    "\n",
    "        # Add the positional embeddings and apply dropout.\n",
    "        embeddings = x + self.position_embeddings\n",
    "        embeddings = self.dropout(embeddings)\n",
    "\n",
    "        return embeddings\n",
    "\n",
    "\n",
    "class Encoder1D(nn.Module):\n",
    "    \"\"\"\n",
    "    The Encoder1D class is a custom encoder module designed for handling 1D data sequences. It's structured similarly to\n",
    "    the encoder part of the Transformer architecture but is adapted for 1D data. The module uses a custom embedding layer\n",
    "    for 1D data and a series of Transformer layers to process the sequence.\n",
    "    \"\"\"\n",
    "    def __init__(\n",
    "        self, \n",
    "        ver_size, \n",
    "        embed_size, \n",
    "        dropout,\n",
    "        num_layers=4, \n",
    "        expansion_factor=4, \n",
    "        heads=8\n",
    "    ) -> None:\n",
    "        \"\"\"\n",
    "        Initializes the Encoder1D module.\n",
    "\n",
    "        Args:\n",
    "        ver_size (int): Size of the variable (number of unique tokens or bins).\n",
    "        embed_size (int): Size of the embedding vectors.\n",
    "        dropout (float): Dropout rate.\n",
    "        num_layers (int): Number of Transformer layers in the encoder.\n",
    "        expansion_factor (int): Expansion factor for the feed-forward network within the Transformer layers.\n",
    "        heads (int): Number of heads in the multi-head attention mechanism.\n",
    "        \"\"\"\n",
    "        super(Encoder1D, self).__init__()\n",
    "\n",
    "        # Embedding layer specifically designed for 1D data.\n",
    "        self.embedding_layer = Embedding1D(ver_size, embed_size, dropout)\n",
    "\n",
    "        # Stack of Transformer layers.\n",
    "        self.layers = nn.ModuleList(\n",
    "            [\n",
    "                TransformerLayer(\n",
    "                    embed_size,\n",
    "                    heads,\n",
    "                    dropout=dropout,\n",
    "                    forward_expansion=expansion_factor,\n",
    "                )\n",
    "                for _ in range(num_layers)\n",
    "            ]\n",
    "        )\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward pass through the encoder.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Input tensor representing a sequence of 1D data.\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Output tensor after processing through the embedding layer and Transformer layers.\n",
    "        \"\"\"\n",
    "        # Process the input through the embedding layer.\n",
    "        x = self.embedding_layer(x)\n",
    "\n",
    "        # Pass the embedded input through each Transformer layer.\n",
    "        for layer in self.layers:\n",
    "            x = layer(x, x, x, None) # In the Encoder, the input serves as Q, K, V.\n",
    "\n",
    "        return x\n",
    "    \n",
    "\n",
    "class Fog_Transformer(nn.Module):\n",
    "    \"\"\"\n",
    "    The Fog_Transformer class is a neural network model designed for processing 1D data sequences, particularly aimed at\n",
    "    tasks like fog prediction. It is based on the Transformer architecture, utilizing a custom 1D data encoder.\n",
    "    \"\"\"\n",
    "    def __init__(self, ver_size, \n",
    "                 embed_size=512, \n",
    "                 dropout=0,\n",
    "                 num_layers=6, \n",
    "                 forward_expansion=4, \n",
    "                 heads=8, \n",
    "                 num_class=2):\n",
    "        \"\"\"\n",
    "        Initializes the Fog_Transformer model.\n",
    "\n",
    "        Args:\n",
    "        ver_size (int): Size of the variable (number of unique tokens or bins for 1D data).\n",
    "        embed_size (int): Size of the embedding vectors.\n",
    "        dropout (float): Dropout rate.\n",
    "        num_layers (int): Number of layers in the encoder.\n",
    "        forward_expansion (int): Expansion factor for the feed-forward network within the encoder layers.\n",
    "        heads (int): Number of heads in the multi-head attention mechanism in the encoder.\n",
    "        num_class (int): Number of output classes for the classification task.\n",
    "        \"\"\"\n",
    "        super(Fog_Transformer, self).__init__()\n",
    "\n",
    "        self.pool = None # Optional pooling strategy (e.g., 'mean' pooling)\n",
    "        \n",
    "        # Encoder module specifically designed for 1D data.\n",
    "        self.encoder = Encoder1D(\n",
    "            ver_size, \n",
    "            embed_size, \n",
    "            dropout,\n",
    "            num_layers, \n",
    "            heads, \n",
    "            forward_expansion\n",
    "        )\n",
    "        \n",
    "        # Classifier linear layer for the final output.\n",
    "        self.classifier = nn.Linear(embed_size, num_class)\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward pass through the Fog_Transformer model.\n",
    "\n",
    "        Args:\n",
    "        x (Tensor): Input tensor representing a sequence of 1D data.\n",
    "\n",
    "        Returns:\n",
    "        Tensor: Logits tensor after processing through the encoder and classifier. Shape: (batch_size, num_class).\n",
    "        \"\"\"\n",
    "        # Process the input through the encoder.\n",
    "        x = self.encoder(x)\n",
    "\n",
    "        # Apply pooling strategy (mean pooling or taking the first element).\n",
    "        x = x.mean(dim=1) if self.pool == 'mean' else x[:, 0]\n",
    "\n",
    "        # Pass through the classifier to get the final logits.\n",
    "        logits = self.classifier(x)\n",
    "\n",
    "        return logits\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Fog_Transformer(\n",
    "    ver_size = 89,\n",
    "    embed_size = 512, \n",
    "    dropout = 0.3,\n",
    "    num_layers = 6, \n",
    "    forward_expansion = 4, \n",
    "    heads = 8, \n",
    "    num_class = 2\n",
    "    ).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = optim.AdamW(model.parameters(), \n",
    "                        lr=1e-3, \n",
    "                        weight_decay=1e-2)\n",
    "loss_func  = torch.nn.BCEWithLogitsLoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 001: | Train Loss: 0.0979 | Val Loss: 0.0978 | Time(s): 46.811\n",
      "===================== Best model Saved! Val Loss: 0.0978 ====================\n",
      "Epoch 002: | Train Loss: 0.0892 | Val Loss: 0.1102 | Time(s): 48.212\n",
      "Model is not saved! Current Val Loss: 0.1102\n",
      "Epoch 003: | Train Loss: 0.0930 | Val Loss: 0.0976 | Time(s): 48.291\n",
      "===================== Best model Saved! Val Loss: 0.0976 ====================\n",
      "Epoch 004: | Train Loss: 0.0889 | Val Loss: 0.1007 | Time(s): 47.203\n",
      "Model is not saved! Current Val Loss: 0.1007\n",
      "Epoch 005: | Train Loss: 0.0906 | Val Loss: 0.0997 | Time(s): 47.274\n",
      "Model is not saved! Current Val Loss: 0.0997\n",
      "Epoch 006: | Train Loss: 0.0903 | Val Loss: 0.0986 | Time(s): 47.373\n",
      "Model is not saved! Current Val Loss: 0.0986\n",
      "Epoch 007: | Train Loss: 0.0893 | Val Loss: 0.0990 | Time(s): 47.487\n",
      "Model is not saved! Current Val Loss: 0.0990\n",
      "Epoch 008: | Train Loss: 0.0887 | Val Loss: 0.0985 | Time(s): 47.222\n",
      "Model is not saved! Current Val Loss: 0.0985\n",
      "Epoch 009: | Train Loss: 0.0898 | Val Loss: 0.1011 | Time(s): 47.338\n",
      "Model is not saved! Current Val Loss: 0.1011\n",
      "Epoch 010: | Train Loss: 0.0888 | Val Loss: 0.1014 | Time(s): 47.489\n",
      "Model is not saved! Current Val Loss: 0.1014\n",
      "Epoch 011: | Train Loss: 0.0879 | Val Loss: 0.1027 | Time(s): 47.255\n",
      "Model is not saved! Current Val Loss: 0.1027\n",
      "Epoch 012: | Train Loss: 0.0878 | Val Loss: 0.0994 | Time(s): 47.053\n",
      "Model is not saved! Current Val Loss: 0.0994\n",
      "Epoch 013: | Train Loss: 0.0874 | Val Loss: 0.0981 | Time(s): 46.943\n",
      "Model is not saved! Current Val Loss: 0.0981\n",
      "Epoch 014: | Train Loss: 0.0883 | Val Loss: 0.0977 | Time(s): 47.276\n",
      "Model is not saved! Current Val Loss: 0.0977\n",
      "Epoch 015: | Train Loss: 0.0890 | Val Loss: 0.0986 | Time(s): 47.021\n",
      "Model is not saved! Current Val Loss: 0.0986\n",
      "Epoch 016: | Train Loss: 0.0883 | Val Loss: 0.1023 | Time(s): 45.280\n",
      "Model is not saved! Current Val Loss: 0.1023\n",
      "Epoch 017: | Train Loss: 0.0880 | Val Loss: 0.1009 | Time(s): 40.716\n",
      "Model is not saved! Current Val Loss: 0.1009\n",
      "Epoch 018: | Train Loss: 0.0892 | Val Loss: 0.0979 | Time(s): 40.739\n",
      "Model is not saved! Current Val Loss: 0.0979\n",
      "Epoch 019: | Train Loss: 0.0916 | Val Loss: 0.0984 | Time(s): 40.733\n",
      "Model is not saved! Current Val Loss: 0.0984\n",
      "Epoch 020: | Train Loss: 0.0889 | Val Loss: 0.0977 | Time(s): 45.341\n",
      "Model is not saved! Current Val Loss: 0.0977\n",
      "Epoch 021: | Train Loss: 0.0870 | Val Loss: 0.1081 | Time(s): 46.994\n",
      "Model is not saved! Current Val Loss: 0.1081\n",
      "Epoch 022: | Train Loss: 0.0883 | Val Loss: 0.0979 | Time(s): 47.473\n",
      "Model is not saved! Current Val Loss: 0.0979\n",
      "Epoch 023: | Train Loss: 0.0886 | Val Loss: 0.0977 | Time(s): 47.616\n",
      "Model is not saved! Current Val Loss: 0.0977\n",
      "Epoch 024: | Train Loss: 0.0878 | Val Loss: 0.0978 | Time(s): 47.555\n",
      "Model is not saved! Current Val Loss: 0.0978\n",
      "Epoch 025: | Train Loss: 0.0883 | Val Loss: 0.0981 | Time(s): 46.928\n",
      "Model is not saved! Current Val Loss: 0.0981\n",
      "Epoch 026: | Train Loss: 0.0882 | Val Loss: 0.0977 | Time(s): 46.752\n",
      "Model is not saved! Current Val Loss: 0.0977\n",
      "Epoch 027: | Train Loss: 0.0920 | Val Loss: 0.0977 | Time(s): 46.831\n",
      "Model is not saved! Current Val Loss: 0.0977\n",
      "Epoch 028: | Train Loss: 0.0874 | Val Loss: 0.0981 | Time(s): 47.207\n",
      "Model is not saved! Current Val Loss: 0.0981\n",
      "Epoch 029: | Train Loss: 0.0888 | Val Loss: 0.1005 | Time(s): 47.153\n",
      "Model is not saved! Current Val Loss: 0.1005\n",
      "Epoch 030: | Train Loss: 0.0870 | Val Loss: 0.0989 | Time(s): 46.950\n",
      "Model is not saved! Current Val Loss: 0.0989\n"
     ]
    }
   ],
   "source": [
    "best_val_loss = float('inf')  # Initialize to infinity for better practice\n",
    "epochs = 30\n",
    "\n",
    "for epoch in range(1, epochs + 1):\n",
    "    training_start_time = time.time()\n",
    "\n",
    "    # Initialize epoch losses\n",
    "    train_epoch_loss = 0.0\n",
    "    val_epoch_loss = 0.0\n",
    "\n",
    "    # Training Phase\n",
    "    model.train()\n",
    "    for sample in data_loader_training:\n",
    "        xtrain = sample['input'].to(device).long()\n",
    "        ytrain_true = sample['label'].to(device).float()\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        ytrain_pred = model(xtrain)\n",
    "        train_loss = loss_func(ytrain_pred[:, 1], ytrain_true)\n",
    "        train_loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        train_epoch_loss += train_loss.item()\n",
    "\n",
    "    # Validation Phase\n",
    "    model.eval()\n",
    "    with torch.no_grad():\n",
    "        for sample in data_loader_validate:\n",
    "            xvalid = sample['input'].to(device).float()\n",
    "            yvalid_true = sample['label'].to(device).float()\n",
    "            yvalid_pred = model(xvalid)\n",
    "\n",
    "            val_loss = loss_func(yvalid_pred[:, 1], yvalid_true)\n",
    "            val_epoch_loss += val_loss.item()\n",
    "\n",
    "    # Calculate time taken for the epoch\n",
    "    training_duration_time = time.time() - training_start_time\n",
    "\n",
    "    # Compute average losses\n",
    "    avg_train_loss = train_epoch_loss / len(data_loader_training)\n",
    "    avg_val_loss = val_epoch_loss / len(data_loader_validate)\n",
    "\n",
    "    # Print training stats\n",
    "    print(f'Epoch {epoch:03}: | Train Loss: {avg_train_loss:.4f} | Val Loss: {avg_val_loss:.4f} | Time(s): {training_duration_time:.3f}') \n",
    "\n",
    "    # Model Saving Logic\n",
    "    if avg_val_loss < best_val_loss:\n",
    "        best_val_loss = avg_val_loss\n",
    "        torch.save(model.state_dict(), './best_model.pth')\n",
    "        print(f'===================== Best model Saved! Val Loss: {best_val_loss:.4f} ====================')\n",
    "    else:\n",
    "        print(f'Model is not saved! Current Val Loss: {avg_val_loss:.4f}') \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise 1: Evaluation the Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise 2: High Temp Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_dataset = pd.read_csv('kdfw_processed_sub_data.csv',index_col=0).sort_values(by='date') ##kdfw - Dallas-Ft Worth, TX\n",
    "total_dataset = total_dataset.replace('********', np.nan).replace(np.inf, np.nan).dropna(how='any',axis=1)\n",
    "total_dataset.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Observation data\n",
    "total_label_data = total_dataset.filter(like='OBS')\n",
    "\n",
    "#remove obs data and date from the feature set\n",
    "dropCols = list(total_label_data.columns) # + ['date']\n",
    "total_feature_data = total_dataset.copy(deep=True)\n",
    "total_feature_data = total_feature_data.drop(dropCols,axis=1)\n",
    "\n",
    "subsample_feature = total_feature_data.iloc[:,:51] #date + 50 features\n",
    "\n",
    "subsample_label = total_label_data['OBS_tmpf_max'][:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "subsample_label_class = pd.DataFrame()\n",
    "subsample_label_class['class'] = np.where(subsample_label > 100, 1, 0)\n",
    "# Binarize the 'temperature' column and label classes\n",
    "subsample_label_class['class_name'] = np.where(subsample_label > 100, 'Severe Days', 'Common Days')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_data_by_date(features, labels, date_ranges):\n",
    "    \"\"\"\n",
    "    Splits a DataFrame into multiple DataFrames based on specified date ranges.\n",
    "\n",
    "    Args:\n",
    "    df (pd.DataFrame): The original DataFrame.\n",
    "    time_col (str): The name of the column containing time data.\n",
    "    date_ranges (list of tuples): A list of tuples where each tuple contains start and end dates as strings.\n",
    "                                  Example: [('2009-01-01', '2012-12-31'), ('2013-01-01', '2017-12-31')]\n",
    "\n",
    "    Returns:\n",
    "    list: A list containing the split DataFrames.\n",
    "    \"\"\"\n",
    "\n",
    "    features_list = pd.to_datetime(features['date'])\n",
    "    date_mask = (features_list > date_ranges[0]) & (features_list <= date_ranges[1])\n",
    "    out_data = features.loc[date_mask,:].drop(['date'], axis=1)\n",
    "    out_labels = labels.loc[date_mask]\n",
    "\n",
    "    return out_data, out_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "date_ranges = [('2011-01-01','2017-12-31'), ('2018-01-01','2019-12-31')]\n",
    "time_train_x, time_train_y = split_data_by_date(subsample_feature, subsample_label_class,  date_ranges[0])\n",
    "time_test_x, time_test_y = split_data_by_date(subsample_feature, subsample_label_class, date_ranges[1])\n",
    "print(f\"train dataset samples: {time_train_x.shape} | test dataset samples: {time_test_x.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "def scale_data(train_x, test_x):\n",
    "    \"\"\"\n",
    "    Scales training and testing data using MinMaxScaler.\n",
    "\n",
    "    Parameters:\n",
    "    train_x (array-like): Training data features.\n",
    "    test_x (array-like): Testing data features.\n",
    "\n",
    "    Returns:\n",
    "    tuple: Scaled training and testing data.\n",
    "    \"\"\"\n",
    "\n",
    "    # Fit/find minimum and maximum value for each input variable\n",
    "    min_max_model = MinMaxScaler().fit(train_x)\n",
    "\n",
    "    # This saves the values to apply or 'transform' testing data\n",
    "    train_x_minmax = min_max_model.transform(train_x)\n",
    "    #print('Training data range \\nMax:', np.nanmax(train_x_minmax), ', Min:', np.nanmin(train_x_minmax))\n",
    "\n",
    "    # Apply the min/max values to testing data\n",
    "    test_x_minmax = min_max_model.transform(test_x)\n",
    "    #print('\\nTesting data range \\nMax:', np.nanmax(test_x_minmax), ', Min:', np.nanmin(test_x_minmax))\n",
    "\n",
    "    return train_x_minmax, test_x_minmax\n",
    "\n",
    "def split_data_by_date(features, labels, date_ranges):\n",
    "    \"\"\"\n",
    "    Splits a DataFrame into multiple DataFrames based on specified date ranges.\n",
    "\n",
    "    Args:\n",
    "    df (pd.DataFrame): The original DataFrame.\n",
    "    time_col (str): The name of the column containing time data.\n",
    "    date_ranges (list of tuples): A list of tuples where each tuple contains start and end dates as strings.\n",
    "                                  Example: [('2009-01-01', '2012-12-31'), ('2013-01-01', '2017-12-31')]\n",
    "\n",
    "    Returns:\n",
    "    list: A list containing the split DataFrames.\n",
    "    \"\"\"\n",
    "\n",
    "    features_list = pd.to_datetime(features['date'])\n",
    "    date_mask = (features_list > date_ranges[0]) & (features_list <= date_ranges[1])\n",
    "    out_data = features.loc[date_mask,:].drop(['date'], axis=1)\n",
    "    out_labels = labels.loc[date_mask]\n",
    "\n",
    "    return out_data, out_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_time_train_x, scaled_time_test_x = scale_data(time_train_x, time_test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = pd.DataFrame()\n",
    "train_df['label'] = time_train_y['class'].values\n",
    "new_rows = pd.DataFrame(scaled_time_train_x)\n",
    "train_df = pd.concat((train_df, new_rows), axis = 1)\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.DataFrame()\n",
    "test_df['label'] = time_test_y['class'].values\n",
    "new_rows_test = pd.DataFrame(scaled_time_test_x)\n",
    "test_df = pd.concat((test_df, new_rows_test), axis = 1)\n",
    "test_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "class dataloader(): \n",
    "\n",
    "    def __init__(self, dataframe):\n",
    "\n",
    "        self.dataframe = dataframe\n",
    "    def __len__(self):\n",
    "        return len(self.dataframe)\n",
    "\n",
    "    def __getitem__(self, idx): \n",
    "\n",
    "        input = self.dataframe.iloc[idx][1:]\n",
    "        input = torch.as_tensor(input, dtype=torch.float32)\n",
    "\n",
    "        label = self.dataframe.iloc[idx]['label'] \n",
    "        label = torch.as_tensor(label) #, dtype= torch.long\n",
    "\n",
    "        sample = {\n",
    "            \"input\": input, \n",
    "            \"label\": label}\n",
    "\n",
    "        return sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.reset_index()\n",
    "test_df = test_df.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = dataloader(train_df)\n",
    "test_dataset = dataloader(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_loader_training = torch.utils.data.DataLoader(train_dataset, batch_size= 32, \n",
    "                                                shuffle = True,  num_workers=0) \n",
    "data_loader_testing  = torch.utils.data.DataLoader(test_dataset, batch_size= 32, \n",
    "                                                shuffle = False,  num_workers=0) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_val_loss = float('inf')  # Initialize to infinity for better practice\n",
    "epochs = 30\n",
    "\n",
    "for epoch in range(1, epochs + 1):\n",
    "    training_start_time = time.time()\n",
    "\n",
    "    # Initialize epoch losses\n",
    "    train_epoch_loss = 0.0\n",
    "    val_epoch_loss = 0.0\n",
    "\n",
    "    # Training Phase\n",
    "    model.train()\n",
    "    for sample in data_loader_training:\n",
    "        xtrain = sample['input'].to(device).float()\n",
    "        ytrain_true = sample['label'].to(device).float()\n",
    "\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        ytrain_pred = model(xtrain)\n",
    "        train_loss = loss_func(ytrain_pred[:, 1], ytrain_true)\n",
    "        train_loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        train_epoch_loss += train_loss.item()\n",
    "\n",
    "    # Validation Phase\n",
    "    model.eval() \n",
    "    with torch.no_grad():\n",
    "        for sample in data_loader_testing:\n",
    "            xvalid = sample['input'].to(device).float()\n",
    "            yvalid_true = sample['label'].to(device).float()\n",
    "            yvalid_pred = model(xvalid)\n",
    "\n",
    "            val_loss = loss_func(yvalid_pred[:, 1], yvalid_true)\n",
    "            val_epoch_loss += val_loss.item()\n",
    "\n",
    "    # Calculate time taken for the epoch\n",
    "    training_duration_time = time.time() - training_start_time\n",
    "\n",
    "    # Compute average losses\n",
    "    avg_train_loss = train_epoch_loss / len(data_loader_training)\n",
    "    avg_val_loss = val_epoch_loss / len(data_loader_validate)\n",
    "\n",
    "    # Print training stats\n",
    "    print(f'Epoch {epoch:03}: | Train Loss: {avg_train_loss:.4f} | Val Loss: {avg_val_loss:.4f} | Time(s): {training_duration_time:.3f}') \n",
    "\n",
    "    # Model Saving Logic\n",
    "    if avg_val_loss < best_val_loss:\n",
    "        best_val_loss = avg_val_loss\n",
    "        torch.save(model.state_dict(), './best_model.pth')\n",
    "        print(f'===================== Best model Saved! Val Loss: {best_val_loss:.4f} ====================')\n",
    "    else:\n",
    "        print(f'Model is not saved! Current Val Loss: {avg_val_loss:.4f}') \n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "env",
   "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": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}