inspect_att_scores.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from models import GraphEmbedding\n",
    "from data_utils import GraphSetDataset\n",
    "import torch\n",
    "from torch_geometric.data import Batch\n",
    "import numpy as np\n",
    "\n",
    "dataset = GraphSetDataset('data/hea_youngs.pkl')\n",
    "result_path = 'results/att_score/'\n",
    "model_checkpoint = 'results/checkpoints/GraphConv_2_h32_youngs.pt'\n",
    "\n",
    "class SetsAttScore(torch.nn.Module):\n",
    "    def __init__(self, n_node_features, gnn_dim, molsets_dim, conv, n_conv_layers, after_readout='tanh', activation='relu', norm=None):\n",
    "        super().__init__()\n",
    "\n",
    "        match activation:\n",
    "            case 'relu': self.act = torch.nn.ReLU()\n",
    "            case 'silu': self.act = torch.nn.SiLU()\n",
    "            case 'gelu': self.act = torch.nn.GELU()\n",
    "            case 'lrelu': self.act = torch.nn.LeakyReLU()\n",
    "            case _: raise NotImplementedError\n",
    "        \n",
    "        emb_dim = n_node_features if conv == 'CGConv' else gnn_dim\n",
    "        self.phi = GraphEmbedding(n_node_features, emb_dim, conv, n_conv_layers, after_readout, self.act, norm=norm)\n",
    "\n",
    "        self.att_q_net = torch.nn.Linear(emb_dim, molsets_dim)\n",
    "        self.att_k_net = torch.nn.Linear(emb_dim, molsets_dim)\n",
    "        self.att_v_net = torch.nn.Linear(emb_dim, emb_dim)\n",
    "        self.rho = torch.nn.Sequential(\n",
    "            torch.nn.Linear(emb_dim, molsets_dim),\n",
    "            self.act,\n",
    "            torch.nn.Linear(molsets_dim, molsets_dim),\n",
    "            self.act,\n",
    "            torch.nn.Linear(molsets_dim, 1)\n",
    "        )\n",
    "\n",
    "    def forward(self, graph_list):\n",
    "        # graph_list is a batch of graphs in one datapoint\n",
    "        embeddings = self.phi(graph_list)  # n_graphs * emb_dim\n",
    "\n",
    "        frac = graph_list.y.unsqueeze(0)  # n_graphs * 1\n",
    "\n",
    "        # Permutation-invariant aggregation using attention mechanism\n",
    "        att_queries = self.att_q_net(embeddings)  # n_graphs * att_dim\n",
    "        att_values = self.att_v_net(embeddings)  # n_graphs * emb_dim\n",
    "        att_keys = self.att_k_net(embeddings)  # n_graphs * att_dim\n",
    "        att_scores = torch.matmul(att_queries, att_keys.transpose(0, 1)) / (att_keys.size(1) ** 0.5) # n_graphs * n_graphs\n",
    "        att_outputs = torch.matmul(torch.softmax(att_scores, dim=0), att_values) # n_graphs * emb_dim\n",
    "        \n",
    "        # Changed from here\n",
    "        return torch.div(torch.norm(att_outputs, dim=1), torch.norm(embeddings, dim=1))  # n_graphs\n",
    "    \n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = SetsAttScore(n_node_features=19, gnn_dim=32, molsets_dim=32, conv='GraphConv', n_conv_layers=2, after_readout='tanh', activation='relu', norm=None).to(device)\n",
    "\n",
    "model.load_state_dict(torch.load(model_checkpoint))\n",
    "scores = []\n",
    "model.eval()\n",
    "\n",
    "with torch.no_grad():\n",
    "    for sample in dataset:\n",
    "        inputs, target = sample\n",
    "        inputs = [data.to(device) for data in inputs]\n",
    "        inputs = Batch.from_data_list(inputs)\n",
    "        out = model(inputs)\n",
    "        scores.append(out.cpu().numpy())\n",
    "\n",
    "scores = np.stack(scores, axis=0)\n",
    "\n",
    "#%%\n",
    "np.savetxt(result_path + 'att_scores_youngs.csv', scores, fmt='%.3f', delimiter=',')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get compositions and youngs moduli"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "\n",
    "result_path = 'results/att_score/'\n",
    "\n",
    "with open('./data/hea_full.txt', 'r') as f:\n",
    "    data = json.loads(f.read())\n",
    "\n",
    "with open(result_path + 'comps.csv', 'w', newline='') as f:\n",
    "    # writer = csv.writer(f, quoting=csv.QUOTE_NONE, escapechar='\\\\', delimiter=',')\n",
    "    for dp in data:\n",
    "        bm = dp['Bulk modulus']\n",
    "        if bm < 0 or bm > 320:\n",
    "            continue\n",
    "        if 'Youngs modulus' in dp:\n",
    "            youngs = dp['Youngs modulus']\n",
    "            if youngs < 0 or youngs > 900:\n",
    "                continue\n",
    "            f.write(dp['Composition'].replace(' ', ',')+'\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Analytics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "result_path = 'results/att_score/'\n",
    "\n",
    "comps = pd.read_csv(result_path + 'comps.csv', header=None)\n",
    "scores = pd.read_csv(result_path + 'att_scores_youngs.csv', header=None)\n",
    "\n",
    "select_triple = []\n",
    "select_most = []\n",
    "select_both = []\n",
    "for i in range(1909):\n",
    "    row = scores.loc[i]\n",
    "    for j in range(3):\n",
    "        if row[j] > 3*min(row):\n",
    "            select_triple.append((i,j))\n",
    "            if row[j] == max(row):\n",
    "                select_both.append((i,j))\n",
    "        if row[j] == max(row):\n",
    "            select_most.append((i,j))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 15 artists>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "el_list = ['Al', 'Ti', 'V0', 'Cr', 'Mn', 'Fe', 'Co', 'Ni', 'Cu', 'Zr', 'Nb', 'Mo', 'Hf', 'Ta', 'W0']\n",
    "el_name_list = ['Al', 'Ti', 'V', 'Cr', 'Mn', 'Fe', 'Co', 'Ni', 'Cu', 'Zr', 'Nb', 'Mo', 'Hf', 'Ta', 'W']\n",
    "\n",
    "key_elements_triple = [comps.loc[i,j][0:2] for i,j in select_triple]\n",
    "\n",
    "key_dict_triple = {el_name_list[i]:key_elements_triple.count(el) for i, el in enumerate(el_list)}\n",
    "plt.rcParams['font.size'] = 16\n",
    "fig, ax = plt.subplots(figsize=(8, 3))\n",
    "ax.bar(*zip(*key_dict_triple.items()), color='#93b5cf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 15 artists>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "key_elements_both = [comps.loc[i,j][0:2] for i,j in select_both]\n",
    "\n",
    "key_dict_both = {el_name_list[i]:key_elements_both.count(el) for i, el in enumerate(el_list)}\n",
    "fig, ax = plt.subplots(figsize=(8, 3))\n",
    "ax.bar(*zip(*key_dict_both.items()), color='#8fb2c9')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Co Ta\n",
      "Ni Ta\n",
      "Cu Ta\n",
      "Hf Ta\n",
      "Ta Co\n",
      "Ta Ni\n",
      "Ta Cu\n",
      "Ta Hf\n"
     ]
    }
   ],
   "source": [
    "def find_scores_interact(el1, el2):\n",
    "    scores_with_el2 = []\n",
    "    scores_wo_el2 = []\n",
    "    for i, row in comps.iterrows():\n",
    "        row_els = [row[i][0:2] for i in range(4)]\n",
    "        if el1 in row_els:\n",
    "            row_score = scores.loc[i, row_els.index(el1)]\n",
    "            if el2 in row_els:\n",
    "                scores_with_el2.append(row_score)\n",
    "            else:\n",
    "                scores_wo_el2.append(row_score)\n",
    "    if min(len(scores_with_el2), len(scores_wo_el2)) < 1:\n",
    "        print(el1, el2)\n",
    "        return 0, 0\n",
    "    return np.mean(scores_with_el2), np.mean(scores_wo_el2)\n",
    "\n",
    "interact_map = np.zeros((15, 15))\n",
    "\n",
    "for i, el1 in enumerate(el_list):\n",
    "    for j, el2 in enumerate(el_list):\n",
    "        if i != j:\n",
    "            score_with, score_wo = find_scores_interact(el1, el2)\n",
    "            score_diff = score_with - score_wo\n",
    "            interact_map[i,j] = score_diff\n",
    "\n",
    "interactions = pd.DataFrame(interact_map, columns=el_name_list, index=el_name_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(78.25, 0.5, 'Element 1')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "plt.figure(figsize=(10,8))\n",
    "sns.set_theme(font_scale=1.8)\n",
    "sns.heatmap(interactions, cmap='coolwarm', center=0)\n",
    "plt.xlabel('Element 2')\n",
    "plt.ylabel('Element 1')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "gnn",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}