(14) Graph embeddings , DeepWalk, Node2Vec, Hierarchical softmax ,GraRep.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "omVMB32vJNgB"
   },
   "source": [
    "# Assignment — Graph embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "0jw1Bh-fJNgF"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.manifold import TSNE\n",
    "from torch.optim import Adam\n",
    "from scipy.sparse import csr_matrix\n",
    "from torch.utils.data import DataLoader\n",
    "from sklearn.cluster import k_means\n",
    "from scipy.sparse.linalg import svds\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm.notebook import tqdm, trange\n",
    "from sklearn.decomposition import PCA\n",
    "import networkx as nx\n",
    "from sklearn.metrics import mutual_info_score\n",
    "import requests\n",
    "from torch.utils.data import Dataset\n",
    "import torch.nn as nn\n",
    "import torch\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from scipy import sparse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "FZvmCa1k_nK-"
   },
   "source": [
    "In this assignment, we will evaluate node embedding methods on the facebook graph where nodes are pages and edges are links. Each node has a category: government, tv-show, company, politician."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "-I2DhszTyVg3",
    "outputId": "96776a6b-afab-433c-926b-841c635283da"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3873, array([0, 0, 0, ..., 3, 3, 3]))"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = 'https://raw.githubusercontent.com/netspractice/network-science/main/datasets/musae_facebook_ego_802.gml'\n",
    "open('musae_facebook_ego_802.gml', 'wb').write(requests.get(url).content)\n",
    "G = nx.read_gml('musae_facebook_ego_802.gml')\n",
    "G = nx.convert_node_labels_to_integers(G)\n",
    "_labels = np.array(list(nx.get_node_attributes(G, 'value').values()))\n",
    "unique = list(set(_labels))\n",
    "labels = np.array([unique.index(l) for l in _labels])\n",
    "len(G), labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Ssx8XznhJNgQ"
   },
   "source": [
    "### Task 1. DeepWalk (0 points)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "CT9EHbHPJNgR"
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VFEVAy9NmX1_"
   },
   "source": [
    "Deepwalk is an approach for learning latent representations of nodes in a network. The motivation of DeepWalk is based on an observation that the frequency of nodes occurrence in the short random walks in social networks is similar to the frequency of words occurrence in sentences in natural languages.\n",
    "\n",
    "<img src=\"https://raw.githubusercontent.com/netspractice/network-science/main/images/node_word_powerlaw.png\" width=600>\n",
    "\n",
    "The both follow a power-law distribution, therefore NLP word embedding models can help to represent nodes in networks. DeepWalk is based on SkipGram model that is trained to predict the context for a given word.\n",
    "\n",
    "<img src='https://lena-voita.github.io/resources/lectures/word_emb/w2v/window_prob1-min.png' width=600>\n",
    "\n",
    "(an image is taken from [NLP Course For You](https://lena-voita.github.io/nlp_course.html))\n",
    "\n",
    "DeepWalk uses nodes instead of words and random walks on a network instead of sentences. Let a central word be a start node of a random walk, context words be nodes in a tail of a random walk."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GNEIjHkD2Ess"
   },
   "source": [
    "Write a function `sample_random_walks` that takes a graph, number of walks per node and the length of walks, returns np.array of the shape (total number of walks) x (length)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "deletable": false,
    "id": "9Ow1aAoq0N1V",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "d3ce58f440fdc8e196671a9575dd819b",
     "grade": false,
     "grade_id": "cell-6dae7ab8bdd8bca1",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def sample_random_walks(G, walks_per_node, length):\n",
    "    walks = []\n",
    "    for node in G.nodes():\n",
    "      for _ in range(walks_per_node):\n",
    "        walk = [node]\n",
    "        for i in range(length -1):\n",
    "          walk.append(np.random.choice(list(G.neighbors(walk[-1]))))\n",
    "        walks.append(walk)\n",
    "    return np.array(walks)\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "ium8TB601-Ea",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "bf580c164d7ab96868a44d09bf855390",
     "grade": true,
     "grade_id": "cell-6804ce30d354612f",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "walks_per_node = 5\n",
    "length = 5\n",
    "rwalks = sample_random_walks(G, walks_per_node, length)\n",
    "assert rwalks.shape == (len(G) * walks_per_node, length)\n",
    "assert np.all(rwalks[:, 0][::5] == np.arange(len(G)))\n",
    "A = nx.to_numpy_array(G)\n",
    "assert np.all(A[rwalks[0, :-1], rwalks[0, 1:]] == 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ECHOTLkSbmaO",
    "outputId": "787c2311-9e0c-4760-cb36-5b868380949f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[   0,   33,   53,   47,   26],\n",
       "       [   0,   18,   62,  343,  279],\n",
       "       [   0,   39,   17,  291,  303],\n",
       "       ...,\n",
       "       [3872, 1495,  117,  292,  309],\n",
       "       [3872, 1495, 2121, 3179, 2121],\n",
       "       [3872, 1980, 1861,  169,   78]])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rwalks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "FZH3kd2Tvvfh"
   },
   "source": [
    "Consider a graph $G=(V,E)$. Let the first node in each random walk be _start node_ and others be _context nodes_. DeepWalk uses two embedding layers:\n",
    "* $v_i: \\{0, 1\\}^{|V|} \\to \\mathbb R^{d}$ embeds the one-hot encoded vector of the start node $i$ to latent space\n",
    "* $u_j: \\{0, 1\\}^{|V|} \\to \\mathbb R^{d}$ embeds the one-hot encoded vector of the context node $j$ to latent space\n",
    "\n",
    "The objective is to maximize the probability that $i$ and $j$ co-occur on a random walk over the network. Maximizing of the probability is equivalent to minimizing negative log-likelihood:\n",
    "\n",
    "$$\\mathcal L = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\log P(j|i)$$\n",
    "\n",
    "where $N$ is the number of walks per node, $L$ is the length of a random walk excluding start node. $P(j|i)$ can be modelled by softmax with dot product similarity score $\\text{sim}(i, j) = u_i^\\top v_j$ as follows:\n",
    "\n",
    "$$\\mathcal L = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\log \\frac{\\exp(v_i^\\top u_j)}{\\sum_{k=1}^{|V|}\\exp(u_k^\\top v_i)}$$\n",
    "\n",
    "However, calculating $\\sum_{k=1}^{|V|}\\exp(v_i^\\top u_j)$ is computationally expensive in a large network. To overcome such an obstacle, we can approximate softmax by binary cross-entropy with _negative sampling_. Instead of calculating softmax, we draw some random (negative) context and minimize binary cross-entropy using sigmoid function:\n",
    "\n",
    "$$\\mathcal L^{\\text{pos}}_{ij} = -\\log P(j|i) = - \\log \\sigma(v_i^\\top u_j) \\\\\n",
    "\\mathcal L^\\text{neg}_{ij} = -\\sum_{k=1}^K \\log P(k|i) = -\\sum_{k=1}^K \\log (1 - \\sigma(v_i^\\top u_k)) \\\\ \n",
    "\\mathcal L = \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\left(\\mathcal L^{\\text{pos}}_{ij} + \\mathcal L^\\text{neg}_{ij}\\right)$$\n",
    "\n",
    "where $K$ is the number of negative nodes for each context node."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "WWO3FS8f2BXA"
   },
   "source": [
    "Let us create a dataset for computing a such loss function. Write a class `NodeContextDataset`. \n",
    "\n",
    "Function `__init__` takes random walks, the number of nodes in a graph and the number of negative nodes per each context node.\n",
    "\n",
    "Function `__len__` returns the number of random walks.\n",
    "\n",
    "Function `__getitem__` takes an index of random walk, sample a negative context and returns a tuple:\n",
    "\n",
    "* start node, torch.int64\n",
    "* positive context nodes, torch.tensor of the shape (number of context nodes)\n",
    "* negative context nodes, torch.tensor of the shape (number of context nodes, number of negative nodes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "deletable": false,
    "id": "65SsWhyu2SZd",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "f37ac528f7f9c27b886cd738c11ed47f",
     "grade": false,
     "grade_id": "cell-92d80df6a92f6fc2",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "class NodeContextDataset(Dataset):\n",
    "    def __init__(self, rwalks, n_nodes, n_neg):\n",
    "        self.starting_node = torch.LongTensor(rwalks[:,0])\n",
    "        self.node_context = torch.LongTensor(rwalks[:,1:])\n",
    "        self.context_length = self.node_context.shape[1]\n",
    "        self.n_nodes = n_nodes\n",
    "        self.n_neg = n_neg\n",
    "        # YOUR CODE HERE\n",
    "        #raise NotImplementedError()\n",
    "\n",
    "    def __len__(self):\n",
    "        return self.starting_node.shape[0]\n",
    "        # YOUR CODE HERE\n",
    "        #raise NotImplementedError()\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        neg_samples = torch.randint(\n",
    "            low = 0,\n",
    "            high = self.n_nodes,\n",
    "            size = (self.context_length,self.n_neg)\n",
    "        )\n",
    "        return self.starting_node[idx],self.node_context[idx] , neg_samples\n",
    "        # YOUR CODE HERE\n",
    "        #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "me3MM4yK4NSb",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "33e459957a271d9fbd1451657741be76",
     "grade": true,
     "grade_id": "cell-a745361ad8b8827a",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "dataset = NodeContextDataset(rwalks, len(G), n_neg=5)\n",
    "assert len(dataset) == rwalks.shape[0]\n",
    "start_node, pos_context, neg_context0 = dataset[0]\n",
    "start_node, pos_context, neg_context1 = dataset[0]\n",
    "assert start_node == 0\n",
    "assert start_node.dtype == torch.int64\n",
    "assert start_node.shape == ()\n",
    "assert pos_context.shape == (4, )\n",
    "assert neg_context0.shape == neg_context1.shape ==(4, 5)\n",
    "assert not torch.all(neg_context0 == neg_context1)\n",
    "dloader = DataLoader(dataset, batch_size=2)\n",
    "for start_nodes, pos_context, neg_context in dloader:\n",
    "    break\n",
    "assert start_nodes.shape == (2,)\n",
    "assert pos_context.shape == (2, 4)\n",
    "assert neg_context.shape == (2, 4, 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XPaJWYeEOyH9"
   },
   "source": [
    "Write a function `cross_entropy` that takes vectors $v$, positive $u$, negative $u$ and returns the binary cross-entropy loss before reduction $\\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L$.\n",
    "\n",
    "_Remark: to prevent $-\\infty$ in log, add $\\varepsilon=1^{-6}$ as follows `torch.log(x + 1e-6)`_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "deletable": false,
    "id": "Gc4yYrdtTysM",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "8c52f1255915a57c9e9e136063d7a1bb",
     "grade": false,
     "grade_id": "cell-e2f4ba57342fd5e8",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def cross_entropy(v, u_pos, u_neg):\n",
    "    pos_prob = -torch.log(torch.sigmoid((v.unsqueeze(dim=1)*u_pos).sum(dim = 2))+1e-6)\n",
    "    neg_prob = -torch.log(1- torch.sigmoid((v.unsqueeze(dim=1).unsqueeze(dim=1)*u_neg).sum(dim = 3))+1e-6).sum(dim=2)\n",
    "    return pos_prob +neg_prob\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "5cJcyDIqaUTL",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "6560bbf935e37888e3e3b1bf47bd0881",
     "grade": true,
     "grade_id": "cell-588d17c4193a2d02",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "start_node_emb = torch.randn(len(G), 16)\n",
    "context_emb = torch.randn(len(G), 16)\n",
    "v = start_node_emb[start_nodes]\n",
    "u_pos = context_emb[pos_context]\n",
    "u_neg = context_emb[neg_context]\n",
    "loss = cross_entropy(v, u_pos, u_neg)\n",
    "assert loss.shape == (2, 4)\n",
    "lpos = -torch.log(torch.sigmoid(v[0] @ u_pos[0, 0]) + 1e-6)\n",
    "lneg = -torch.log(1 - torch.sigmoid(torch.tensor([v[0] @ u_neg[0, 0, i] for i in range(5)])) + 1e-6).sum()\n",
    "assert round((lpos + lneg).item(), 2) == round(loss[0, 0].item(), 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gH1kaJd4Sccq"
   },
   "source": [
    "Here is SkipGram model with negative sampling. It takes start nodes and positive, negative context nodes, returns cross-entropy loss."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "_nAOYIwr8_ZQ",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "fc264bd4d5ca49b71b8ebd673adcb1d7",
     "grade": false,
     "grade_id": "cell-79e82ef853f9c439",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "class SkipGramNegativeSampling(nn.Module):\n",
    "    def __init__(self, n_nodes, dim):\n",
    "        super().__init__()\n",
    "        self.start_node_emb = nn.Embedding(n_nodes, dim)\n",
    "        self.context_emb = nn.Embedding(n_nodes, dim)\n",
    "    def forward(self, start_nodes, pos_context, neg_context):\n",
    "        v = self.start_node_emb(start_nodes)\n",
    "        u_pos = self.context_emb(pos_context)\n",
    "        u_neg = self.context_emb(neg_context)\n",
    "        return cross_entropy(v, u_pos, u_neg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3YDBr9IfdMc_"
   },
   "source": [
    "Let us train the model using Adam optimizer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297,
     "referenced_widgets": [
      "ac2f589466aa4adc8410c3a33e4e8b12",
      "8cd3b25c74b14cef95da4ae10778c67c",
      "1142b2d3ecd1418ea80eb1bb3d1dcbf4",
      "700b1b90991c4d44839f64b1d10025ff",
      "43595dbc885742558dd80b1469442550",
      "4074a73f78534eb2a09caf3f35463e27",
      "6375042745a24584a55575854e37b65a",
      "eceaf917f28948a19eff59e769ca7c57",
      "5a80e1647f254196b24a2c80ee5adfc8",
      "410edff388c9415aa57058379782035e",
      "f23d5a29aa064f4a8123be43f70095d9"
     ]
    },
    "id": "0CnMaY-iBNot",
    "outputId": "988d4129-cd81-42f2-eee9-cb6de47ebb62"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ac2f589466aa4adc8410c3a33e4e8b12",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sgmodel = SkipGramNegativeSampling(n_nodes=len(G), dim=16)\n",
    "epoch_loss = []\n",
    "opt = Adam(sgmodel.parameters(), lr=0.1)\n",
    "dloader = DataLoader(dataset, batch_size=len(G))\n",
    "for epoch in trange(50):\n",
    "    for start_nodes, pos_context, neg_context in dloader:\n",
    "        loss = sgmodel(start_nodes, pos_context, neg_context).sum(dim=1).mean()\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        opt.step()\n",
    "    epoch_loss.append(loss.item())\n",
    "plt.plot(epoch_loss);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fOEAkiCUUfsu"
   },
   "source": [
    "We evaluate the model by mutual information between ground truth labels and cluster indices."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "1mdPsAiWFZ0Y",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "95d47fc2fc4c52f43987f446abb71cb2",
     "grade": true,
     "grade_id": "cell-1685a473a18d47f8",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "with torch.no_grad():\n",
    "    emb = sgmodel.start_node_emb(torch.arange(len(G)))\n",
    "_, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "mi = mutual_info_score(labels, pred_labels)\n",
    "assert mi > 0.15"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "As_bLu2mVGae"
   },
   "source": [
    "Let us plot the t-SNE visualization of node embeddings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 499
    },
    "id": "BrZKP5i3UzFs",
    "outputId": "1ef5ddf4-5a25-49c5-9494-473ccb7873dc"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "decomposition = TSNE(n_components=2)\n",
    "xy_emb = decomposition.fit_transform(emb)\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "scatter = plt.scatter(xy_emb[:, 0], xy_emb[:, 1], c=labels, s=10, cmap=plt.cm.Set2)\n",
    "\n",
    "plt.legend(handles=scatter.legend_elements()[0], labels=unique)\n",
    "plt.title(f'MI: {mi:.4f}');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NbvuXKkgJNgr"
   },
   "source": [
    "### Task 2. Node2Vec (3 points)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5q-lgag_YnQ3"
   },
   "source": [
    "In this task, we will consider Node2Vec embedding algorithm. In Node2Vec, we learn a mapping of nodes to a low-dimensional space of features that maximizes the likelihood of preserving network neighborhoods of nodes. It is similar to DeepWalk, but uses *biased random walk procedure* which efficiently explores diverse neighborhoods. There are two parameters:\n",
    "\n",
    "* Return parameter $p$ controls the likelihood of immediately revisiting a node in the walk. Setting it to a high value ensures that we are less likely to sample an already-visited node in the following two steps.\n",
    "\n",
    "* In-out parameter $q$ allows the search to differentiate between “inward” and “outward” nodes. If $q > 1$, the random walk is biased towards nodes close to previous node. In contrast, if $q < 1$, the walk is more inclined to visit nodes which are further away from the previous node.\n",
    "\n",
    "<img src='https://raw.githubusercontent.com/netspractice/network-science/main/images/biased_random_walk.png' width=300>\n",
    "\n",
    "Consider that we just moved from the node $t$ to $v$ and now we want to decide on the next step so it evaluates the transition probabilities on edges ($v$, $x$) leading from $v$. Then the *unnormalized* probability is\n",
    "\n",
    "$$\\alpha = \\begin{cases}\n",
    "\\frac{1}{p} &  \\text{ if } d_{tx} = 0\\\\\n",
    "1 & \\text{ if } d_{tx} = 1\\\\\n",
    "\\frac{1}{q} & \\text{ if } d_{tx} = 2\\\\\n",
    "\\end{cases}$$\n",
    "\n",
    "where $d_{tx}$ is the shortest path distance between nodes $t$ and $x$. To compute the true probability, we need to normalize values so that the sum is 1."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9LFfr6fy18Gb"
   },
   "source": [
    "Write a function `biased_random_walk` that takes a graph, node for which we start random walk, length of walk, parameters `p` and `q` and returns a list with random walks.\n",
    "\n",
    "*Hint: do not use `nx.shortest_path` to calcule shortest paths, it is too expensive. Look at the image above — we can explicitly calculate probabilities for all neighbors of $v$ using neighborhood of $t$*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "deletable": false,
    "id": "HGXAsOHw-aUV",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "248810a536b6d7a59e0b93a8937cf7a8",
     "grade": false,
     "grade_id": "cell-982226ce9aa30f91",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def biased_random_walk(G, node, length, p, q):\n",
    "    walk = [node]\n",
    "    nn = list (G.neighbors(node))\n",
    "    prop = np.zeros (len(nn)) \n",
    "    for i in range(len(nn)):\n",
    "      prop[i] = 1/p\n",
    "    norm_prop = [float(i)/sum(prop) for i in prop]\n",
    "    for i in range(length-1 ):\n",
    "      new_node = np.random.choice(nn, p=norm_prop)\n",
    "      last_node =  walk[-1]\n",
    "      nn = list (G.neighbors(new_node))\n",
    "      prop = np.zeros (len(nn))\n",
    "\n",
    "      for index, value in enumerate(nn):\n",
    "        if value == last_node:\n",
    "          prop[index] = 1/p\n",
    "        elif value in list (G.neighbors(last_node)):\n",
    "          prop[index] = 1\n",
    "        else:\n",
    "          prop[index] = 1/q\n",
    "      norm_prop = [float(i)/sum(prop) for i in prop]\n",
    "      walk.append (new_node)\n",
    "    return walk\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "uzi9dGUUcYMh",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "2435b88e56db99f5401d6df7bd8aac19",
     "grade": true,
     "grade_id": "cell-0404594e2e4226e7",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "rwalks = biased_random_walk(G, 0, 10, 0.1, 0.5)\n",
    "assert len(rwalks) == 10\n",
    "rwalks = biased_random_walk(G, node=0, length=4, p=0.001, q=1000)\n",
    "assert rwalks[0] == rwalks[2]\n",
    "assert rwalks[1] == rwalks[3]\n",
    "rwalks = biased_random_walk(G, node=0, length=4, p=1000, q=0.001)\n",
    "assert len(set(rwalks)) >= 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LGqNp_axcyFk"
   },
   "source": [
    "Let us generate biased random walks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "cmys-WiMGIAz"
   },
   "outputs": [],
   "source": [
    "def sample_biased_random_walks(G, walks_per_node, length, p, q):\n",
    "    walks = []\n",
    "    for node in tqdm(G.nodes, leave=False, desc='Sampling biased random walks'):\n",
    "        for _ in range(walks_per_node):\n",
    "            walk_from_node = biased_random_walk(G, node, length, p, q)\n",
    "            walks.append(walk_from_node)\n",
    "    return np.array(walks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 139,
     "referenced_widgets": [
      "8a39372c0d7944d4bbd7781e7fdb2eb1",
      "ac6f5904475b4c2c9f1454682acbaf45",
      "6e763d19798f410795a810b339b1371c",
      "7ce3b52703a040d38ab4030c80324aef",
      "e0c970d3b6154d78a1c08800151b34a0",
      "5f1120e664cc4305b78cefafa63c9b93",
      "e5fc97fd76ca425a973d96ede1d08678",
      "f0e5c7ec07e94626b2c7e1a4f7cfe2e2",
      "6e7e497548184b90b4f8a3f6dc871b5d",
      "41dd7d4d614a495d8140efc13ae3144b",
      "f4714f36d2b6441f8de4e9cd0256c4ff"
     ]
    },
    "id": "z5GQKyl3GXEy",
    "outputId": "c8ce9aee-e078-425a-ae70-babd22270408"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8a39372c0d7944d4bbd7781e7fdb2eb1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sampling biased random walks:   0%|          | 0/3873 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([[   0,   12,  160,  301,  337],\n",
       "       [   0,   38,  399,  452,  679],\n",
       "       [   0,    4,  134,   32,  361],\n",
       "       ...,\n",
       "       [3872, 1495,  294, 1831,  145],\n",
       "       [3872, 1980,  169, 1980, 3177],\n",
       "       [3872, 1980, 1861, 3181, 1861]])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "walks_per_node = 5\n",
    "length = 5\n",
    "rwalks = sample_biased_random_walks(G, walks_per_node, length, p=0.25, q=0.25)\n",
    "rwalks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "MhZic6vYQbs3"
   },
   "source": [
    "Train the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297,
     "referenced_widgets": [
      "c6e64987178b42dd8553e72c3ab92137",
      "a8981e199f414d4686fef5df98403ffb",
      "5767cde28d524f3682b470b585417f11",
      "c704472dadda4263896ec31feed1b2c0",
      "f29f288d77054f4f96bbe7ec4a052cd2",
      "cb5baf8c7891461f893e827b040223d7",
      "660c47186f5f469c9fa56b3bc12dbafb",
      "773ceaa7ee8c4101892473f7d33170d5",
      "8c11267bc29c47929f659e1003a1dfcf",
      "e5a9bb19066d47caa04cc8b0ea9ccc5c",
      "4ebcffe97f234827b88f8c4d94762262"
     ]
    },
    "id": "EQnnsFnwGlGz",
    "outputId": "6dc42514-6c0f-4920-9866-ed35b0078c43"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c6e64987178b42dd8553e72c3ab92137",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "e_loss = []\n",
    "dataset = NodeContextDataset(rwalks, len(G), n_neg=5)\n",
    "dloader = DataLoader(dataset, batch_size=len(G))\n",
    "sgmodel = SkipGramNegativeSampling(n_nodes=len(G), dim=16)\n",
    "opt = Adam(sgmodel.parameters(), lr=0.1)\n",
    "for e in trange(50):\n",
    "    for start_nodes, pos_context, neg_context in dloader:\n",
    "        loss = sgmodel(start_nodes, pos_context, neg_context).sum(dim=1).mean()\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        opt.step()\n",
    "    e_loss.append(loss.item())\n",
    "plt.plot(e_loss);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "MUkb4k-mdsj2",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "7005fe5ef24ada019070d2e796d90abd",
     "grade": true,
     "grade_id": "cell-d1d23ef4a88759f7",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "with torch.no_grad():\n",
    "    emb = sgmodel.start_node_emb(torch.arange(len(G)))\n",
    "_, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "mi = mutual_info_score(labels, pred_labels)\n",
    "assert mi > 0.15"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3J8GA322eAtB"
   },
   "source": [
    "Find the best $p$ and $q$. Write a function `grid_search_pq` that takes a graph, node labels and returns the best $p$ and $q$ by grid search in $\\{0.01, 0.2, 1, 2, 8\\}$. Sample biased random walks with 5 walks per node and the length 5. Evaluate the model by mutual information score between ground truth labels and k-means with 8 clusters. To pass time limits, calculate the best `p` and `q` and then rewrite the function as\n",
    "\n",
    "```python\n",
    "def grid_search_pq(graph):\n",
    "    return best_p, best_q\n",
    "    # your grid search implementation ...\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "deletable": false,
    "id": "drn8aJkje0D1",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "2381e148a50dffea83ac5792805d7c31",
     "grade": false,
     "grade_id": "cell-63bf0f1902240663",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def grid_search_pq(G, labels):\n",
    "    return (2, 0.01)\n",
    "    #output = {}\n",
    "    #for p in [0.01,0.2,1,2,8]:\n",
    "    #  for q in [0.01,0.2,1,2,8]:\n",
    "    #    rwalks = sample_biased_random_walks(G, 5, 5, p, q)\n",
    "    #    e_loss = []\n",
    "    #    dataset = NodeContextDataset(rwalks, len(G), n_neg=5)\n",
    "    #    dloader = DataLoader(dataset, batch_size=len(G))\n",
    "    #    sgmodel = SkipGramNegativeSampling(n_nodes=len(G), dim=16)\n",
    "    #    opt = Adam(sgmodel.parameters(), lr=0.1)\n",
    "    #    for e in trange(50):\n",
    "    #      for start_nodes, pos_context, neg_context in dloader:\n",
    "    #        loss = sgmodel(start_nodes, pos_context, neg_context).sum(dim=1).mean()\n",
    "    #        opt.zero_grad()\n",
    "    #        loss.backward()\n",
    "    #        opt.step()\n",
    "    #      e_loss.append(loss.item())\n",
    "    ##    #plt.plot(e_loss);\n",
    "     #   with torch.no_grad():\n",
    "     #     emb = sgmodel.start_node_emb(torch.arange(len(G)))\n",
    "     #     _, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "     #     mi = mutual_info_score(labels, pred_labels)\n",
    "     #   output [(p,q)] = mi\n",
    "    #return output\n",
    "\n",
    "        #assert mi > 0.15\n",
    "    \n",
    "    #parameters = {\n",
    "    #    'G':G,\n",
    "    #    'walks_per_node':5,\n",
    "    #    'length': 5,\n",
    "    #    'p': [0.01,0.2,1,2,8],\n",
    "    #    'q': [0.01,0.2,1,2,8]\n",
    "    #}\n",
    "    #grid_obj = GridSearchCV(estimator=sample_biased_random_walks, param_grid = parameters, cv = 2, n_jobs=-1)\n",
    "    #return grid_obj.best_params_\n",
    "\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 49,
     "referenced_widgets": [
      "8915d8e524e9456e89d00a4b4dffd2c5",
      "8a5553788326463a96a766cf42681954",
      "d1bc6631c4324aa09fc01e7a492726b9",
      "cc6d3ae8b7434d639e9e7a3c2ffe7316",
      "d207ec10a648446699f88344b8727589",
      "5baae0dec59047cf9814f8290614a5b2",
      "b20c8e72d7cc4a8db0a7aadba2f68dfe",
      "3a87a74057794059b1bda580834dbb2f",
      "dbe76ecf5c1248f091b376d980629f83",
      "e7586c08cee34cf699ba2248c01abdf6",
      "1564b81b7ecc450fbdf15b1323a4d565",
      "f02a455e777f4f4997477f341565ab37",
      "c5cdcc04ae8f4b6f8955aa51f2c5fb84",
      "9d4cca973d4f497991327c530dbf42ba",
      "eff6e8c5a87e4339adce3342553e1060",
      "0f4371fe938141b9a9131af8c2173ab0",
      "ae61cb27fa874099a6fea9321cc89cf3",
      "80244e511a504c929f1d29bd5d148a57",
      "3ada169d117048dd828c2b0eb7d7e5d0",
      "fd62483c077847869732558fd48cc692",
      "dbcf92c1a11d428ea730de23cd19f209",
      "2d597356e6c940afba19c5166d5660b3"
     ]
    },
    "deletable": false,
    "editable": false,
    "id": "igTGn7GALOJN",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "8e792a5d6956322dfacef58e8503a355",
     "grade": true,
     "grade_id": "cell-6102dd78a7156238",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "outputId": "0a1a0212-8115-4873-f3b5-dd56400413e2"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8915d8e524e9456e89d00a4b4dffd2c5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sampling biased random walks:   0%|          | 0/3873 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f02a455e777f4f4997477f341565ab37",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "best_p, best_q = grid_search_pq(G, labels)\n",
    "rwalks = sample_biased_random_walks(G, 5, 5, best_p, best_q)\n",
    "dataset = NodeContextDataset(rwalks, len(G), n_neg=5)\n",
    "dloader = DataLoader(dataset, batch_size=len(G))\n",
    "sgmodel = SkipGramNegativeSampling(n_nodes=len(G), dim=16)\n",
    "opt = Adam(sgmodel.parameters(), lr=0.1)\n",
    "dloader = DataLoader(dataset, batch_size=len(G))\n",
    "for epoch in trange(50):\n",
    "    for start_nodes, pos_context, neg_context in dloader:\n",
    "        loss = sgmodel(start_nodes, pos_context, neg_context).sum(dim=1).mean()\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        opt.step()\n",
    "with torch.no_grad():\n",
    "    emb = sgmodel.start_node_emb(torch.arange(len(G)))\n",
    "_, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "mi = mutual_info_score(labels, pred_labels)\n",
    "assert mi > 0.21"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4diPYQKCMdoT"
   },
   "source": [
    "Plot t-SNE visualization of node embeddings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 499
    },
    "id": "qkzOkF3NGzbR",
    "outputId": "859c912b-2472-4f66-b4b6-7519e745e95c"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "decomposition = TSNE(n_components=2)\n",
    "xy_emb = decomposition.fit_transform(emb)\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "scatter = plt.scatter(xy_emb[:, 0], xy_emb[:, 1], c=labels, s=10, cmap=plt.cm.Set2)\n",
    "\n",
    "plt.legend(handles=scatter.legend_elements()[0], labels=unique)\n",
    "plt.title(f'MI: {mi:.4f}');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "MsgOj9TWJNg2"
   },
   "source": [
    "### Task 3. Hierarchical softmax (4 points)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "I_93vIz0nsfT"
   },
   "source": [
    "Hierarchical softmax is an alternative way to approximate softmax in DeepWalk and similar models such as Walklets, Node2Vec. Remind the regular loss function with softmax:\n",
    "\n",
    "$$\\mathcal L = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\log P(j|i) = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\log \\frac{\\exp(v_i^\\top u_j)}{\\sum_{k=1}^{|V|}\\exp(v_i^\\top u_k)}$$\n",
    "\n",
    "We have observed that computing $\\sum_{k=1}^{|V|}\\exp(v_i^\\top u_k)$ is expensive in a large network. Indeed, it takes $O(|V|)$ since we need to compute $v_i^\\top u_k$ for each $k$. \n",
    "\n",
    "Hierarchical softmax allows to approximate softmax by propagation probabilities in a binary tree where leaves are nodes in the original network. Consider a simple network on 6 nodes and the corresponding binary tree.\n",
    "\n",
    "<img src=\"https://raw.githubusercontent.com/netspractice/network-science/main/images/hs_btree.png\" width=600>\n",
    "\n",
    "We observe that leaves correspond to nodes in the network: the leaf 5 corresponds to the node 0, the leaf 6 corresponds to the node 1 and so on. Let us call all other white nodes in the binary tree units.\n",
    "\n",
    "Write a function `get_binary_tree` that takes number of leaves and returns a binary tree.\n",
    "\n",
    "_Hint: create a binary tree by `nx.balanced_tree(2, depth, nx.DiGraph)` and drop extra leaves. The depth of the tree can be calculated as $\\lceil \\log_2 |V| \\rceil$._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "deletable": false,
    "id": "TTpjUbSMjORQ",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "01c8b6fa4bd256a22b699e36c1556bbd",
     "grade": false,
     "grade_id": "cell-fa6e506ce6741f45",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "import math\n",
    "def get_binary_tree(n_leaves):\n",
    "    depth = math.log(n_leaves,2)\n",
    "    print (depth)\n",
    "    G = nx.balanced_tree(2,round(depth),create_using=nx.DiGraph())\n",
    "    leaves = [n for n in G.nodes if G.degree[n] == 1]\n",
    "    while len(leaves) != n_leaves:\n",
    "      for node in leaves[n_leaves-1 :]:\n",
    "        G.remove_node(node)\n",
    "      leaves = [n for n in G.nodes if G.degree[n] == 1]\n",
    "    return G \n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "deletable": false,
    "editable": false,
    "id": "VinKbEWplrvy",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "8fcd40389d09de393cba575495e2513d",
     "grade": true,
     "grade_id": "cell-12931d335d97bb1c",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "outputId": "4ad5990e-ad89-4121-ada4-f87b5ee713e8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11.919235786028347\n"
     ]
    }
   ],
   "source": [
    "btree = get_binary_tree(n_leaves=len(G))\n",
    "leaves = [n for n in btree.nodes if btree.degree[n] == 1]\n",
    "assert len(leaves) == len(G)\n",
    "assert list(btree.nodes)[0] == 0\n",
    "assert nx.is_weakly_connected(btree)\n",
    "assert btree.in_degree[0] == 0\n",
    "assert np.all([btree.in_degree[n] == 1 for n in range(1, len(btree))])\n",
    "assert np.all([btree.out_degree[n] == 2 for n in range(len(btree) - len(leaves))])\n",
    "assert np.all([btree.out_degree[n] == 0 for n in range(len(btree) - len(leaves), len(btree))])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8KVUtUtm8OIy"
   },
   "source": [
    "If we assign a probability conditioned by the start node for each unit, we can calculate a probability of each context node as the product of probabilities of all units in the path from the root unit to a corresponding leaf taking $p_i$ if we go to the left and $1 - p_i$ if we go to the right from the unit $i$. For example, consider a path from the root unit to the context node 5 (leaf 10):\n",
    "\n",
    "<img src=\"https://raw.githubusercontent.com/netspractice/network-science/main/images/hs_btree_proba.png\" width=600>\n",
    "\n",
    "$$P(5|i) = p_0\\cdot (1-p_1) \\cdot (1-p_4)$$\n",
    "\n",
    "We can observe following thigs:\n",
    "* Each unit divides probability space into two subspace, so that probabilities on leaves are summed up to 1.\n",
    "* We make only 3 multiplucations instead of 6 summations in regular softmax. Indeed, the number of multiplications is equal to the depth of the binary tree $O(\\log_2|V|)$. The number of summations in regular softmax is $O(|V|)$. That is, the cost of calculating probabilities for all context nodes is reduced from $O(|V|^2)$ to $O(|V|\\log_2|V|)$.\n",
    "\n",
    "We condition probabilities by the start node as follows:\n",
    "\n",
    "$$p_j(i) = \\sigma(v_i^\\top u_j)$$\n",
    "\n",
    "where $v_i$ is embedding of the start node $i$ and $u_j$ is embedding of the unit $j$. Thus, we have no embeddings for context nodes in this model. Note that $1-\\sigma(v_i^\\top u_j)$ can be substituted to $\\sigma(-v_i^\\top u_j)$. Let us call a sign under the sigmoid a turn $t_k$: $-1$ if we go to the left and $1$ if we go to the right. Thus the probability of co-occurance of the context node 5 and the start node, say, 1 is\n",
    "\n",
    "$$P(5|1) = \\sigma(t_0 v_1^\\top u_0) \\cdot \\sigma(t_1 v_1^\\top u_1) \\cdot \\sigma(t_2 v_1^\\top u_4)$$\n",
    "\n",
    "where $t_0=1$, $t_1=-1$, $t_2=-1$. Our objective is to maximize the probability of co-occurance of start nodes and context nodes by minimizing the negative log-likelihood, thus the loss function is turned out to\n",
    "\n",
    "$$\\mathcal L = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\sum_{k=1}^K\\log \\sigma(t_{[jk]} v_i^\\top u_{[jk]})$$\n",
    "\n",
    "where $K$ is the path length from the root to the leaf excluding the leaf. Subscript $[jk]$ means the index corresponded to k-th step of the path to j-th context node. Note that the path length depends on a leaf, it makes calculating loss in mini-batch training inconvenient, so let us set $K$ as the depth of the tree and multiply the loss by _training mask_ $z_{jk}$: $0$ for all $k$ that greater than path length and $1$ otherwise. The final loss function is\n",
    "\n",
    "$$\\mathcal L = - \\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L \\sum_{k=1}^K\\log \\sigma(t_{[jk]} v_i^\\top u_{[jk]}) z_{[jk]}$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lk8UBAd214CG"
   },
   "source": [
    "Let us create a dataset for computing a such loss function. Write a class `NodePathDataset`. \n",
    "\n",
    "Function `__init__` takes random walks, binary tree, depth of the binary tree and finds paths for each leaf.\n",
    "\n",
    "Function `__len__` returns the number of random walks.\n",
    "\n",
    "Function `__getitem__` takes an index of random walk and returns a tuple:\n",
    "\n",
    "* start node, torch.int64\n",
    "* context nodes, torch.tensor of the shape (number of context nodes)\n",
    "* paths to the context nodes excluding a leaf: torch.tensor of the shape (number of context nodes, depth). Padded with zeros if the depth is greater than the path length.\n",
    "* turns on the paths: torch.tensor of the shape (number of context nodes, depth). Padded with zeros if the depth is greater than the path length. Define the left turn as a link to the even unit.\n",
    "* training mask on the paths: torch.tensor of the shape (number of context nodes, depth)\n",
    "\n",
    "_Hint: use `nx.shortest_path(btree, 0, n)` to find a path from the root unit to the leaf `n` in the tree._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "deletable": false,
    "id": "gNmgoZW1l9Wg",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "292835343292a61ef9503e0f113a2a32",
     "grade": false,
     "grade_id": "cell-48de35ab799a5afd",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "class NodePathDataset(Dataset):\n",
    "    def __init__(self, rwalks, btree, depth):\n",
    "        self.start_nodes = torch.tensor(rwalks[:, 0]).long()\n",
    "        self.context = torch.tensor(rwalks[:, 1:]).long()\n",
    "        \n",
    "        self.paths = []\n",
    "        self.turns = []\n",
    "        self.masks = []\n",
    "        leaves = [n for n in btree.nodes if btree.degree[n] == 1]\n",
    "        for node in leaves:\n",
    "          tmp_path = nx.shortest_path (btree,0,node)[:-1]\n",
    "          tmp_mask = [1] * len(tmp_path)\n",
    "          tmp_turn = []\n",
    "          for n in tmp_path:\n",
    "            if (n % 2 == 0 ):\n",
    "              tmp_turn.append(1)\n",
    "            else:\n",
    "              tmp_turn.append(-1)\n",
    "          if len(tmp_path) < depth:\n",
    "            n_extra_units = depth - len(tmp_path)\n",
    "            tmp_path = tmp_path + [0] * n_extra_units\n",
    "            tmp_mask = tmp_mask + [0] * n_extra_units\n",
    "            tmp_turn = tmp_turn + [0] * n_extra_units\n",
    "          self.paths.append(tmp_path)\n",
    "          self.masks.append(tmp_mask)\n",
    "          self.turns.append(tmp_turn)\n",
    "\n",
    "        self.paths = torch.tensor(self.paths)\n",
    "        self.turns = torch.tensor(self.turns)\n",
    "        self.masks = torch.tensor(self.masks)\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.start_nodes)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        return (\n",
    "            self.start_nodes[idx], \n",
    "            self.context[idx], \n",
    "            self.paths[self.context[idx]], \n",
    "            self.turns[self.context[idx]], \n",
    "            self.masks[self.context[idx]]\n",
    "        )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 17,
     "referenced_widgets": [
      "402a8d91fe6547b19f632e7aadf299f7",
      "609cc7ef350f4b8499c385f1db465ca2",
      "252e9d40ef774c9d9148f8e58747b5b2",
      "8574c57a8e50471fa75d796638a3b19b",
      "0f43bdee0f87447c8d0d401f39101139",
      "2cd5657d6b3243a0b5d7ddc2ec1125f8",
      "27c72c8216fe40f2a9f0abb71d00e52f",
      "fdec3af90eab4faea793d307453d9f69",
      "16f4a6b2e6044c63a73c9419223c5eca",
      "3e88b235dd8b4461bf056ae966cca361",
      "9888aaf95c70421895d3c26c3282043d"
     ]
    },
    "deletable": false,
    "editable": false,
    "id": "mColU52_WOJi",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "d8ce7f617444f4b2c156cc851907e46d",
     "grade": true,
     "grade_id": "cell-41adebff43c8f34d",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "outputId": "1c3918d1-261e-4566-ce19-268f0fc1ba01"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "402a8d91fe6547b19f632e7aadf299f7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sampling biased random walks:   0%|          | 0/3873 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "depth = 12\n",
    "rwalks = sample_biased_random_walks(G, 5, 5, best_p, best_q)\n",
    "dataset = NodePathDataset(rwalks, btree, depth)\n",
    "start_node, context, paths, turns, masks = dataset[0]\n",
    "assert start_node.shape == ()\n",
    "assert context.shape == (4,)\n",
    "assert paths.shape == turns.shape == masks.shape == (4, 12)\n",
    "assert start_node == 0\n",
    "leaves = [n for n in btree.nodes if btree.degree[n] == 1]\n",
    "shortest_path = nx.shortest_path(btree, 0, leaves[context[1]])[:-1]\n",
    "mask = [1] * len(shortest_path)\n",
    "if len(shortest_path) < depth:\n",
    "    n_extra_units = depth - len(shortest_path)\n",
    "    shortest_path = shortest_path + [0] * n_extra_units\n",
    "    mask = mask + [0] * n_extra_units\n",
    "assert np.all(shortest_path == paths[1].numpy())\n",
    "assert np.all(mask == masks[1].numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "A66EbScvpz6z",
    "outputId": "ff8cf3a9-abed-4a19-f82a-4e1001bcf666"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([2]),\n",
       " torch.Size([2, 4]),\n",
       " torch.Size([2, 4, 12]),\n",
       " torch.Size([2, 4, 12]),\n",
       " torch.Size([2, 4, 12]))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dloader = DataLoader(dataset, batch_size=2)\n",
    "for start_nodes, context, paths, turns, masks in dloader:\n",
    "    break\n",
    "start_nodes.shape, context.shape, paths.shape, turns.shape, masks.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eoXn4BU1oEBd"
   },
   "source": [
    "Write a function `hierarchical_softmax_loss` that takes vectors $v$, $u$ and returns the hierarchical softmax loss before reduction $\\frac{1}{|V|\\times N} \\sum_{i=1}^{|V|\\times N} \\sum_{j=1}^L$.\n",
    "\n",
    "_Remark: to prevent $-\\infty$ in log, add $\\varepsilon=1^{-6}$ as follows `torch.log(x + 1e-6)`_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "deletable": false,
    "id": "0haP3i7hni4c",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "2d24227f9a5b502d535b0beff12e17f1",
     "grade": false,
     "grade_id": "cell-5580ea00c3649ace",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def hierarchical_softmax_loss(v, u, turns, masks):\n",
    "    l = -torch.log(torch.sigmoid((v.unsqueeze(dim=1).unsqueeze(dim=1)*u).sum(dim=3) * turns )+ 1e-6) * masks                              \n",
    "    return l.sum(dim=2)\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "vf0XAmFQpVhP",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "77571345c5f803b3d52b802c082392e0",
     "grade": true,
     "grade_id": "cell-8c8320b2c7c030eb",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "start_node_emb = torch.randn(len(G), 16)\n",
    "path_emb = torch.randn(len(G) - 1, 16)\n",
    "v = start_node_emb[start_nodes]\n",
    "u = path_emb[paths]\n",
    "loss = hierarchical_softmax_loss(v, u, turns, masks)\n",
    "assert loss.shape == (2, 4)\n",
    "l = -torch.log(torch.sigmoid(torch.tensor([v[0] @ u[0, 0, i] * turns[0, 0, i] for i in range(12)])) + 1e-6) * masks[0, 0]\n",
    "assert round(l.sum().item(), 2) == round(loss[0, 0].item(), 2)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KrmcWlWDxyjm"
   },
   "source": [
    "Here is SkipGram model with hierarchical softmax. The number of vectors in `path_emb` is `n_nodes - 1`, since it can be shown that the number of units is equal to $|V|-1$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "gfQ3aA-Bvxer",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "8535c3166b0b7c5bc2e57ab2960ddc75",
     "grade": false,
     "grade_id": "cell-39c948726656b5ba",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "class SkipGramHierarchicalSoftmax(nn.Module):\n",
    "    def __init__(self, n_nodes, dim):\n",
    "        super().__init__()\n",
    "        self.start_node_emb = nn.Embedding(n_nodes, dim)\n",
    "        self.path_emb = nn.Embedding(n_nodes - 1, dim)\n",
    "    def forward(self, start_nodes, paths, turns, masks):\n",
    "        v = self.start_node_emb(start_nodes)\n",
    "        u = self.path_emb(paths)\n",
    "        loss = hierarchical_softmax_loss(v, u, turns, masks)\n",
    "        return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uJgoNM6qz3ra"
   },
   "source": [
    "Train the model using Adam optimizer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297,
     "referenced_widgets": [
      "60d5de0bd3e84072840e647fea80cfe4",
      "26f87406367e4050852787044e6e56ef",
      "31a7b4f5fc424ea4a69582db3ba3f345",
      "7e9b54c6689d40f2a2cfe9d886a96f8d",
      "234c9f52c20b455faffd6d2dd18fdcd1",
      "67fa78a581814c5fb4350ad9dc363e28",
      "0927d030aa19439a913de34318beca65",
      "7b292585bbe846ea84343de2dbb3596a",
      "2e018158da9c4362817489ff0fef11e3",
      "9b0cdbf346114368bdad72acf84bfb99",
      "c0124b803b5e4e61ae20245eb9777113"
     ]
    },
    "id": "qyUFLJnx1npz",
    "outputId": "7f33e4a7-0502-43b6-a6c0-9f11874a31b7"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "60d5de0bd3e84072840e647fea80cfe4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sgmodel = SkipGramHierarchicalSoftmax(n_nodes=len(G), dim=16)\n",
    "epoch_loss = []\n",
    "opt = Adam(sgmodel.parameters(), lr=0.1)\n",
    "dloader = DataLoader(dataset, batch_size=len(G))\n",
    "for epoch in trange(50):\n",
    "    for start_nodes, context, paths, turns, masks in dloader:\n",
    "        loss = sgmodel(start_nodes, paths, turns, masks).sum(dim=1).mean()\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        opt.step()\n",
    "    epoch_loss.append(loss.item())\n",
    "plt.plot(epoch_loss);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jFwKNMO3z8X2"
   },
   "source": [
    "Evaluate the model by mutual information between ground truth labels and cluster indices."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 201
    },
    "deletable": false,
    "editable": false,
    "id": "Y0ugT6Ep173h",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "163de70b349c978e9503141a12c90a5e",
     "grade": true,
     "grade_id": "cell-736c8f9bf1ffccd8",
     "locked": true,
     "points": 1,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "outputId": "46c287b3-5c38-48f4-d1d9-ec3fbbfd2f4e"
   },
   "outputs": [
    {
     "ename": "AssertionError",
     "evalue": "ignored",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-38-87e0f9d46b21>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0m_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mk_means\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0memb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mn_clusters\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m8\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mmi\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmutual_info_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_labels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0;32massert\u001b[0m \u001b[0mmi\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0.15\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAssertionError\u001b[0m: "
     ]
    }
   ],
   "source": [
    "with torch.no_grad():\n",
    "    emb = sgmodel.start_node_emb(torch.arange(len(G)))\n",
    "_, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "mi = mutual_info_score(labels, pred_labels)\n",
    "assert mi > 0.15"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cPql7hMu0HF8"
   },
   "source": [
    "Plot t-SNE visualization of node embeddings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "dZnRfklrzFsx"
   },
   "outputs": [],
   "source": [
    "decomposition = TSNE(n_components=2)\n",
    "xy_emb = decomposition.fit_transform(emb)\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "scatter = plt.scatter(xy_emb[:, 0], xy_emb[:, 1], c=labels, s=10, cmap=plt.cm.Set2)\n",
    "plt.legend(handles=scatter.legend_elements()[0], labels=unique)\n",
    "plt.title(f'MI: {mi:.4f}');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lCmzKtCAJ-zD"
   },
   "source": [
    "The hierarchical softmax can be improved by building a binary tree where hubs are closer to the root unit. In this task we have used the randomly builded binary tree and we can see that it works pretty well with respect to negative sampling."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aa-SAa-XJNhI"
   },
   "source": [
    "### Task 4. GraRep (3 points)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aviJdvivWW6D"
   },
   "source": [
    "The idea behind GraRep is to account for multiscale relationships between nodes. For example,a student can have 1-hop neighborhood of friends, 2-hop neighborhood of classmates, 3-hop neighborhood of society and so on.\n",
    "\n",
    "<img src=\"https://raw.githubusercontent.com/netspractice/network-science/main/images/multiscale_relationships.png\" width=350>\n",
    "\n",
    "Unlike DeepWalk, GraRep defines different objective functions for capturing the different $k$-step local relational information by manipulating transition matrices in the $k$-th powers defined over the graph. Recall that the transition matrix is $P = D^{-1}A$ where $D$ is a degree matrix. $P_{ij}$ refers to a probability to move from the node $i$ to the node $j$ in one step of a random walk. Therefore, $P^k_{ij}$ is a probability to move from the node $i$ to the node $j$ in $k$ steps of a random walk. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vdylqNJVreik"
   },
   "source": [
    "Write a function `csr_transition_matrix` that takes a graph and returns a transition matrix. To save memory and speed up the following calculation of powers, convert transition matrix to sparse format by `scipy.sparse.csr_matrix`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "id": "ijUKHWFDLKUc",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "bec7a132a8e91292a33f9ce803d9e648",
     "grade": false,
     "grade_id": "cell-f0ec11b1c4da2c21",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def csr_transition_matrix(G):\n",
    "    A = nx.to_numpy_array(G)\n",
    "    D = np.zeros(A.shape)\n",
    "    r,c = A.shape\n",
    "    for i in range(0,r):\n",
    "      for j in range (0,c):\n",
    "        if i == j:\n",
    "          D[i,i] = np.sum(A[i])\n",
    "    P = np.dot(np.linalg.inv(D),A)\n",
    "    return sparse.csr_matrix(P)\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "LKqHuKwhLvI6",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "89c5d34c88f598b589ecaf7c68858755",
     "grade": true,
     "grade_id": "cell-2d2f99d8bac11893",
     "locked": true,
     "points": 0.75,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "P = csr_transition_matrix(G)\n",
    "assert type(P) == csr_matrix\n",
    "assert P.shape == (len(G), len(G))\n",
    "assert round(P[3, 9], 4) == 0.0417"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NAQLGKBAH8tf"
   },
   "source": [
    "Similar to DeepWalk, there are two types of embeddings:\n",
    "* $v_i$ is a vector that represents the start node $i$\n",
    "* $u_j$ is a vector that represents the context node $j$\n",
    "\n",
    "Our objective aims to maximize: 1) the probability that these pairs come from the graph, and 2) the probability that all other pairs do not come from the graph (negative samples). The objective function for nodes $i, j$ in the transition matrix of the power $k$ is\n",
    "\n",
    "$$\\mathcal L^k_{ij} = P^k_{ij}\\log \\sigma(v_i^\\top u_j) + \\frac{\\lambda}{|V|}\\sum_{t=1}^{|V|}P^k_{tj}\\log \\sigma(-v_t^\\top u_j)$$\n",
    "\n",
    "where $\\lambda$ is the hyperparameter indicating the number of negative samples. Optimizing this objective makes observed pairs $(i, j)$ have similar embeddings, while scattering unobserved pairs $(t, j)$. Substitution $v_i^\\top u_j = X_{ij}$ and setting $\\frac{\\partial \\mathcal L_{ij}}{ \\partial X_{ij}} = 0$ yields\n",
    "\n",
    "$$v_i^\\top u_j = X_{ij}^k = \\log \\frac{P^k_{ij}}{\\sum_{t=1}^{|V|}P^k_{tj}} - \\log\\frac{\\lambda}{|V|}$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X7bwqmWGs-QC"
   },
   "source": [
    "Let us call $X^k$ log probabilistic matrix. Write a function `log_probabilistic_matrix` that takes a transition matrix, hyperparameter $\\lambda$ and returns a log probabilistic matrix.\n",
    "\n",
    "_Remark: to prevent $-\\infty$ in log, add $\\varepsilon=1^{-6}$ as follows `np.log(x + 1e-6)`_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "id": "XjVAyoUiMq_8",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "63684003d32189034bf77a52da99cd2b",
     "grade": false,
     "grade_id": "cell-c949fe784127c31a",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def log_probabilistic_matrix(P, lambd):\n",
    "  n = P.shape[0]\n",
    "  X = np.log(P/P.sum(axis=0) + 1e-6) - np.log(lambd/n + 1e-6)\n",
    "  return X\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "KTl_DISFt8mA",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "f81afad0435cd21fc8814f7974fd775b",
     "grade": true,
     "grade_id": "cell-624f6d1384718d6f",
     "locked": true,
     "points": 0.75,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "X = log_probabilistic_matrix(P, lambd=5)\n",
    "assert type(X) == np.matrix\n",
    "assert X.shape == (len(G), len(G))\n",
    "assert round(X[3, 9], 3) == 3.182"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "y56r1-evIBAJ"
   },
   "source": [
    "The log probabilistic matrix $X^k$ can be approximated by truncated SVD so that\n",
    "\n",
    "$$X^k \\approx X_d^k = U_d^k \\Sigma_d^k (V_d^k)^\\top$$\n",
    "\n",
    "where $d$ is the number of singular values of truncated SVD. Thus, embeddings of start nodes can be obtained by \n",
    "\n",
    "$$v_i = \\left[U_d^k \\sqrt{\\Sigma_d^k}\\right]_i$$ \n",
    "\n",
    "and embeddings of context nodes by \n",
    "\n",
    "$$u_j = \\left[\\sqrt{\\Sigma_d^k} (V_d^k)^\\top\\right]_j$$ \n",
    "\n",
    "_Remark: the power 1/2 makes SVD symmetric. It is empirically shown that the symmetric SVD is better for social graphs and makes embeddings more \"similar\" to corresponding embeddings of SkipGram with negative sampling in a sence of some matrix properties._"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "M5gDGuACuy-5"
   },
   "source": [
    "Write a function `svd_emb` that takes a log probabilistic matrix, number of dimensions (number of largest singular values) in truncated SVD and returns embeddings of start nodes.\n",
    "\n",
    "_Hint: use `scipy.sparse.linalg.svds` to calculate truncated SVD for a given number of singular values._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "id": "S9yoD3FxOXbw",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "64aa1dc62ca5b5035ec1eedfc7372ab7",
     "grade": false,
     "grade_id": "cell-683cb8f26df5fd3a",
     "locked": false,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def svd_emb(X, dim):\n",
    "    u, s, vT = svds(X, k=dim)\n",
    "    #print (u.shape)\n",
    "    #print (s.shape)\n",
    "    #print (vT.shape)\n",
    "    return u @ np.sqrt(np.diag(s)) \n",
    "\n",
    "    # YOUR CODE HERE\n",
    "    #raise NotImplementedError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "fVF6rZ9dJNhS",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "b79ac0db377db5bae1f2923b7877e2dc",
     "grade": true,
     "grade_id": "cell-426bb9037fe96aba",
     "locked": true,
     "points": 0.75,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "_emb = svd_emb(X, dim=4)\n",
    "assert type(_emb) == np.ndarray\n",
    "assert _emb.shape == (3873, 4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DMgCQ_9GxjKg"
   },
   "source": [
    "\n",
    "The resulting node embeddings are concatenated from the log probabilistic matrices $X, X^2, ..., X^k$.\n",
    "\n",
    "In `grarep_emb`, we calculate powers of transition matrix $P, P^2, \\dots, P^k$, calculate and concatenate embeddings of start nodes, then we compress them to the original number of dimensions by PCA."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "V76RS5EmJNhT",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "08744ba53988cb545f953cce607f67c6",
     "grade": false,
     "grade_id": "cell-fe1d6156e8d730ad",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def grarep_emb(G, dim, k, lambd):\n",
    "    emb = []\n",
    "    P_1 = csr_transition_matrix(G)\n",
    "    P = P_1.copy()\n",
    "    for i in trange(k):\n",
    "        X = log_probabilistic_matrix(P, lambd=lambd)\n",
    "        _emb = svd_emb(X, dim=dim)\n",
    "        emb.append(_emb)\n",
    "        if i == (k - 1):\n",
    "            continue\n",
    "        P = P @ P_1\n",
    "    emb = PCA(n_components=dim).fit_transform(np.concatenate(emb, axis=1))\n",
    "    return emb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "SitFJ7ZcJNhU"
   },
   "source": [
    "Finally, we evaluate the model by mutual information between ground truth labels and cluster indices."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "deletable": false,
    "editable": false,
    "id": "0NvbTouIFSm7",
    "nbgrader": {
     "cell_type": "code",
     "checksum": "a8be8f00f7fc401b45409cd399e91d28",
     "grade": true,
     "grade_id": "cell-a016b73d556e34c9",
     "locked": true,
     "points": 0.75,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "emb = grarep_emb(G, dim=16, k=4, lambd=5)\n",
    "assert emb.shape == (len(G), 16)\n",
    "_, pred_labels, _ = k_means(emb, n_clusters=8)\n",
    "mi = mutual_info_score(labels, pred_labels)\n",
    "assert mi > 0.22"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VkLVoqDV0L93"
   },
   "source": [
    "Plot t-SNE visualization of node embeddings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ur-bWmtED-DE"
   },
   "outputs": [],
   "source": [
    "decomposition = TSNE(n_components=2)\n",
    "xy_emb = decomposition.fit_transform(emb)\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "scatter = plt.scatter(xy_emb[:, 0], xy_emb[:, 1], c=labels, s=10, cmap=plt.cm.Set2)\n",
    "plt.legend(handles=scatter.legend_elements()[0], labels=unique)\n",
    "plt.title(f'MI: {mi:.4f}');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "GG8kwRNRJNhY"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "final try.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "0927d030aa19439a913de34318beca65": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "0f4371fe938141b9a9131af8c2173ab0": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "0f43bdee0f87447c8d0d401f39101139": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1142b2d3ecd1418ea80eb1bb3d1dcbf4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_eceaf917f28948a19eff59e769ca7c57",
      "max": 50,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_5a80e1647f254196b24a2c80ee5adfc8",
      "value": 50
     }
    },
    "1564b81b7ecc450fbdf15b1323a4d565": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "16f4a6b2e6044c63a73c9419223c5eca": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "234c9f52c20b455faffd6d2dd18fdcd1": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "252e9d40ef774c9d9148f8e58747b5b2": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_fdec3af90eab4faea793d307453d9f69",
      "max": 3873,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_16f4a6b2e6044c63a73c9419223c5eca",
      "value": 3873
     }
    },
    "26f87406367e4050852787044e6e56ef": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_67fa78a581814c5fb4350ad9dc363e28",
      "placeholder": "​",
      "style": "IPY_MODEL_0927d030aa19439a913de34318beca65",
      "value": "100%"
     }
    },
    "27c72c8216fe40f2a9f0abb71d00e52f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "2cd5657d6b3243a0b5d7ddc2ec1125f8": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2d597356e6c940afba19c5166d5660b3": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "2e018158da9c4362817489ff0fef11e3": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "31a7b4f5fc424ea4a69582db3ba3f345": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_7b292585bbe846ea84343de2dbb3596a",
      "max": 50,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_2e018158da9c4362817489ff0fef11e3",
      "value": 50
     }
    },
    "3a87a74057794059b1bda580834dbb2f": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "3ada169d117048dd828c2b0eb7d7e5d0": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "3e88b235dd8b4461bf056ae966cca361": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "402a8d91fe6547b19f632e7aadf299f7": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_609cc7ef350f4b8499c385f1db465ca2",
       "IPY_MODEL_252e9d40ef774c9d9148f8e58747b5b2",
       "IPY_MODEL_8574c57a8e50471fa75d796638a3b19b"
      ],
      "layout": "IPY_MODEL_0f43bdee0f87447c8d0d401f39101139"
     }
    },
    "4074a73f78534eb2a09caf3f35463e27": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "410edff388c9415aa57058379782035e": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "41dd7d4d614a495d8140efc13ae3144b": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "43595dbc885742558dd80b1469442550": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "4ebcffe97f234827b88f8c4d94762262": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "5767cde28d524f3682b470b585417f11": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_773ceaa7ee8c4101892473f7d33170d5",
      "max": 50,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_8c11267bc29c47929f659e1003a1dfcf",
      "value": 50
     }
    },
    "5a80e1647f254196b24a2c80ee5adfc8": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "5baae0dec59047cf9814f8290614a5b2": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5f1120e664cc4305b78cefafa63c9b93": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "609cc7ef350f4b8499c385f1db465ca2": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2cd5657d6b3243a0b5d7ddc2ec1125f8",
      "placeholder": "​",
      "style": "IPY_MODEL_27c72c8216fe40f2a9f0abb71d00e52f",
      "value": "Sampling biased random walks: 100%"
     }
    },
    "60d5de0bd3e84072840e647fea80cfe4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_26f87406367e4050852787044e6e56ef",
       "IPY_MODEL_31a7b4f5fc424ea4a69582db3ba3f345",
       "IPY_MODEL_7e9b54c6689d40f2a2cfe9d886a96f8d"
      ],
      "layout": "IPY_MODEL_234c9f52c20b455faffd6d2dd18fdcd1"
     }
    },
    "6375042745a24584a55575854e37b65a": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "660c47186f5f469c9fa56b3bc12dbafb": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "67fa78a581814c5fb4350ad9dc363e28": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "6e763d19798f410795a810b339b1371c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_f0e5c7ec07e94626b2c7e1a4f7cfe2e2",
      "max": 3873,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_6e7e497548184b90b4f8a3f6dc871b5d",
      "value": 3873
     }
    },
    "6e7e497548184b90b4f8a3f6dc871b5d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "700b1b90991c4d44839f64b1d10025ff": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_410edff388c9415aa57058379782035e",
      "placeholder": "​",
      "style": "IPY_MODEL_f23d5a29aa064f4a8123be43f70095d9",
      "value": " 50/50 [00:35&lt;00:00,  1.18it/s]"
     }
    },
    "773ceaa7ee8c4101892473f7d33170d5": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "7b292585bbe846ea84343de2dbb3596a": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "7ce3b52703a040d38ab4030c80324aef": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_41dd7d4d614a495d8140efc13ae3144b",
      "placeholder": "​",
      "style": "IPY_MODEL_f4714f36d2b6441f8de4e9cd0256c4ff",
      "value": " 3866/3873 [01:27&lt;00:00, 94.87it/s]"
     }
    },
    "7e9b54c6689d40f2a2cfe9d886a96f8d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_9b0cdbf346114368bdad72acf84bfb99",
      "placeholder": "​",
      "style": "IPY_MODEL_c0124b803b5e4e61ae20245eb9777113",
      "value": " 50/50 [00:45&lt;00:00,  1.08it/s]"
     }
    },
    "80244e511a504c929f1d29bd5d148a57": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "8574c57a8e50471fa75d796638a3b19b": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_3e88b235dd8b4461bf056ae966cca361",
      "placeholder": "​",
      "style": "IPY_MODEL_9888aaf95c70421895d3c26c3282043d",
      "value": " 3866/3873 [01:22&lt;00:00, 87.39it/s]"
     }
    },
    "8915d8e524e9456e89d00a4b4dffd2c5": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_8a5553788326463a96a766cf42681954",
       "IPY_MODEL_d1bc6631c4324aa09fc01e7a492726b9",
       "IPY_MODEL_cc6d3ae8b7434d639e9e7a3c2ffe7316"
      ],
      "layout": "IPY_MODEL_d207ec10a648446699f88344b8727589"
     }
    },
    "8a39372c0d7944d4bbd7781e7fdb2eb1": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_ac6f5904475b4c2c9f1454682acbaf45",
       "IPY_MODEL_6e763d19798f410795a810b339b1371c",
       "IPY_MODEL_7ce3b52703a040d38ab4030c80324aef"
      ],
      "layout": "IPY_MODEL_e0c970d3b6154d78a1c08800151b34a0"
     }
    },
    "8a5553788326463a96a766cf42681954": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5baae0dec59047cf9814f8290614a5b2",
      "placeholder": "​",
      "style": "IPY_MODEL_b20c8e72d7cc4a8db0a7aadba2f68dfe",
      "value": "Sampling biased random walks: 100%"
     }
    },
    "8c11267bc29c47929f659e1003a1dfcf": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "8cd3b25c74b14cef95da4ae10778c67c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_4074a73f78534eb2a09caf3f35463e27",
      "placeholder": "​",
      "style": "IPY_MODEL_6375042745a24584a55575854e37b65a",
      "value": "100%"
     }
    },
    "9888aaf95c70421895d3c26c3282043d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "9b0cdbf346114368bdad72acf84bfb99": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "9d4cca973d4f497991327c530dbf42ba": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_3ada169d117048dd828c2b0eb7d7e5d0",
      "max": 50,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_fd62483c077847869732558fd48cc692",
      "value": 50
     }
    },
    "a8981e199f414d4686fef5df98403ffb": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_cb5baf8c7891461f893e827b040223d7",
      "placeholder": "​",
      "style": "IPY_MODEL_660c47186f5f469c9fa56b3bc12dbafb",
      "value": "100%"
     }
    },
    "ac2f589466aa4adc8410c3a33e4e8b12": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_8cd3b25c74b14cef95da4ae10778c67c",
       "IPY_MODEL_1142b2d3ecd1418ea80eb1bb3d1dcbf4",
       "IPY_MODEL_700b1b90991c4d44839f64b1d10025ff"
      ],
      "layout": "IPY_MODEL_43595dbc885742558dd80b1469442550"
     }
    },
    "ac6f5904475b4c2c9f1454682acbaf45": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5f1120e664cc4305b78cefafa63c9b93",
      "placeholder": "​",
      "style": "IPY_MODEL_e5fc97fd76ca425a973d96ede1d08678",
      "value": "Sampling biased random walks: 100%"
     }
    },
    "ae61cb27fa874099a6fea9321cc89cf3": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "b20c8e72d7cc4a8db0a7aadba2f68dfe": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "c0124b803b5e4e61ae20245eb9777113": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "c5cdcc04ae8f4b6f8955aa51f2c5fb84": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_ae61cb27fa874099a6fea9321cc89cf3",
      "placeholder": "​",
      "style": "IPY_MODEL_80244e511a504c929f1d29bd5d148a57",
      "value": "100%"
     }
    },
    "c6e64987178b42dd8553e72c3ab92137": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_a8981e199f414d4686fef5df98403ffb",
       "IPY_MODEL_5767cde28d524f3682b470b585417f11",
       "IPY_MODEL_c704472dadda4263896ec31feed1b2c0"
      ],
      "layout": "IPY_MODEL_f29f288d77054f4f96bbe7ec4a052cd2"
     }
    },
    "c704472dadda4263896ec31feed1b2c0": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e5a9bb19066d47caa04cc8b0ea9ccc5c",
      "placeholder": "​",
      "style": "IPY_MODEL_4ebcffe97f234827b88f8c4d94762262",
      "value": " 50/50 [00:20&lt;00:00,  2.36it/s]"
     }
    },
    "cb5baf8c7891461f893e827b040223d7": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "cc6d3ae8b7434d639e9e7a3c2ffe7316": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e7586c08cee34cf699ba2248c01abdf6",
      "placeholder": "​",
      "style": "IPY_MODEL_1564b81b7ecc450fbdf15b1323a4d565",
      "value": " 3861/3873 [01:32&lt;00:00, 77.55it/s]"
     }
    },
    "d1bc6631c4324aa09fc01e7a492726b9": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_3a87a74057794059b1bda580834dbb2f",
      "max": 3873,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_dbe76ecf5c1248f091b376d980629f83",
      "value": 3873
     }
    },
    "d207ec10a648446699f88344b8727589": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "dbcf92c1a11d428ea730de23cd19f209": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "dbe76ecf5c1248f091b376d980629f83": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "e0c970d3b6154d78a1c08800151b34a0": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e5a9bb19066d47caa04cc8b0ea9ccc5c": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e5fc97fd76ca425a973d96ede1d08678": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "e7586c08cee34cf699ba2248c01abdf6": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "eceaf917f28948a19eff59e769ca7c57": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "eff6e8c5a87e4339adce3342553e1060": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_dbcf92c1a11d428ea730de23cd19f209",
      "placeholder": "​",
      "style": "IPY_MODEL_2d597356e6c940afba19c5166d5660b3",
      "value": " 50/50 [00:20&lt;00:00,  2.41it/s]"
     }
    },
    "f02a455e777f4f4997477f341565ab37": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_c5cdcc04ae8f4b6f8955aa51f2c5fb84",
       "IPY_MODEL_9d4cca973d4f497991327c530dbf42ba",
       "IPY_MODEL_eff6e8c5a87e4339adce3342553e1060"
      ],
      "layout": "IPY_MODEL_0f4371fe938141b9a9131af8c2173ab0"
     }
    },
    "f0e5c7ec07e94626b2c7e1a4f7cfe2e2": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f23d5a29aa064f4a8123be43f70095d9": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "f29f288d77054f4f96bbe7ec4a052cd2": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f4714f36d2b6441f8de4e9cd0256c4ff": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "fd62483c077847869732558fd48cc692": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "fdec3af90eab4faea793d307453d9f69": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}