notebooks/Neo4j_Example2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cytoscape Visualization using a remote Neo4j Database\n",
    "This notebook demonstrates the visualization of subgraphs from the [Neo4j](https://neo4j.com/) Graph Database. It uses the [py2neo](https://py2neo.org/) library to access a Neo4j database instance.\n",
    "\n",
    "The examples in this notebook access the [COVID-19-Net](https:/covid-19-net/covid-19-community) Knowledge Graph.\n",
    "\n",
    "Author: Peter W. Rose (pwrose@ucsd.edu)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import random\n",
    "import ipycytoscape\n",
    "from py2neo import Graph"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Node and edge styles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# style for nodes with node names\n",
    "node_centered = {'selector': 'node',\n",
    "                 'style': {'font-size': '10px',\n",
    "                           'label': 'data(name)',\n",
    "                           'height': '60px',\n",
    "                           'width': '80px',\n",
    "                           'text-max-width': '80px',\n",
    "                           'text-wrap': 'wrap',\n",
    "                           'text-valign': 'center',\n",
    "                           'background-color': 'blue',\n",
    "                           'background-opacity': 0.6}\n",
    "                }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# style for directed graph with arrows\n",
    "edge_directed = {'selector': 'edge',\n",
    "                 'style':  {'line-color': '#9dbaea',\n",
    "                            'curve-style': 'bezier',\n",
    "                            'target-arrow-shape': 'triangle',\n",
    "                            'target-arrow-color': '#9dbaea'}\n",
    "                }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# style for directed graph with arrows and relationship names\n",
    "edge_directed_named = {'selector': 'edge',\n",
    "                       'style':  {'font-size': '8px',\n",
    "                                  'label': 'data(name)',\n",
    "                                  'line-color': '#9dbaea',\n",
    "                                  'text-rotation': 'autorotate',\n",
    "                                  'curve-style': 'bezier',\n",
    "                                  'target-arrow-shape': 'triangle',\n",
    "                                  'target-arrow-color': '#9dbaea'}\n",
    "                      }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# style for undirected graph\n",
    "edge_undirected = {'selector': 'edge',\n",
    "                   'style':  {'line-color': '#9dbaea'}\n",
    "                  }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Node colors\n",
    "Change seed to select a different color palette."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_color_palette(n_colors, seed=3):\n",
    "    \"\"\" \n",
    "    Creates a random color palette of n_colors \n",
    "    See https://stackoverflow.com/questions/28999287/generate-random-colors-rgb)\n",
    "    \n",
    "    \"\"\"\n",
    "    random.seed(seed)\n",
    "    return ['#'+''.join([random.choice('0123456789ABCDEF') for j in range(6)]) for i in range(n_colors)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Connect to a Neo4j Database\n",
    "Here we use the [COVID-19-Net](https:/covid-19-net/covid-19-community) Knowledge Graph to demonstrate how to run a Neo4j Cypher query and pass the resulting subgraph into Cytoscape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "DEMO_SERVER_URL = os.getenv(\"DEMO_SERVER_URL\")\n",
    "DEMO_SERVER_UN = os.getenv(\"DEMO_SERVER_UN\")\n",
    "DEMO_SERVER_PW = os.getenv(\"DEMO_SERVER_PW\")\n",
    "graph = Graph(DEMO_SERVER_URL, user=DEMO_SERVER_UN, password=DEMO_SERVER_PW)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Example 1: Find all cities with the name \"San Francisco\"\n",
    "This query demonstrates the geographic hierachy in COVID-19-Net."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "query1 = \"\"\"\n",
    "MATCH p=(:City{name:'San Francisco'})-[:IN*]->(:World) RETURN p\n",
    "\"\"\"\n",
    "subgraph1 = graph.run(query1).to_subgraph()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add Neo4j subgraph to Cytoscape Widget"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget1 = ipycytoscape.CytoscapeWidget()\n",
    "widget1.graph.add_graph_from_neo4j(subgraph1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Set node and edge styles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "style1 = [node_centered, edge_directed]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Set node specific colors based on node labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Node labels: ['City', 'USRegion', 'Location', 'Admin2', 'UNRegion', 'Admin1', 'UNSubRegion', 'USDivision', 'World', 'UNIntermediateRegion', 'Country']\n"
     ]
    }
   ],
   "source": [
    "labels1 = list(subgraph1.labels())\n",
    "print('Node labels:', labels1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors1 = random_color_palette(len(labels1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label, color in zip(labels1, colors1):\n",
    "    style1.append({'selector': 'node[label = \"' + label + '\"]', 'style': {'background-color': color}})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget1.set_style(style1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget1.set_layout(name='dagre', padding=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When a Neo4j subgraph is added to a Cytoscape graph, a ```tooltip``` attribute is generated that contains all Neo4j node properties."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget1.set_tooltip_source('tooltip')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Click on a node to show the tooltip."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8c67618d3734466290a4066e47895767",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "CytoscapeWidget(cytoscape_layout={'name': 'dagre', 'padding': 0}, cytoscape_style=[{'selector': 'node', 'style…"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "widget1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Example 2: Find all proteins that interact with the SARS-CoV-2 Spike protein\n",
    "Here we run an undirected query (without \"->\" arrow) since the direction of interaction is arbitrary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "query2 = \"\"\"\n",
    "MATCH p=(:Protein{name: 'Spike glycoprotein', taxonomyId: 'taxonomy:2697049'})-[:INTERACTS_WITH]-(:Protein) RETURN p\n",
    "\"\"\"\n",
    "subgraph2 = graph.run(query2).to_subgraph()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget2 = ipycytoscape.CytoscapeWidget()\n",
    "widget2.graph.add_graph_from_neo4j(subgraph2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "style2 = [node_centered, edge_undirected]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Node labels: ['Protein']\n"
     ]
    }
   ],
   "source": [
    "labels2 = list(subgraph2.labels())\n",
    "print('Node labels:', labels2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors2 = random_color_palette(len(labels2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label, color in zip(labels2, colors2):\n",
    "    style2.append({'selector': 'node[label = \"' + label + '\"]', 'style': {'background-color': color}})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget2.set_style(style2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget2.set_layout(name='concentric', padding=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget2.set_tooltip_source('tooltip')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO relationships are not rendered correctly ???"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Click on a node to show the tooltip."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8de31d6dad234a67a14eef4edcde94f7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "CytoscapeWidget(cytoscape_layout={'name': 'concentric', 'padding': 0}, cytoscape_style=[{'selector': 'node', '…"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "widget2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Example 3: Explore the Data Sources used to create the COVID-19-Net Knowledge Graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "query3 = \"\"\"\n",
    "MATCH p=(:MetaNode)-[:ETL_FROM]->(:DataSource) RETURN p  // ETL_FROM: Extracted, transformed, and loaded FROM\n",
    "\"\"\"\n",
    "subgraph3 = graph.run(query3).to_subgraph()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget3 = ipycytoscape.CytoscapeWidget()\n",
    "widget3.graph.add_graph_from_neo4j(subgraph3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "style3 = [node_centered, edge_directed]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Node labels: ['DataSource', 'MetaNode']\n"
     ]
    }
   ],
   "source": [
    "labels3 = list(subgraph3.labels())\n",
    "print('Node labels:', labels3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors3 = random_color_palette(len(labels3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label, color in zip(labels3, colors3):\n",
    "    style3.append({'selector': 'node[label = \"' + label + '\"]', 'style': {'background-color': color}})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget3.set_style(style3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget3.set_layout(name='klay', padding=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget3.set_tooltip_source('tooltip')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Click on a node to show the tooltip."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a8cd215b67354875bdf5d7ec02b638ba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "CytoscapeWidget(cytoscape_layout={'name': 'klay', 'padding': 0}, cytoscape_style=[{'selector': 'node', 'style'…"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "widget3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Example 4: Create a Metagraph that shows the Nodes and their Relationships in the COVID-19-Net Knowledge Graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "query4 = \"\"\"\n",
    "MATCH p=(a:MetaNode)-[:META_RELATIONSHIP]->(b:MetaNode) \n",
    "WHERE a.name <> 'Location' AND b.name <> 'Location' // exclude Location nodes since they make the graph too crowded\n",
    "RETURN p\n",
    "\"\"\"\n",
    "subgraph4 = graph.run(query4).to_subgraph()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget4 = ipycytoscape.CytoscapeWidget()\n",
    "widget4.graph.add_graph_from_neo4j(subgraph4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "style4 = [node_centered, edge_directed_named]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Node labels: ['MetaNode']\n"
     ]
    }
   ],
   "source": [
    "labels4 = list(subgraph4.labels())\n",
    "print('Node labels:', labels4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors4 = random_color_palette(len(labels4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label, color in zip(labels4, colors4):\n",
    "    style4.append({'selector': 'node[label = \"' + label + '\"]', 'style': {'background-color': color}})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget4.set_style(style4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cola layout [options](https:/cytoscape/cytoscape.js-cola#api)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget4.set_layout(name='cola', padding=0, nodeSpacing=65, nodeDimensionsIncludeLabels=True, unconstrIter=5000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "widget4.set_tooltip_source('tooltip')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Click on a node to show the tooltip**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cd4901bce4b74fafb3cf830f333dac8f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "CytoscapeWidget(cytoscape_layout={'name': 'cola', 'padding': 0, 'nodeSpacing': 65, 'nodeDimensionsIncludeLabel…"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "widget4"
   ]
  }
 ],
 "metadata": {
  "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}