examples/impact.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "7067d845-9121-47c1-8bc2-65db8f4d7004",
   "metadata": {},
   "source": [
    "# Senzing + Neo4j: Analyze impact of ERKG"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78939f8c-3e2c-49b0-b447-213cccfd5ea6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-03-15T19:36:02.310335Z",
     "iopub.status.busy": "2024-03-15T19:36:02.303320Z",
     "iopub.status.idle": "2024-03-15T19:36:02.352513Z",
     "shell.execute_reply": "2024-03-15T19:36:02.348131Z",
     "shell.execute_reply.started": "2024-03-15T19:36:02.310230Z"
    }
   },
   "source": [
    "## Set up the Python environment"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea7cc699-a3f7-4768-966a-a7fd77b825c4",
   "metadata": {},
   "source": [
    "First, we need to import the Python library dependencies which are required for the code we'll be running."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c55880de-2f5f-425d-bdfc-6a9cd19f9777",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:39:58.866312Z",
     "iopub.status.busy": "2024-04-12T04:39:58.865710Z",
     "iopub.status.idle": "2024-04-12T04:40:01.290553Z",
     "shell.execute_reply": "2024-04-12T04:40:01.289360Z",
     "shell.execute_reply.started": "2024-04-12T04:39:58.866285Z"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "from graphdatascience import GraphDataScience \n",
    "from icecream import ic\n",
    "from tqdm import tqdm\n",
    "import dotenv\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import pyvis\n",
    "import seaborn as sns\n",
    "import watermark\n",
    "\n",
    "%load_ext watermark"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11331aa0-be80-4de3-bed3-24beb12a57bc",
   "metadata": {},
   "source": [
    "Show a \"watermark\" of which versions are being used for system componenents and library dependencies. This may help in case you need to troubleshoot the dependencies on your system, e.g., if there's some conflict during installation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e6a97a2f-e4c4-4086-9dec-bfbf680fd9b9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:01.292159Z",
     "iopub.status.busy": "2024-04-12T04:40:01.291715Z",
     "iopub.status.idle": "2024-04-12T04:40:01.332613Z",
     "shell.execute_reply": "2024-04-12T04:40:01.330839Z",
     "shell.execute_reply.started": "2024-04-12T04:40:01.292133Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Last updated: 2024-04-11T21:40:01.293760-07:00\n",
      "\n",
      "Python implementation: CPython\n",
      "Python version       : 3.11.0\n",
      "IPython version      : 8.23.0\n",
      "\n",
      "Compiler    : Clang 13.0.0 (clang-1300.0.29.30)\n",
      "OS          : Darwin\n",
      "Release     : 21.6.0\n",
      "Machine     : x86_64\n",
      "Processor   : i386\n",
      "CPU cores   : 8\n",
      "Architecture: 64bit\n",
      "\n",
      "pandas    : 2.2.1\n",
      "matplotlib: 3.8.4\n",
      "seaborn   : 0.13.2\n",
      "watermark : 2.4.3\n",
      "pyvis     : 0.3.2\n",
      "\n"
     ]
    }
   ],
   "source": [
    "%watermark\n",
    "%watermark --iversions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "915d34b2-c582-412f-86ad-8c569387f5ca",
   "metadata": {},
   "source": [
    "## Connect the GDS library to Neo4j Desktop"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "184191bb-8761-4d96-b311-21df568df292",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-03-26T21:56:48.062843Z",
     "iopub.status.busy": "2024-03-26T21:56:48.062487Z",
     "iopub.status.idle": "2024-03-26T21:56:48.111484Z",
     "shell.execute_reply": "2024-03-26T21:56:48.100460Z",
     "shell.execute_reply.started": "2024-03-26T21:56:48.062818Z"
    }
   },
   "source": [
    "Set up a GDS connection using our credentials for Neo4j Desktop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1bc0cacc-5534-4535-ab7b-1790f0c45e78",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:01.336090Z",
     "iopub.status.busy": "2024-04-12T04:40:01.335506Z",
     "iopub.status.idle": "2024-04-12T04:40:02.070411Z",
     "shell.execute_reply": "2024-04-12T04:40:02.069853Z",
     "shell.execute_reply.started": "2024-04-12T04:40:01.336059Z"
    }
   },
   "outputs": [],
   "source": [
    "dotenv.load_dotenv(dotenv.find_dotenv())\n",
    "\n",
    "bolt_uri: str = os.environ.get(\"NEO4J_BOLT\")\n",
    "database: str = os.environ.get(\"NEO4J_DBMS\")\n",
    "username: str = os.environ.get(\"NEO4J_USER\")\n",
    "password: str = os.environ.get(\"NEO4J_PASS\")\n",
    "\n",
    "gds:GraphDataScience = GraphDataScience(\n",
    "    bolt_uri,\n",
    "    auth = ( username, password, ),\n",
    "    database = database,\n",
    "    aura_ds = False,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ae500c9-3f7c-48c8-b734-e89a88e05706",
   "metadata": {},
   "source": [
    "## Analyze the impact of ER"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da90a1fa-ab92-4976-bee9-a1c467d78173",
   "metadata": {},
   "source": [
    "Now let's analyze the Senzing results, measuring how much the process of _entity resolution_ has consolidated records among the input datasets.\n",
    "In particular, it's helpful to understand:\n",
    "\n",
    "  - total number of entities\n",
    "  - number of entities with linked reference (should be 100% in the graph)\n",
    "  - number of linked records per entity\n",
    "  - number of related entities per entity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b2905552-40dc-4e19-8acb-78a8179d7611",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:02.071831Z",
     "iopub.status.busy": "2024-04-12T04:40:02.071241Z",
     "iopub.status.idle": "2024-04-12T04:40:06.183294Z",
     "shell.execute_reply": "2024-04-12T04:40:06.182662Z",
     "shell.execute_reply.started": "2024-04-12T04:40:02.071804Z"
    }
   },
   "outputs": [],
   "source": [
    "df_ent: pd.DataFrame = gds.run_cypher(\n",
    "  \"\"\"\n",
    "MATCH (ent:Entity)\n",
    "RETURN ent.uid AS ent_uid, COUNT { (ent)-[:RESOLVES]->(:Record) } AS num_rec, COUNT { (ent)-[:RELATED]->(:Entity) } AS num_rel\n",
    "ORDER BY num_rec DESC\n",
    "  \"\"\"\n",
    ")\n",
    "\n",
    "df_rec: pd.DataFrame = gds.run_cypher(\n",
    "  \"\"\"\n",
    "MATCH (ent:Entity)\n",
    "RETURN COUNT(ent.uid) AS count_ent, COUNT { (ent)-[:RESOLVES]->(:Record) } AS num_rec\n",
    "ORDER BY num_rec DESC\n",
    "  \"\"\"\n",
    ")\n",
    "\n",
    "df_rel: pd.DataFrame = gds.run_cypher(\n",
    "  \"\"\"\n",
    "MATCH (ent:Entity)\n",
    "RETURN COUNT(ent.uid) AS count_ent, COUNT { (ent)-[:RELATED]->(:Entity) } AS num_rel\n",
    "ORDER BY num_rel DESC\n",
    "  \"\"\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "0e8c6d36-a02e-45ee-b704-e5cf7e75ba26",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:06.184565Z",
     "iopub.status.busy": "2024-04-12T04:40:06.184040Z",
     "iopub.status.idle": "2024-04-12T04:40:06.346230Z",
     "shell.execute_reply": "2024-04-12T04:40:06.345081Z",
     "shell.execute_reply.started": "2024-04-12T04:40:06.184544Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ic| total_entities: 83386\n",
      "    total_records: 84988\n",
      "    unref_ents: 0\n",
      "    dupl_rate: 0.02\n"
     ]
    }
   ],
   "source": [
    "total_entities: int = df_ent.ent_uid.count()\n",
    "total_records: int = df_ent.num_rec.sum()\n",
    "unref_ents: int = df_rec.loc[df_rec[\"num_rec\"] < 1].count().iloc[0]\n",
    "dupl_rate: float = round((total_records - total_entities) / float(total_entities), 2)\n",
    "\n",
    "ic(total_entities, total_records, unref_ents, dupl_rate);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d65c19ae-ae59-4efb-9556-4f1186cea67d",
   "metadata": {},
   "source": [
    "Great, 100% of the input records have been resolved as entities – as expected.\n",
    "A rate of ~2% duplicates was detected among these records."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c936201-c852-41b2-abff-8194a19e2552",
   "metadata": {},
   "source": [
    "Now let's run more detailed statistical analysis, visualizing a histogram of the resolved entities versus their related records from the input datasets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "80f4c0dd-d318-49eb-b8eb-06a798771e76",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:06.350521Z",
     "iopub.status.busy": "2024-04-12T04:40:06.350201Z",
     "iopub.status.idle": "2024-04-12T04:40:06.385038Z",
     "shell.execute_reply": "2024-04-12T04:40:06.384540Z",
     "shell.execute_reply.started": "2024-04-12T04:40:06.350500Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ent_uid</th>\n",
       "      <th>num_rec</th>\n",
       "      <th>num_rel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>83386.000000</td>\n",
       "      <td>83386.000000</td>\n",
       "      <td>83386.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>45603.987204</td>\n",
       "      <td>1.019212</td>\n",
       "      <td>4.134459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>36222.739725</td>\n",
       "      <td>0.173885</td>\n",
       "      <td>6.380714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>20905.250000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>41770.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>62639.750000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>6.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>203487.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>326.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             ent_uid       num_rec       num_rel\n",
       "count   83386.000000  83386.000000  83386.000000\n",
       "mean    45603.987204      1.019212      4.134459\n",
       "std     36222.739725      0.173885      6.380714\n",
       "min         1.000000      1.000000      0.000000\n",
       "25%     20905.250000      1.000000      1.000000\n",
       "50%     41770.500000      1.000000      3.000000\n",
       "75%     62639.750000      1.000000      6.000000\n",
       "max    203487.000000     14.000000    326.000000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_ent.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ac11fdfb-3d10-4f81-b992-ff1801cbc690",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:06.386114Z",
     "iopub.status.busy": "2024-04-12T04:40:06.385941Z",
     "iopub.status.idle": "2024-04-12T04:40:07.010585Z",
     "shell.execute_reply": "2024-04-12T04:40:07.009839Z",
     "shell.execute_reply.started": "2024-04-12T04:40:06.386097Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "plt.rcParams[\"font.family\"] = \"sans-serif\"\n",
    "\n",
    "y = sns.barplot(df_rec, y = \"count_ent\", x = \"num_rec\")\n",
    "y.tick_params(axis = \"y\", size = 9, colors = \"gray\")\n",
    "y.bar_label(y.containers[0], padding = 3, color = \"gray\", fontsize = 7)\n",
    "\n",
    "plt.xlabel(\"records per entity\", size = 10, fontstyle = \"italic\")\n",
    "plt.ylabel(\"entity count\", size = 10, fontstyle = \"italic\")\n",
    "\n",
    "rec_mean: float = df_ent.num_rec.mean()\n",
    "plt.axvline(x = rec_mean, color = \"red\", lw = 2.5)\n",
    "\n",
    "sns.despine(bottom = True, left = True)\n",
    "plt.yscale(\"log\")\n",
    "\n",
    "fig.savefig(\"../article/img/graphs.plot.ent_rec.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e015a2b9-71e6-4844-890f-a1af043d1619",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:07.011909Z",
     "iopub.status.busy": "2024-04-12T04:40:07.011633Z",
     "iopub.status.idle": "2024-04-12T04:40:07.381592Z",
     "shell.execute_reply": "2024-04-12T04:40:07.381035Z",
     "shell.execute_reply.started": "2024-04-12T04:40:07.011859Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "plt.rcParams[\"font.family\"] = \"sans-serif\"\n",
    "\n",
    "y = sns.scatterplot(df_rel, y = \"count_ent\", x = \"num_rel\")\n",
    "\n",
    "plt.xlabel(\"related entities per entity\", size = 10, fontstyle = \"italic\")\n",
    "plt.ylabel(\"entity count\", size = 10, fontstyle = \"italic\")\n",
    "\n",
    "rel_mean: float = df_ent.num_rel.mean()\n",
    "plt.axvline(x = rel_mean, color = \"red\", lw = 2.5)\n",
    "\n",
    "sns.despine(bottom = True, left = True)\n",
    "plt.yscale(\"log\")\n",
    "\n",
    "fig.savefig(\"../article/img/graphs.plot.ent_rel.png\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0e7e638e-0778-48a6-a273-11df768d446f",
   "metadata": {},
   "source": [
    "## Graph visualizations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "002b8a7c-76c4-4d31-908d-8f648d9f3bf5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:07.383269Z",
     "iopub.status.busy": "2024-04-12T04:40:07.383031Z",
     "iopub.status.idle": "2024-04-12T04:40:11.335267Z",
     "shell.execute_reply": "2024-04-12T04:40:11.334336Z",
     "shell.execute_reply.started": "2024-04-12T04:40:07.383251Z"
    }
   },
   "outputs": [],
   "source": [
    "df_link_rec = gds.run_cypher(\n",
    "  \"\"\"\n",
    "MATCH (ent:Entity)-[rel:RESOLVES]->(rec:Record)\n",
    "RETURN ent.uid AS ent_uid, ent.name AS ent_name, COUNT { (ent)-[:RESOLVES]->(:Record) } as num_rec, rec.uid AS rec_uid, rec.name AS rec_name, rel.match_key AS match_key\n",
    "  \"\"\"\n",
    ")\n",
    "\n",
    "MIN_DEGREE: int = 3  # adjust lower to view a larger graph\n",
    "\n",
    "df_link_rec = df_link_rec.loc[df_link_rec[\"num_rec\"] >= MIN_DEGREE]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6fb44630-12e8-42be-bbbf-7bfb8c64694e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:11.336749Z",
     "iopub.status.busy": "2024-04-12T04:40:11.336507Z",
     "iopub.status.idle": "2024-04-12T04:40:11.394104Z",
     "shell.execute_reply": "2024-04-12T04:40:11.393274Z",
     "shell.execute_reply.started": "2024-04-12T04:40:11.336727Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: When  cdn_resources is 'local' jupyter notebook has issues displaying graphics on chrome/safari. Use cdn_resources='in_line' or cdn_resources='remote' if you have issues viewing graphics in a notebook.\n",
      "vegas.1.html\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"100%\"\n",
       "            height=\"600px\"\n",
       "            src=\"vegas.1.html\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "            \n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x1260bd7d0>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "net: pyvis.network.Network = pyvis.network.Network(notebook = True)\n",
    "\n",
    "for _, row in df_link_rec.iterrows():\n",
    "    net.add_node(\n",
    "        row.rec_uid,\n",
    "        label = row.rec_name,\n",
    "        color = \"blue\",\n",
    "        shape = \"square\",\n",
    "    )\n",
    "\n",
    "net.toggle_physics(True)\n",
    "net.show(\"vegas.1.html\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f0cf23d5-9d7b-4f9d-86d8-e934fe20c795",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:11.395435Z",
     "iopub.status.busy": "2024-04-12T04:40:11.395234Z",
     "iopub.status.idle": "2024-04-12T04:40:24.334510Z",
     "shell.execute_reply": "2024-04-12T04:40:24.333633Z",
     "shell.execute_reply.started": "2024-04-12T04:40:11.395419Z"
    }
   },
   "outputs": [],
   "source": [
    "df_link_rel: pd.DataFrame = gds.run_cypher(\n",
    "  \"\"\"\n",
    "MATCH (ent:Entity)-[rel:RELATED]->(rel_ent:Entity)\n",
    "RETURN ent.uid AS ent_uid, rel.match_key AS match_key, rel_ent.uid AS rel_uid, rel_ent.name AS rel_name\n",
    "  \"\"\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "1ab87df2-b5bd-412a-82fb-7d5496b73745",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-12T04:40:24.335966Z",
     "iopub.status.busy": "2024-04-12T04:40:24.335756Z",
     "iopub.status.idle": "2024-04-12T04:40:53.047365Z",
     "shell.execute_reply": "2024-04-12T04:40:53.043119Z",
     "shell.execute_reply.started": "2024-04-12T04:40:24.335946Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: When  cdn_resources is 'local' jupyter notebook has issues displaying graphics on chrome/safari. Use cdn_resources='in_line' or cdn_resources='remote' if you have issues viewing graphics in a notebook.\n",
      "vegas.2.html\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"100%\"\n",
       "            height=\"600px\"\n",
       "            src=\"vegas.2.html\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "            \n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x126a45310>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "net: pyvis.network.Network = pyvis.network.Network(notebook = True)\n",
    "\n",
    "for _, row in df_link_rec.iterrows():\n",
    "    net.add_node(\n",
    "        row.rec_uid,\n",
    "        label = row.rec_name,\n",
    "        color = \"blue\",\n",
    "        shape = \"square\",\n",
    "        size = 10,\n",
    "    )\n",
    "    \n",
    "    net.add_node(\n",
    "        row.ent_uid,\n",
    "        label = row.ent_name,\n",
    "        color = \"red\",\n",
    "        shape = \"star\",\n",
    "        size = row.num_rec,\n",
    "    )\n",
    "    \n",
    "    edge = net.add_edge(\n",
    "        row.rec_uid,\n",
    "        row.ent_uid,\n",
    "        label = row.match_key,\n",
    "        color = \"green\",\n",
    "        #value = 1.0,\n",
    "    )\n",
    "\n",
    "# enumerate the nodes, so we can do efficient lookups later\n",
    "node_ids: dict = {\n",
    "    node[\"id\"]: i\n",
    "    for i, node in enumerate(net.nodes)\n",
    "}\n",
    "\n",
    "# link the nodes/edges for \"RELATED\" entities\n",
    "for _, row in df_link_rel.iterrows():\n",
    "    if row.ent_uid in node_ids:\n",
    "        if row.rel_uid not in node_ids:\n",
    "            net.add_node(\n",
    "                row.rel_uid,\n",
    "                label = row.rel_name,\n",
    "                color = \"gray\",\n",
    "                shape = \"star\",\n",
    "                size = 4,\n",
    "            )\n",
    "\n",
    "        edge = net.add_edge(\n",
    "            row.ent_uid,\n",
    "            row.rel_uid,\n",
    "            title = row.match_key,\n",
    "            color = \"#d3d3d3\",\n",
    "        )\n",
    "\n",
    "# add titles (flyover HTML) to show the properties for each node\n",
    "df_recs: pd.DataFrame = gds.run_cypher(\"\"\"\n",
    "MATCH (ent:Entity)-[rel:RESOLVES]->(rec:Record)\n",
    "RETURN rec.uid AS rec_uid, properties(rec) as props\n",
    "\"\"\")\n",
    "\n",
    "for _, row in df_recs.iterrows():\n",
    "    if row.rec_uid in node_ids:\n",
    "        title: str = \"\\n\".join([\n",
    "            f\"{key}: {val}\"\n",
    "            for key, val in row.props.items()\n",
    "        ])\n",
    "        net.nodes[node_ids[row.rec_uid]][\"title\"] = title\n",
    "\n",
    "# set the font size and color for nodes and edges\n",
    "for node in net.nodes:\n",
    "    if \"label\" in node:\n",
    "        node[\"font\"] = {\n",
    "            \"size\": 9,\n",
    "            \"color\": \"gray\",\n",
    "        }\n",
    "\n",
    "for edge in net.edges:\n",
    "    if \"label\" in edge:\n",
    "        edge[\"font\"] = {\n",
    "            \"size\": 8,\n",
    "            \"color\": \"green\",\n",
    "        }\n",
    "\n",
    "# visualize as HTML\n",
    "net.toggle_physics(True)\n",
    "net.show(\"vegas.2.html\")"
   ]
  }
 ],
 "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}