{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analyzing and Comparing EMLO Collections\n",
    "\n",
    "The EMLO project contains dozens of correspondence collections centered around different historical figures. Each collection is maintained either by a single institute, or is a merging of smaller collections maintained across multiple institutions. \n",
    "\n",
    "The metadata of the correspondences has been mapped to a single schema.\n",
    "\n",
    "Making a comparison of different sets of correspondences, at different scales, draws the focus on different aspects of comparison. At the same time, it brings to the surface some differences in how the digital collections were shaped by selection criteria.\n",
    "\n",
    "At a small scale, it is easy to see for instance that a collection around a historical figure, e.g. Samuel Hartlib or Françoise de Graffigny has not only letters authored by or addressed to that figure, but also some letters between the correspondents in their networks. When working with many correspondence collections with thousands or tens of thousands of letters, this is a detail that is easily lost in overviews of metadata records and most summary statistics.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import glob\n",
    "import matplotlib.pyplot as plt\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's first load the data into a dataframe and inspect a number of rows so we get an idea of what is in there."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>id</th>\n",
       "      <th>type</th>\n",
       "      <th>collection</th>\n",
       "      <th>date</th>\n",
       "      <th>author</th>\n",
       "      <th>addressee</th>\n",
       "      <th>origin</th>\n",
       "      <th>destination</th>\n",
       "      <th>repository</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>577c226e-adfa-43ed-8db2-95792b73f3c3</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bayle, Pierre</td>\n",
       "      <td>21 February 1662</td>\n",
       "      <td>Bayle, Jacob, 1644-1685</td>\n",
       "      <td>Bayle, Jean, 1609-1685</td>\n",
       "      <td>Puylaurens, Occitanie, France</td>\n",
       "      <td>Carla-Bayle, Occitanie, France</td>\n",
       "      <td>Collection d'E. Labrousse\\n      \\n        \\n ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0ccd18c5-de76-4c66-a7b5-fbc344a9dcde</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bayle, Pierre</td>\n",
       "      <td>17 June 1662</td>\n",
       "      <td>Bayle, Jacob, 1644-1685</td>\n",
       "      <td>Bayle, Jean, 1609-1685</td>\n",
       "      <td>Puylaurens, Occitanie, France</td>\n",
       "      <td>Carla-Bayle, Occitanie, France</td>\n",
       "      <td>2 printed editions</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>2e189892-8105-42e9-876c-02aaa97fb877</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bayle, Pierre</td>\n",
       "      <td>25 August 1662</td>\n",
       "      <td>Bayle, Jacob, 1644-1685</td>\n",
       "      <td>Bayle, Jean, 1609-1685</td>\n",
       "      <td>Puylaurens, Occitanie, France</td>\n",
       "      <td>Carla-Bayle, Occitanie, France</td>\n",
       "      <td>Collection d'E. Labrousse\\n      \\n        \\n ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>da35bb30-cfe1-4b4a-957d-42742088e693</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bayle, Pierre</td>\n",
       "      <td>5 March 1663</td>\n",
       "      <td>Bayle, Jacob, 1644-1685</td>\n",
       "      <td>Bayle, Pierre, 1647-1706</td>\n",
       "      <td>Puylaurens, Occitanie, France</td>\n",
       "      <td>Carla-Bayle, Occitanie, France</td>\n",
       "      <td>1 printed edition</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>c5fefa45-f197-4bc3-bf52-1272f7f1d573</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bayle, Pierre</td>\n",
       "      <td>7 April 1665</td>\n",
       "      <td>Bayle, Jacob, 1644-1685</td>\n",
       "      <td>Bayle, Jean, 1609-1685</td>\n",
       "      <td>Puylaurens, Occitanie, France</td>\n",
       "      <td>Carla-Bayle, Occitanie, France</td>\n",
       "      <td>Bibliothèque de la  Société de l'Histoire du P...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132205</th>\n",
       "      <td>23</td>\n",
       "      <td>a9088a3f-3eb2-4039-aeaa-8ad98b9e6f35</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Beeckman, Isaac</td>\n",
       "      <td>12 April 1632</td>\n",
       "      <td>Croix, Jacques, 1579-1655</td>\n",
       "      <td>Beeckman, Isaac, 1588-1637</td>\n",
       "      <td>Delft, South Holland, (United Provinces) Nethe...</td>\n",
       "      <td>Dordrecht, South Holland, (United Provinces) N...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132206</th>\n",
       "      <td>24</td>\n",
       "      <td>1759be67-e3d3-4e30-a987-b0f84b212f03</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Beeckman, Isaac</td>\n",
       "      <td>17 May 1632</td>\n",
       "      <td>Beeckman, Isaac, 1588-1637</td>\n",
       "      <td>Rivet, André, 1572-1651</td>\n",
       "      <td>Dordrecht, South Holland, (United Provinces) N...</td>\n",
       "      <td>The Hague, South Holland, Netherlands</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132207</th>\n",
       "      <td>25</td>\n",
       "      <td>c1c3e71d-9ca1-4953-968a-b5a8e5d3c00b</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Beeckman, Isaac</td>\n",
       "      <td>30 May 1633</td>\n",
       "      <td>Beeckman, Isaac, 1588-1637</td>\n",
       "      <td>Mersenne, Marin, 1588-1648</td>\n",
       "      <td>Dordrecht, South Holland, (United Provinces) N...</td>\n",
       "      <td>Paris, Île-de-France, France</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132208</th>\n",
       "      <td>26</td>\n",
       "      <td>1b82d6d1-8b0c-4bc2-aa57-c856c73c2e66</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Beeckman, Isaac</td>\n",
       "      <td>22 August 1634</td>\n",
       "      <td>Descartes, René, 1596-1650</td>\n",
       "      <td>Beeckman, Isaac, 1588-1637</td>\n",
       "      <td>Amsterdam, North Holland, (United Provinces) N...</td>\n",
       "      <td>Dordrecht, South Holland, (United Provinces) N...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132209</th>\n",
       "      <td>27</td>\n",
       "      <td>06dc5c13-cec1-4fcd-972e-0d6052350645</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Beeckman, Isaac</td>\n",
       "      <td>13 February 1635</td>\n",
       "      <td>Beeckman, Isaac, 1588-1637</td>\n",
       "      <td>Beeckman, Abraham, fl. 1635</td>\n",
       "      <td>Dordrecht, South Holland, (United Provinces) N...</td>\n",
       "      <td>Amsterdam, North Holland, (United Provinces) N...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>132210 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Unnamed: 0                                    id    type  \\\n",
       "0                0  577c226e-adfa-43ed-8db2-95792b73f3c3  Letter   \n",
       "1                1  0ccd18c5-de76-4c66-a7b5-fbc344a9dcde  Letter   \n",
       "2                2  2e189892-8105-42e9-876c-02aaa97fb877  Letter   \n",
       "3                3  da35bb30-cfe1-4b4a-957d-42742088e693  Letter   \n",
       "4                4  c5fefa45-f197-4bc3-bf52-1272f7f1d573  Letter   \n",
       "...            ...                                   ...     ...   \n",
       "132205          23  a9088a3f-3eb2-4039-aeaa-8ad98b9e6f35  Letter   \n",
       "132206          24  1759be67-e3d3-4e30-a987-b0f84b212f03  Letter   \n",
       "132207          25  c1c3e71d-9ca1-4953-968a-b5a8e5d3c00b  Letter   \n",
       "132208          26  1b82d6d1-8b0c-4bc2-aa57-c856c73c2e66  Letter   \n",
       "132209          27  06dc5c13-cec1-4fcd-972e-0d6052350645  Letter   \n",
       "\n",
       "             collection              date                      author  \\\n",
       "0         Bayle, Pierre  21 February 1662     Bayle, Jacob, 1644-1685   \n",
       "1         Bayle, Pierre      17 June 1662     Bayle, Jacob, 1644-1685   \n",
       "2         Bayle, Pierre    25 August 1662     Bayle, Jacob, 1644-1685   \n",
       "3         Bayle, Pierre      5 March 1663     Bayle, Jacob, 1644-1685   \n",
       "4         Bayle, Pierre      7 April 1665     Bayle, Jacob, 1644-1685   \n",
       "...                 ...               ...                         ...   \n",
       "132205  Beeckman, Isaac     12 April 1632   Croix, Jacques, 1579-1655   \n",
       "132206  Beeckman, Isaac       17 May 1632  Beeckman, Isaac, 1588-1637   \n",
       "132207  Beeckman, Isaac       30 May 1633  Beeckman, Isaac, 1588-1637   \n",
       "132208  Beeckman, Isaac    22 August 1634  Descartes, René, 1596-1650   \n",
       "132209  Beeckman, Isaac  13 February 1635  Beeckman, Isaac, 1588-1637   \n",
       "\n",
       "                          addressee  \\\n",
       "0            Bayle, Jean, 1609-1685   \n",
       "1            Bayle, Jean, 1609-1685   \n",
       "2            Bayle, Jean, 1609-1685   \n",
       "3          Bayle, Pierre, 1647-1706   \n",
       "4            Bayle, Jean, 1609-1685   \n",
       "...                             ...   \n",
       "132205   Beeckman, Isaac, 1588-1637   \n",
       "132206      Rivet, André, 1572-1651   \n",
       "132207   Mersenne, Marin, 1588-1648   \n",
       "132208   Beeckman, Isaac, 1588-1637   \n",
       "132209  Beeckman, Abraham, fl. 1635   \n",
       "\n",
       "                                                   origin  \\\n",
       "0                           Puylaurens, Occitanie, France   \n",
       "1                           Puylaurens, Occitanie, France   \n",
       "2                           Puylaurens, Occitanie, France   \n",
       "3                           Puylaurens, Occitanie, France   \n",
       "4                           Puylaurens, Occitanie, France   \n",
       "...                                                   ...   \n",
       "132205  Delft, South Holland, (United Provinces) Nethe...   \n",
       "132206  Dordrecht, South Holland, (United Provinces) N...   \n",
       "132207  Dordrecht, South Holland, (United Provinces) N...   \n",
       "132208  Amsterdam, North Holland, (United Provinces) N...   \n",
       "132209  Dordrecht, South Holland, (United Provinces) N...   \n",
       "\n",
       "                                              destination  \\\n",
       "0                          Carla-Bayle, Occitanie, France   \n",
       "1                          Carla-Bayle, Occitanie, France   \n",
       "2                          Carla-Bayle, Occitanie, France   \n",
       "3                          Carla-Bayle, Occitanie, France   \n",
       "4                          Carla-Bayle, Occitanie, France   \n",
       "...                                                   ...   \n",
       "132205  Dordrecht, South Holland, (United Provinces) N...   \n",
       "132206              The Hague, South Holland, Netherlands   \n",
       "132207                       Paris, Île-de-France, France   \n",
       "132208  Dordrecht, South Holland, (United Provinces) N...   \n",
       "132209  Amsterdam, North Holland, (United Provinces) N...   \n",
       "\n",
       "                                               repository  \n",
       "0       Collection d'E. Labrousse\\n      \\n        \\n ...  \n",
       "1                                      2 printed editions  \n",
       "2       Collection d'E. Labrousse\\n      \\n        \\n ...  \n",
       "3                                       1 printed edition  \n",
       "4       Bibliothèque de la  Société de l'Histoire du P...  \n",
       "...                                                   ...  \n",
       "132205                                                NaN  \n",
       "132206                                                NaN  \n",
       "132207                                                NaN  \n",
       "132208                                                NaN  \n",
       "132209                                                NaN  \n",
       "\n",
       "[132210 rows x 10 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# read the merged letters file into a Pandas dataframe\n",
    "merged_letters_file = '../data/emlo_letters.csv'\n",
    "df = pd.read_csv(merged_letters_file, sep='\\t')\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of distinct authors: 14619\n",
      "number of distinct addressees: 7649\n"
     ]
    }
   ],
   "source": [
    "# show the nnnumber of authors and addressees\n",
    "print('number of distinct authors:', df['author'].nunique())\n",
    "print('number of distinct addressees:', df['addressee'].nunique())\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Bodleian card catalogue     48668\n",
       "Groot, Hugo de               8034\n",
       "Huygens, Constantijn         7120\n",
       "Hartlib, Samuel              4719\n",
       "Andreae, Johann Valentin     3696\n",
       "                            ...  \n",
       "Beeckman, Isaac                28\n",
       "Dudley, Anne                   27\n",
       "Vernon, Margaret               21\n",
       "Baxter, Richard                 8\n",
       "Culpeper, Cheney                3\n",
       "Name: collection, Length: 93, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The correspondence collections in the dataset with the number of letters they contain.\n",
    "df.collection.value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de, 1583-1645                                     4912\n",
       "Huygens, Constantijn, 1596-1687                               3951\n",
       "Plantin, Christophe, 1520-1589                                2331\n",
       "Vossius, Gerardus Joannes, 1577-1649                          2292\n",
       "Peiresc, Nicolas-Claude Fabri de, 1580-1637                   2111\n",
       "                                                              ... \n",
       "Farley (Mr), fl. 1775; Rose (Mr), fl. 1775                       1\n",
       "Newton, James, fl. 1710-1713                                     1\n",
       "Martinus, William, fl. 1621                                      1\n",
       "Council of State of the Republic of the United Netherlands       1\n",
       "Montagu, Edward, 1562-1644                                       1\n",
       "Name: author, Length: 14619, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The authors in the dataset with the number of letters they sent.\n",
    "df.author.value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Huygens, Constantijn, 1596-1687                                     4737\n",
       "Hearne, Thomas, 1678-1735                                           3795\n",
       "Vossius, Gerardus Joannes, 1577-1649                                3498\n",
       "Hartlib, Samuel, 1600-1662                                          3388\n",
       "Groot, Hugo de, 1583-1645                                           3233\n",
       "                                                                    ... \n",
       "Blount, Edward, fl. 1724                                               1\n",
       "Baert, Pieter J., fl. 1676-1691                                        1\n",
       "Werndeley, (Reverend Mr), fl. 1711; Werndeley, sons of, fl. 1711       1\n",
       "Vossius, Gerardus, 1619-1640                                           1\n",
       "Heidelberg, ministers in, fl. 1655                                     1\n",
       "Name: addressee, Length: 7649, dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The addressee in the dataset with the number of letters they received.\n",
    "df.addressee.value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of distinct authors: 14619\n",
      "number of distinct addressees: 7649\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Adjust the default size for figures so that placing two plots \n",
    "# next to each other in a sub plot are still big enough.\n",
    "plt.rcParams['figure.figsize'] = [15, 5]\n",
    "\n",
    "# create a plot canvas with two adjacent subplots\n",
    "plt.subplot(1,2,1)\n",
    "# Distribution of number of letters per author\n",
    "# Sub-plot 1 shows the number of letters by each letter author on normal scaled axes\n",
    "df['author'].value_counts().hist(bins=100)\n",
    "plt.ylabel('Number of authors')\n",
    "plt.xlabel('Number of letters authored')\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "# Sub-plot 1 shows the number of letters by each letter author on a log scaled y-axis\n",
    "df['author'].value_counts().hist(bins=100)\n",
    "plt.ylabel('Number of authors')\n",
    "plt.xlabel('Number of letters authored')\n",
    "plt.yscale('log')\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(1,2,1)\n",
    "# Number of letters by each letter addressee\n",
    "df['addressee'].value_counts().hist(bins=100)\n",
    "plt.xlabel('Number of letters received')\n",
    "plt.ylabel('Number of addressees')\n",
    "\n",
    "\n",
    "# Distribution of number of letters per addressee\n",
    "plt.subplot(1,2,2)\n",
    "df['addressee'].value_counts().hist(bins=100)\n",
    "plt.ylabel('Number of addressees')\n",
    "plt.xlabel('Number of letters received')\n",
    "plt.yscale('log')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(1,2,1)\n",
    "# Number of letters by each letter addressee\n",
    "df['author'].value_counts().hist(bins=100)\n",
    "plt.ylabel('Number of authors')\n",
    "plt.xlabel('Number of letters authored')\n",
    "plt.yscale('log')\n",
    "\n",
    "\n",
    "# Distribution of number of letters per addressee\n",
    "plt.subplot(1,2,2)\n",
    "df['addressee'].value_counts().hist(bins=100)\n",
    "plt.ylabel('Number of addressees')\n",
    "plt.xlabel('Number of letters received')\n",
    "plt.yscale('log')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import Counter\n",
    "\n",
    "author_dist = Counter([count for count in df['author'].value_counts()])\n",
    "x_author, y_author = zip(*author_dist.items())\n",
    "plt.subplot(1,2,1)\n",
    "plt.scatter(x_author, y_author)\n",
    "plt.xscale('log')\n",
    "plt.yscale('log')\n",
    "plt.xlabel('Number of letters authored')\n",
    "plt.ylabel('Number of authors')\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "addressee_dist = Counter([count for count in df['addressee'].value_counts()])\n",
    "x_addressee, y_addressee = zip(*addressee_dist.items())\n",
    "plt.scatter(x_addressee, y_addressee)\n",
    "plt.xscale('log')\n",
    "plt.yscale('log')\n",
    "plt.xlabel('Number of letters received')\n",
    "plt.ylabel('Number of addressees')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The plots show typical skewed distributions. The vast majority of correspondents author and/or receive only one or a few letters (the hight bar on the left of each figure represents all authors/addressees authoring or receiving only one letter). Only a handful of people author or receive more than a thousand letters. \n",
    "\n",
    "Who are the most prolific authors?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de, 1583-1645                      4912\n",
       "Huygens, Constantijn, 1596-1687                3951\n",
       "Plantin, Christophe, 1520-1589                 2331\n",
       "Vossius, Gerardus Joannes, 1577-1649           2292\n",
       "Peiresc, Nicolas-Claude Fabri de, 1580-1637    2111\n",
       "Oldenburg, Henry, 1619-1677                    2092\n",
       "Scaliger, Joseph Justus, 1540-1609             1847\n",
       "Huygens, Christiaan, 1629-1695                 1423\n",
       "Hearne, Thomas, 1678-1735                      1382\n",
       "Wallis, John (Dr), 1616-1703                   1290\n",
       "Smith, Thomas (Dr), 1638-1710                  1114\n",
       "Dury, John, 1596-1680                          1003\n",
       "Bayle, Pierre, 1647-1706                        938\n",
       "Bourignon, Antoinette, 1616-1680                887\n",
       "Montagu, Mary Wortley (Lady), 1689-1762         860\n",
       "Aubrey, John, 1626-1697                         854\n",
       "Heinsius, Nicolaas, 1620-1681                   833\n",
       "Descartes, René, 1596-1650                      796\n",
       "Vossius, Isaac (Dr), 1618-1689                  793\n",
       "Brett, Thomas, 1667-1744                        742\n",
       "Name: author, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['author'].value_counts().head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the list above, most of the authors are the central figure or eponym of one of the EMLO collections.\n",
    "\n",
    "Exceptions are:\n",
    "\n",
    "- John Dury (1596-1650): Preacher and ecumenist\n",
    "- August II of Braunschweig-Wolfenbüttel (1579-1666): Duke (Herzog) of Braunschweig-Wolfenbüttel\n",
    "\n",
    "These are prolific authors in collections centred on someone else. \n",
    "\n",
    "We first look at the letters of August II. Which collections are they part of?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collections with August II of Braunschweig-Wolfenbüttel's letters:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Andreae, Johann Valentin                        582\n",
       "Kircher, Athanasius                              21\n",
       "Bodleian card catalogue                           1\n",
       "Braunschweig-Wolfenbüttel, Sophia Hedwig von      1\n",
       "Name: collection, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Collections with August II of Braunschweig-Wolfenbüttel's letters:\")\n",
    "df[df['author'] == 'August II of Braunschweig-Wolfenbüttel, 1579-1666']['collection'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we look at who these letters are addressed to:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Addressees of August II of Braunschweig-Wolfenbüttel's letters:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Andreae, Johann Valentin, 1586-1654                        581\n",
       "Kircher, Athanasius, 1601-1680                              21\n",
       "Württemberg, Eberhard III von, 1614-1674                     1\n",
       "Braunschweig-Lüneburg, Georg, 1582-1641                     1\n",
       "Braunschweig-Wolfenbüttel, Sophia Hedwig von, 1592-1642      1\n",
       "Name: addressee, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Addressees of August II of Braunschweig-Wolfenbüttel's letters:\")\n",
    "df[df['author'] == 'August II of Braunschweig-Wolfenbüttel, 1579-1666']['addressee'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These two queries reveal a typical pattern in these collections. August II has 582 letters in the collection of Johann Valentin Andreae, of which 581 are also addressed to Andreae. Letters in a collection around a certain person tend be either authored or addressed to this person, which makes sense from a recordkeeping perspective. But there is one letter addressed to someone else, i.e. Eberhard III von Württemberg.\n",
    "\n",
    "Now, let us look at the same queries for John Dury's letters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collections with John Dury's letters:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Hartlib, Samuel                837\n",
       "Bodleian card catalogue        149\n",
       "Ussher, James                    6\n",
       "Mede, Joseph                     3\n",
       "Huygens, Constantijn             2\n",
       "Culpeper, Cheney                 2\n",
       "Boyle, Robert                    2\n",
       "Vossius, Gerardus Joannes        1\n",
       "Bisterfeld, Johann Heinrich      1\n",
       "Name: collection, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Collections with John Dury's letters:\")\n",
    "df[df['author'] == 'Dury, John, 1596-1680']['collection'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "John Dury has letters in eight different collections, but in seven of those, it is only a handful of letters. We can also see who he addressed those letters to:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Addressees of John Dury's letters:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Hartlib, Samuel, 1600-1662                      528\n",
       "Roe, Thomas (Sir), 1581-1644                     31\n",
       "Culpeper, Cheney, 1601-1663                      11\n",
       "Borthwick, Eleazar, fl. 1633-1642                 7\n",
       "St Amand, Joseph, fl. 1636-1643                   7\n",
       "                                               ... \n",
       "House of Commons (1641-1712)                      1\n",
       "St Gallen and Appenzell, Clergy in, fl. 1654      1\n",
       "Cecil, Elizabeth, fl. 1640                        1\n",
       "Coysh, Joseph, fl. 1652                           1\n",
       "Rusdorf, Johann Joachim von, 1589-1640            1\n",
       "Name: addressee, Length: 135, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Addressees of John Dury's letters:\")\n",
    "df[df['author'] == 'Dury, John, 1596-1680']['addressee'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we see a differennt pattern. Samuel Hartlib is by far the most frequent addressee of John Dury's letters in these collections. But looking at the two sets of counts above, we note that John Dury authored 837 letters in the Samuel Hartlib collections, of which only 528 are addressed to Samuel Hartlib. Who are the other 309 letters in the Samuel Hartlib collection addressed to? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Addressees of John Dury's letters in the Samuel Hartlib:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Hartlib, Samuel, 1600-1662           528\n",
       "Roe, Thomas (Sir), 1581-1644          30\n",
       "Culpeper, Cheney, 1601-1663            9\n",
       "Waller, William (Sir), 1598-1668       7\n",
       "Borthwick, Eleazar, fl. 1633-1642      7\n",
       "                                    ... \n",
       "Ames, William, 1576-1633               1\n",
       "Ancelin, fl. 1660                      1\n",
       "Bedell, William, 1572-1642             1\n",
       "Figulus, Petr, 1619-1670               1\n",
       "Palmer, Herbert, 1601-1647             1\n",
       "Name: addressee, Length: 127, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Addressees of John Dury's letters in the Samuel Hartlib:\")\n",
    "df[(df['author'] == 'Dury, John, 1596-1680') & (df['collection'] == 'Hartlib, Samuel')]['addressee'].value_counts()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apparently, some collections also contains hundreds of letters that are not authored by or addressed to the collection eponym.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyzing the Addressees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Huygens, Constantijn, 1596-1687           4737\n",
       "Hearne, Thomas, 1678-1735                 3795\n",
       "Vossius, Gerardus Joannes, 1577-1649      3498\n",
       "Hartlib, Samuel, 1600-1662                3388\n",
       "Groot, Hugo de, 1583-1645                 3233\n",
       "Lhwyd, Edward, 1659-1709                  3226\n",
       "Charlett, Arthur (Reverend), 1655-1722    3066\n",
       "Andreae, Johann Valentin, 1586-1654       2953\n",
       "Noble, Mark (Reverend), 1754-1827         2809\n",
       "Sancroft, William, 1617-1693              2797\n",
       "Kircher, Athanasius, 1601-1680            2209\n",
       "Oldenburg, Henry, 1619-1677               2127\n",
       "D'Orville, Jacques Philippe, 1696-1751    2066\n",
       "Brett, Thomas, 1667-1744                  1767\n",
       "Lister, Martin, 1639-1712                 1704\n",
       "Vossius, Isaac (Dr), 1618-1689            1690\n",
       "Solms-Braunfels, Amalia von, 1602-1675    1660\n",
       "Smith, Thomas (Dr), 1638-1710             1637\n",
       "Wood, Anthony, 1632-1695                  1547\n",
       "Scaliger, Joseph Justus, 1540-1609        1512\n",
       "Name: addressee, dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['addressee'].value_counts().head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the list above, most of the addressees are the central figure or eponym of one of the EMLO collections.\n",
    "\n",
    "Exceptions are:\n",
    "\n",
    "- Nicolaas Reigersberch (1584-1654): brother-in-law of Hugo de Groot; Jurist\n",
    "- Willem de Groot (1597-1662): brother of Hugo de Groot (1583-1645); Dutch jurist\n",
    "\n",
    "These are prolific authors in collections centred on someone else. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collections with letters to Nicolaas Reigersberch:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de               881\n",
       "Vossius, Gerardus Joannes      8\n",
       "Bodleian card catalogue        6\n",
       "Name: collection, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Collections with letters to Nicolaas Reigersberch:\")\n",
    "df[df['addressee'] == 'Reigersberch, Nicolaas, 1584-1654']['collection'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Authors of letters to Nicolaas Reigersberch:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de, 1583-1645               862\n",
       "Reigersberch, Maria, 1589-1653           18\n",
       "Vossius, Gerardus Joannes, 1577-1649     14\n",
       "Groot, Willem de, 1597-1662               1\n",
       "Name: author, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Authors of letters to Nicolaas Reigersberch:\")\n",
    "df[df['addressee'] == 'Reigersberch, Nicolaas, 1584-1654']['author'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collections with letters to Willem de Groot:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de               732\n",
       "Vossius, Gerardus Joannes      3\n",
       "Bodleian card catalogue        1\n",
       "Name: collection, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Collections with letters to Willem de Groot:\")\n",
    "df[df['addressee'] == 'Groot, Willem de, 1597-1662']['collection'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Authors of letters to Willem de Groot:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Groot, Hugo de, 1583-1645                     726\n",
       "Vossius, Gerardus Joannes, 1577-1649            4\n",
       "Groot, Johan Hugo de, 1554-1640                 4\n",
       "Groot van Kraayenburg, Dirck de, 1618-1661      2\n",
       "Name: author, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Authors of letters to Willem de Groot:\")\n",
    "df[df['addressee'] == 'Groot, Willem de, 1597-1662']['author'].value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again, we see some letters between persons who are not the central figure in any of the EMLO collections. \n",
    "\n",
    "How many letters in each collection do not involve the eponym as either author or addressee?\n",
    "\n",
    "First, we map the name of the collection to the name as used as author or addressee:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Bodleian card catalogue\n"
     ]
    }
   ],
   "source": [
    "eponyms = list(df['collection'].unique())\n",
    "authors = list(df['author'].unique())\n",
    "author_counts = df['author'].value_counts()\n",
    "authors\n",
    "\n",
    "best_map = {}\n",
    "eponym_map = {}\n",
    "for eponym in eponyms:\n",
    "    #print(eponym)\n",
    "    for author in authors:\n",
    "        if not isinstance(author, str) or ';' in author:\n",
    "            continue\n",
    "        if eponym == 'Fermat, Pierre de' and author == 'Fermat, Pierre, 1601-1665':\n",
    "            eponym_map[eponym] = author\n",
    "        if eponym == 'Comenius, Jan Amos' and author == 'Komenský, Jan Amos, 1592-1670':\n",
    "            eponym_map[eponym] = author\n",
    "        if eponym in author[:len(eponym)]:\n",
    "            if eponym not in best_map or author_counts[author] > best_map[eponym]:\n",
    "                best_map[eponym] = author_counts[author]\n",
    "                eponym_map[eponym] = author\n",
    "    if eponym not in eponym_map:\n",
    "        print(eponym)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collection:\t\t\t\t\t\tAll letters\tNon-eponym letters\n",
      "----------------------------------------------------------------------------------------\n",
      "Bayle, Pierre                                     \t1791\t\t133\t(0.07)\n",
      "Sirleto, Guglielmo                                \t1438\t\t15\t(0.01)\n",
      "Seidenbecher, Georg Lorenz                        \t47\t\t0\t(0.00)\n",
      "Swammerdam, Jan                                   \t172\t\t4\t(0.02)\n",
      "Fermat, Pierre de                                 \t121\t\t6\t(0.05)\n",
      "Ortelius, Abraham                                 \t467\t\t0\t(0.00)\n",
      "Reneri, Henricus                                  \t61\t\t0\t(0.00)\n",
      "Spinoza, Baruch                                   \t58\t\t1\t(0.02)\n",
      "Lister, Martin                                    \t1212\t\t2\t(0.00)\n",
      "Wallis, John                                      \t1998\t\t232\t(0.12)\n",
      "Ussher, James                                     \t681\t\t17\t(0.02)\n",
      "Groot, Hugo de                                    \t8034\t\t280\t(0.03)\n",
      "Franckenberg, Abraham von                         \t85\t\t1\t(0.01)\n",
      "Bourignon, Antoinette                             \t940\t\t0\t(0.00)\n",
      "Peiresc, Nicolas-Claude Fabri de                  \t1939\t\t11\t(0.01)\n",
      "Hartlib, Samuel                                   \t4719\t\t1221\t(0.26)\n",
      "Braunschweig-Wolfenbüttel, Sophia Hedwig von      \t169\t\t1\t(0.01)\n",
      "Solms-Braunfels, Amalia von                       \t1184\t\t3\t(0.00)\n",
      "Comenius, Jan Amos                                \t571\t\t41\t(0.07)\n",
      "Selden, John                                      \t355\t\t10\t(0.03)\n",
      "Culpeper, Cheney                                  \t3\t\t0\t(0.00)\n",
      "Montagu, Mary Wortley                             \t963\t\t0\t(0.00)\n",
      "Nierop, Dirck Rembrantsz van                      \t80\t\t4\t(0.05)\n",
      "Plantin, Christophe                               \t3030\t\t324\t(0.11)\n",
      "Oldenburg, Henry                                  \t3176\t\t34\t(0.01)\n",
      "Pontanus, Johannes Isacius                        \t321\t\t2\t(0.01)\n",
      "Dodington, John                                   \t571\t\t8\t(0.01)\n",
      "Stuart, Arbella                                   \t118\t\t2\t(0.02)\n",
      "Vives, Juan Luis                                  \t195\t\t19\t(0.10)\n",
      "Jungius, Joachim                                  \t506\t\t18\t(0.04)\n",
      "Coccejus, Johannes                                \t515\t\t2\t(0.00)\n",
      "Opitz, Martin                                     \t110\t\t0\t(0.00)\n",
      "Agustín, Antonio                                  \t579\t\t5\t(0.01)\n",
      "Andreae, Johann Valentin                          \t3696\t\t81\t(0.02)\n",
      "Anhalt-Dessau, Henriette Amalia von               \t1352\t\t1\t(0.00)\n",
      "Thomson, Richard                                  \t78\t\t2\t(0.03)\n",
      "Schott, Caspar                                    \t180\t\t0\t(0.00)\n",
      "Permeier, Johann                                  \t89\t\t3\t(0.03)\n",
      "Vossius, Isaac                                    \t1703\t\t2\t(0.00)\n",
      "Pascal, Blaise                                    \t49\t\t6\t(0.12)\n",
      "Magini, Giovanni Antonio                          \t100\t\t1\t(0.01)\n",
      "Vossius, Gerardus Joannes                         \t3430\t\t14\t(0.00)\n",
      "Bernegger, Matthias                               \t435\t\t0\t(0.00)\n",
      "Scaliger, Joseph Justus                           \t3338\t\t8\t(0.00)\n",
      "Mengoli, Pietro                                   \t40\t\t0\t(0.00)\n",
      "Plot, Robert                                      \t108\t\t6\t(0.06)\n",
      "Hilchen, David                                    \t98\t\t22\t(0.22)\n",
      "Sidney, Philip                                    \t380\t\t5\t(0.01)\n",
      "Boyle, Robert                                     \t1759\t\t9\t(0.01)\n",
      "Jurin, James                                      \t701\t\t0\t(0.00)\n",
      "Sachs von Löwenheim, Philipp Jakob                \t143\t\t0\t(0.00)\n",
      "Kepler, Johannes                                  \t883\t\t57\t(0.06)\n",
      "Beverland, Hadriaan                               \t305\t\t0\t(0.00)\n",
      "Reland, Adriaan                                   \t211\t\t0\t(0.00)\n",
      "Bisterfeld, Johann Heinrich                       \t121\t\t3\t(0.02)\n",
      "Baxter, Richard                                   \t8\t\t0\t(0.00)\n",
      "Lhwyd, Edward                                     \t2128\t\t74\t(0.03)\n",
      "Euler, Leonhard                                   \t811\t\t35\t(0.04)\n",
      "Halley, Edmond                                    \t245\t\t0\t(0.00)\n",
      "Aubrey, John                                      \t1073\t\t10\t(0.01)\n",
      "Conway, Anne                                      \t296\t\t78\t(0.26)\n",
      "Mersenne, Marin                                   \t1904\t\t784\t(0.41)\n",
      "Milton, John                                      \t66\t\t0\t(0.00)\n",
      "Rabus, Pieter                                     \t30\t\t0\t(0.00)\n",
      "Mede, Joseph                                      \t441\t\t9\t(0.02)\n",
      "Pennant, Thomas                                   \t508\t\t1\t(0.00)\n",
      "Hobbes, Thomas                                    \t223\t\t10\t(0.04)\n",
      "Beale, Robert                                     \t101\t\t27\t(0.27)\n",
      "Huygens, Christiaan                               \t3080\t\t393\t(0.13)\n",
      "Claude, Jean                                      \t117\t\t4\t(0.03)\n",
      "Ruar, Martin                                      \t100\t\t17\t(0.17)\n",
      "Gray, Thomas                                      \t651\t\t7\t(0.01)\n",
      "Oranje-Nassau, Albertine Agnes van                \t782\t\t48\t(0.06)\n",
      "Schurman, Anna Maria van                          \t244\t\t1\t(0.00)\n",
      "Rich, Penelope                                    \t42\t\t2\t(0.05)\n",
      "Newton, Isaac                                     \t1140\t\t135\t(0.12)\n",
      "Clifford, Margaret                                \t131\t\t0\t(0.00)\n",
      "Kircher, Athanasius                               \t2693\t\t3\t(0.00)\n",
      "Dudley, Anne                                      \t27\t\t1\t(0.04)\n",
      "Ashmole, Elias                                    \t764\t\t27\t(0.04)\n",
      "Vernon, Francis                                   \t275\t\t0\t(0.00)\n",
      "Collins, John                                     \t273\t\t88\t(0.32)\n",
      "Descartes, René                                   \t727\t\t7\t(0.01)\n",
      "Huygens, Constantijn                              \t7120\t\t6\t(0.00)\n",
      "Bacon, Anne                                       \t197\t\t2\t(0.01)\n",
      "Gruter, Jan                                       \t136\t\t0\t(0.00)\n",
      "Worthington, John                                 \t174\t\t8\t(0.05)\n",
      "Brahe, Tycho                                      \t505\t\t28\t(0.06)\n",
      "Rubens, Peter Paul                                \t940\t\t558\t(0.59)\n",
      "Hutton, Charles                                   \t133\t\t6\t(0.05)\n",
      "Vernon, Margaret                                  \t21\t\t0\t(0.00)\n",
      "Beeckman, Isaac                                   \t28\t\t1\t(0.04)\n"
     ]
    }
   ],
   "source": [
    "print(\"Collection:\\t\\t\\t\\t\\t\\tAll letters\\tNon-eponym letters\")\n",
    "print(\"----------------------------------------------------------------------------------------\")\n",
    "for eponym in eponym_map:\n",
    "    epo_df = df[df['collection'] == eponym]\n",
    "    #print(eponym, '\\t', eponym_map[eponym])\n",
    "    non_epo_df = df[(df['collection'] == eponym) & (df['author'] != eponym_map[eponym]) & (df['addressee'] != eponym_map[eponym])]\n",
    "    perc = non_epo_df.shape[0] / epo_df.shape[0]\n",
    "    print(f\"{eponym: <50}\\t{epo_df.shape[0]}\\t\\t{non_epo_df.shape[0]}\\t({perc:.2f})\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Most collection have almost exclusively letters involving the eponym, but some collections are very different. In the Peter Paul Rubens collection, the majority (59%) of letters are between other people than Rubens. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>collection</th>\n",
       "      <th>author</th>\n",
       "      <th>addressee</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>131088</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Moretus, Balthasar, 1574-1641</td>\n",
       "      <td>Rubens, Philip, 1574-1611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131089</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Rubens, Philip, 1574-1611</td>\n",
       "      <td>Rubens, Peter Paul, 1577-1640</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131090</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Albert VII, Archduke of Austria, 1559-1621</td>\n",
       "      <td>Richardot, Jean, 1570-1614</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131091</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Gonzaga, Vincenzo I, 1562-1612</td>\n",
       "      <td>Damasceni Peretti, Alessandro, 1571-1623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131092</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Damasceni Peretti, Alessandro, 1571-1623</td>\n",
       "      <td>Gonzaga, Vincenzo I, 1562-1612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131093</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Arrigoni, Lelio, b.1541</td>\n",
       "      <td>Chieppio, Annibal, 1563-1623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131094</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Rubens, Philip, 1574-1611</td>\n",
       "      <td>Rubens, Peter Paul, 1577-1640</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131095</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Arrigoni, Lelio, b.1541</td>\n",
       "      <td>Chieppio, Annibal, 1563-1623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131096</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Richardot, Jean, 1570-1614</td>\n",
       "      <td>Gonzaga, Vincenzo I, 1562-1612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131097</th>\n",
       "      <td>Rubens, Peter Paul</td>\n",
       "      <td>Arrigoni, Lelio, b.1541</td>\n",
       "      <td>Chieppio, Annibal, 1563-1623</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                collection                                      author  \\\n",
       "131088  Rubens, Peter Paul               Moretus, Balthasar, 1574-1641   \n",
       "131089  Rubens, Peter Paul                   Rubens, Philip, 1574-1611   \n",
       "131090  Rubens, Peter Paul  Albert VII, Archduke of Austria, 1559-1621   \n",
       "131091  Rubens, Peter Paul              Gonzaga, Vincenzo I, 1562-1612   \n",
       "131092  Rubens, Peter Paul    Damasceni Peretti, Alessandro, 1571-1623   \n",
       "131093  Rubens, Peter Paul                     Arrigoni, Lelio, b.1541   \n",
       "131094  Rubens, Peter Paul                   Rubens, Philip, 1574-1611   \n",
       "131095  Rubens, Peter Paul                     Arrigoni, Lelio, b.1541   \n",
       "131096  Rubens, Peter Paul                  Richardot, Jean, 1570-1614   \n",
       "131097  Rubens, Peter Paul                     Arrigoni, Lelio, b.1541   \n",
       "\n",
       "                                       addressee  \n",
       "131088                 Rubens, Philip, 1574-1611  \n",
       "131089             Rubens, Peter Paul, 1577-1640  \n",
       "131090                Richardot, Jean, 1570-1614  \n",
       "131091  Damasceni Peretti, Alessandro, 1571-1623  \n",
       "131092            Gonzaga, Vincenzo I, 1562-1612  \n",
       "131093              Chieppio, Annibal, 1563-1623  \n",
       "131094             Rubens, Peter Paul, 1577-1640  \n",
       "131095              Chieppio, Annibal, 1563-1623  \n",
       "131096            Gonzaga, Vincenzo I, 1562-1612  \n",
       "131097              Chieppio, Annibal, 1563-1623  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df['collection'] == 'Rubens, Peter Paul'][['collection','author','addressee']].head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Normalization and Classification for Creating and Comparing Groups\n",
    "\n",
    "The metadata is fairly minimal when considering just the fields that are in the dataset. But there are more things that can be done.\n",
    "\n",
    "- The names of senders and recipients have the birth and death years (in most cases), so we could use these to group persons by age at death, or birth decade. \n",
    "\n",
    "- The dates that the letters were sent are often exact down to the specific day, but sometimes only a month was known or an earliest and latest probable dates. We can normalise those dates to get an insight in when letters were sent, in which year or month. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalizing and scale\n",
    "\n",
    "At a small scale, there is no need to normalize data, as the researcher can do that mentally while working with the materials.\n",
    "\n",
    "At an intermediate scale of hundreds or thousands of documents, the variations in names of persons and places, ways in which dates are recorded are becoming a hurdle to analysis. For topical analysis, this is also an issue, as many connections between documents are hard to bring to the surface because of morphological and spelling variations. \n",
    "\n",
    "At a large scale with hundreds of thousands or millions of documents, the textual variations become less of a hurdle, as there is enough data to identify and map variants. \n",
    "\n",
    "At a very large scale with tens or hundreds of millions of documents, the textual variations become meaningful and allow measuring contextual nuance in how word variants are used to convey different aspects."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9 January 1638      14\n",
       "Unknown date        14\n",
       "12 December 1643    12\n",
       "20 June 1643        11\n",
       "6 February 1644     11\n",
       "                    ..\n",
       "4 April 1634         1\n",
       "3 November 1628      1\n",
       "9 June 1628          1\n",
       "20 August 1627       1\n",
       "11 January 1609      1\n",
       "Name: date, Length: 4067, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.collection == 'Groot, Hugo de'].date.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The are 4067 different values for the dates, with the most common date being 9 January 1638. There are also 14 unknown dates. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "def is_day_month_year(sent_date):\n",
    "    return re.match(r'^\\d+ (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\\w* \\d{4}$', sent_date) != None\n",
    "\n",
    "def is_month_year(sent_date):\n",
    "    return re.match(r'^(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\\w* \\d{4}$', sent_date) != None\n",
    "\n",
    "def is_year(sent_date):\n",
    "    return re.match(r'^\\d{4}$', sent_date) != None\n",
    "\n",
    "def is_day_month(sent_date):\n",
    "    return re.match(r'^\\d+ (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\\w*$', sent_date) != None\n",
    "\n",
    "def is_month(sent_date):\n",
    "    return re.match(r'^(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\\w*$', sent_date) != None\n",
    "\n",
    "def is_day_year(sent_date):\n",
    "    return re.match(r'^\\d+ \\d{4}$', sent_date) != None\n",
    "\n",
    "def get_year(sent_date):\n",
    "    if is_day_month_year(sent_date) or is_year(sent_date) or is_day_year(sent_date) or is_month_year(sent_date):\n",
    "        return int(sent_date[-4:])\n",
    "    else:\n",
    "        return None\n",
    "    \n",
    "def get_month(sent_date):\n",
    "    if is_day_month_year(sent_date) or is_month(sent_date) or is_day_month(sent_date) or is_month_year(sent_date):\n",
    "        match = re.match(r'.*((Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\\w*).*', sent_date)\n",
    "        return match.group(1)\n",
    "    else:\n",
    "        return None\n",
    "    \n",
    "    \n",
    "def get_date_type(sent_date):\n",
    "    if is_day_month_year(sent_date):\n",
    "        return 'day_month_year'\n",
    "    if is_month_year(sent_date):\n",
    "        return 'month_year'\n",
    "    if is_year(sent_date):\n",
    "        return 'year'\n",
    "    if is_day_month(sent_date):\n",
    "        return 'day_month'\n",
    "    if is_month(sent_date):\n",
    "        return 'day_month'\n",
    "    if is_day_year(sent_date):\n",
    "        return 'day_year'\n",
    "    if 'Between' in sent_date:\n",
    "        return 'range_between'\n",
    "    if 'On or before' in sent_date:\n",
    "        return 'range_before'\n",
    "    if 'On or after' in sent_date:\n",
    "        return 'range_after'\n",
    "    if 'Unknown date' in sent_date:\n",
    "        return 'unknown'\n",
    "    else:\n",
    "        return 'invalid format'\n",
    "\n",
    "#df['date_type'] = df.date.apply(get_date_type)\n",
    "df['date_year'] = df.date.apply(get_year)\n",
    "df['date_month'] = df.date.apply(get_month)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "day_month_year    114636\n",
       "unknown             5748\n",
       "month_year          3778\n",
       "year                3103\n",
       "day_month           3053\n",
       "range_between       1700\n",
       "range_before         104\n",
       "range_after           44\n",
       "day_year              37\n",
       "invalid format         7\n",
       "Name: date_type, dtype: int64"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.date_type.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "322.0"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.date_year.max() - df.date_year.min() + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.date_year.hist(bins=322)#.value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "March        11031\n",
       "August       10507\n",
       "July         10364\n",
       "May          10335\n",
       "April        10239\n",
       "January      10139\n",
       "September     9957\n",
       "June          9951\n",
       "October       9884\n",
       "February      9855\n",
       "November      9640\n",
       "December      9565\n",
       "Name: date_month, dtype: int64"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.date_month.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "author                                       addressee                                      \n",
       "Albert VII, Archduke of Austria, 1559-1621   Gonzaga, Vincenzo I, 1562-1612                      1\n",
       "Rubens, Peter Paul, 1577-1640                Vosbergen, Josias van, 1593-1628                    1\n",
       "Mello, Francisco Manuel de, 1608-1666        Chambre des Comptes (Spanish Netherlands)           1\n",
       "Mennes, John (Sir), 1599-1671                Admiralty Court, England                            1\n",
       "Moncada, Francisco, 1586-1635                Philip IV, King of Spain, 1605-1665                 1\n",
       "                                                                                                ..\n",
       "Rubens, Peter Paul, 1577-1640                Fabri, Palamède, 1582-1645                         18\n",
       "                                             Olivares, Gaspar de Guzmán, Conde de, 1587-1645    25\n",
       "Ferdinand, archiduc d'Autriche, 1609-1641    Philip IV, King of Spain, 1605-1665                36\n",
       "Rubens, Peter Paul, 1577-1640                Dupuy, Pierre, 1582-1651                           71\n",
       "Peiresc, Nicolas-Claude Fabri de, 1580-1637  Rubens, Peter Paul, 1577-1640                      85\n",
       "Length: 301, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_rubens = df[df.collection == 'Rubens, Peter Paul']\n",
    "\n",
    "df_rubens[['collection','author','addressee']].head(10)\n",
    "\n",
    "g = df_rubens.groupby(['author', 'addressee']).size()\n",
    "\n",
    "u = g.unstack('author')\n",
    "\n",
    "plt.imshow(u, cmap='hot', interpolation='nearest')\n",
    "\n",
    "g.sort_values()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Connections between collections\n",
    "\n",
    "How many connections are there between collections? This is easy with two collections, but becomes more difficult when there are many collections.\n",
    "\n",
    "Which persons appear in multiple collections?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Unknown                                                                                                                                                     37\n",
       "Oldenburg, Henry, 1619-1677                                                                                                                                 16\n",
       "Huygens, Constantijn, 1596-1687                                                                                                                             14\n",
       "Mersenne, Marin, 1588-1648                                                                                                                                  13\n",
       "Gronovius, Johann Frederick, 1611-1671                                                                                                                      12\n",
       "Groot, Hugo de, 1583-1645                                                                                                                                   11\n",
       "Leibniz, Gottfried Wilhelm, 1646-1716                                                                                                                       11\n",
       "Vossius, Gerardus Joannes, 1577-1649                                                                                                                        11\n",
       "Descartes, René, 1596-1650                                                                                                                                  10\n",
       "Huygens, Christiaan, 1629-1695                                                                                                                              10\n",
       "Rivet, André, 1572-1651                                                                                                                                     10\n",
       "Saumaise, Claude de, 1588-1653                                                                                                                              10\n",
       "Hevelius, Johannes, 1611-1687                                                                                                                                9\n",
       "Boyle, Robert, 1627-1691                                                                                                                                     9\n",
       "Hartlib, Samuel, 1600-1662                                                                                                                                   9\n",
       "Aubrey, John, 1626-1697                                                                                                                                      9\n",
       "Digby, Kenelm (Sir), 1603-1665                                                                                                                               9\n",
       "Gassendi, Pierre, 1592-1655                                                                                                                                  9\n",
       "Wallis, John (Dr), 1616-1703                                                                                                                                 9\n",
       "Dury, John, 1596-1680                                                                                                                                        8\n",
       "Lipsius, Justus, 1547-1606                                                                                                                                   8\n",
       "Komenský, Jan Amos, 1592-1670                                                                                                                                8\n",
       "Bernegger, Matthias, 1582-1640                                                                                                                               8\n",
       "Newton, Isaac (Sir), 1642-1727                                                                                                                               8\n",
       "Peiresc, Nicolas-Claude Fabri de, 1580-1637                                                                                                                  8\n",
       "Kircher, Athanasius, 1601-1680                                                                                                                               8\n",
       "Boulliau, Ismaël, 1605-1694                                                                                                                                  8\n",
       "Heinsius, Daniel, 1580-1655                                                                                                                                  8\n",
       "Sorbière, Samuel, 1615-1670                                                                                                                                 8\n",
       "Christina, Queen of Sweden, 1626-1689                                                                                                                        7\n",
       "                                                                                                                                                            ..\n",
       "Manwaring, Robert, fl. 1637                                                                                                                                  1\n",
       "Smith, John, 1630-1679                                                                                                                                       1\n",
       "Cort, Christiaan de, 1611-1669                                                                                                                               1\n",
       "Aerssen, Cornelis van, 1600-1662                                                                                                                             1\n",
       "Cottereau, N., 1641-1706                                                                                                                                     1\n",
       "Conway, Anne, 1631-1679                                                                                                                                      1\n",
       "Chieppio, Annibal, 1563-1623                                                                                                                                 1\n",
       "Honywood, Robert (Sir), 1601-1686                                                                                                                            1\n",
       "Doublet, Philips, d.1647                                                                                                                                     1\n",
       "Beeck, Anna, fl. 1649                                                                                                                                        1\n",
       "Montagu, Edward Wortley, 1678-1761                                                                                                                           1\n",
       "Martinengo da Barco, Ascanio, 1539-1600                                                                                                                      1\n",
       "Ghilde, Johan Flud van, fl. 1685                                                                                                                             1\n",
       "Thomas, Robert, b.1681                                                                                                                                       1\n",
       "Brassard, Marie, fl. 1685                                                                                                                                    1\n",
       "Dalrymple, James, 1619-1695                                                                                                                                  1\n",
       "Leslie, John (Sir), 1766-1832                                                                                                                                1\n",
       "Bacon, Arthur, fl. 1652                                                                                                                                      1\n",
       "Standfast, William (Reverend), 1683-1754                                                                                                                     1\n",
       "Hoorn, magistrate of, fl. 1616-1617                                                                                                                          1\n",
       "Clenche, Andrew, d.1692                                                                                                                                      1\n",
       "Bachcroft, Thomas, 1571-1662; Bainbridge, Thomas, 1574-1646; Brownrigg, Ralph (Dr), 1592-1659; Collins, Samuel (Dr), 1576-1651; Love, Richard, 1596-1661     1\n",
       "Higgins, Obadiah, 1663-1741                                                                                                                                  1\n",
       "Techmannus, Arnoldus, 1594-1666                                                                                                                              1\n",
       "Klage, Thomas, 1598-1664                                                                                                                                     1\n",
       "Bodecher, J. W., fl. 1643                                                                                                                                    1\n",
       "Bachacius, Martinus, 1539-1612                                                                                                                               1\n",
       "Morgan, Anthony, fl. 1654                                                                                                                                    1\n",
       "Finch, John, fl. 1653                                                                                                                                        1\n",
       "Ellenmeier, Johann, fl. 1635                                                                                                                                 1\n",
       "Name: author, dtype: int64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# which authors occur in multiple collections\n",
    "df[(df[['collection', 'author']].duplicated(keep='first') == False)]['author'].value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Oldenburg, Henry                      1524\n",
       "Wallis, John                           142\n",
       "Boyle, Robert                          107\n",
       "Huygens, Christiaan                    105\n",
       "Lister, Martin                          61\n",
       "Hartlib, Samuel                         46\n",
       "Newton, Isaac                           16\n",
       "Swammerdam, Jan                          8\n",
       "Milton, John                             7\n",
       "Sachs von Löwenheim, Philipp Jakob       5\n",
       "Vossius, Isaac                           4\n",
       "Hobbes, Thomas                           2\n",
       "Ashmole, Elias                           1\n",
       "Comenius, Jan Amos                       1\n",
       "Vossius, Gerardus Joannes                1\n",
       "Coccejus, Johannes                       1\n",
       "Name: collection, dtype: int64"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df['author'] == 'Oldenburg, Henry, 1619-1677']['collection'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Samuel Hartlib is in the top 20 of addressees but not in the top 20 of authors:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Samuel Hartlib\n",
      "\n",
      "\tnumber of letters sent: 401\n",
      "\tnumber of letters received: 3388\n"
     ]
    }
   ],
   "source": [
    "print('Samuel Hartlib\\n')\n",
    "print(f'\\tnumber of letters sent:', df[df['author'] == 'Hartlib, Samuel, 1600-1662'].shape[0])\n",
    "print(f'\\tnumber of letters received:', (df[df['addressee'] == 'Hartlib, Samuel, 1600-1662'].shape[0]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noFdKqRCngV4ppUKcBnqllApx/x96wyWBRjydrQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x145ad7160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Number of letters authored by Hugo de Groot per year\n",
    "hugo = 'Groot, Hugo de, 1583-1645'\n",
    "\n",
    "df['year'] = df['date'].str.extract('(\\d\\d\\d\\d)', expand=False)\n",
    "\n",
    "df_hugo = df[df['author'] == hugo]\n",
    "\n",
    "df_hugo['year'].value_counts().sort_index().plot()\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Reigersberch, Nicolaas, 1584-1654                                862\n",
       "Groot, Willem de, 1597-1662                                      726\n",
       "Oxenstierna, Axel (Count), 1583-1654                             587\n",
       "Camerarius, Ludwig, 1573-1651                                    347\n",
       "Vossius, Gerardus Joannes, 1577-1649                             339\n",
       "Marin, Charles, d.1651                                           122\n",
       "Oxenstierna, Johan Axelsson, 1611-1657                           109\n",
       "Salvius, Johan Adler, 1590-1652                                  106\n",
       "Wicquefort, Joachim van, 1600-1670                                99\n",
       "Heinsius, Daniel, 1580-1655                                       89\n",
       "Aubery du Maurier, Benjamin, 1566-1636                            85\n",
       "Appelboom, Harald Andersson, 1612-1674                            83\n",
       "Christina, Queen of Sweden, 1626-1689                             77\n",
       "Schmalz, Peter Abel, fl. 1635-1638                                61\n",
       "Uyttenbogaert, Johannes (Dr), 1557-1644                           51\n",
       "Jaski, Israel, 1573-1642                                          39\n",
       "Bielke, Sten Svantesson, 1598-1638                                39\n",
       "Spiring Silvercrona, Petter, 1600-1652                            39\n",
       "Groot, Johan Hugo de, 1554-1640                                   37\n",
       "Lingelsheim, Georg Michael, 1556-1636                             36\n",
       "Unknown                                                           33\n",
       "Bernegger, Matthias, 1582-1640                                    29\n",
       "Müller, Georg, d.1639                                             29\n",
       "Sprecher von Bernegg, Fortunatus, 1585-1647                       24\n",
       "Casaubon, Isaac, 1559-1614                                        23\n",
       "Grubbe, Lars, 1601-1642                                           22\n",
       "Camerarius, Joachim, 1603-1687                                    22\n",
       "Meursius, Johannes, 1579-1639                                     22\n",
       "Vossius, Isaac (Dr), 1618-1689                                    20\n",
       "Otto II, 1578-1637                                                19\n",
       "                                                                ... \n",
       "Skytte, Bengt, 1614-1683                                           1\n",
       "Gomaer, François, 1563-1641                                        1\n",
       "Barclay-Debonnaire, Louise, 1585-1652                              1\n",
       "Emporagrius, Erik Gabrielsson, 1606-1674                           1\n",
       "Cappel, Louis, 1585-1658                                           1\n",
       "Jack, Gilbert, 1577-1628                                           1\n",
       "Vossius, Gerardus, 1619-1640                                       1\n",
       "Aligre, Étienne, 1550-1635                                         1\n",
       "Höpfner, Heinrich, 1583-1642                                       1\n",
       "Aa, Anthony Willemsz., 1582-1638                                   1\n",
       "Rigault, Nicolas, 1577-1654                                        1\n",
       "Brederode, Reinoud, 1567-1633                                      1\n",
       "Jungermann, Gottfried, 1577-1610                                   1\n",
       "Wertheim de Rochefort, Johann Dietrich Löwenstein, 1585-1657       1\n",
       "L'Empereur, Constantine, 1591-1648                                 1\n",
       "Stringe, Johan, fl. 1622                                           1\n",
       "Forstner, Christoph von, 1598-1667                                 1\n",
       "Voisin, Joseph, 1610-1685                                          1\n",
       "Orléans, administration of the German Nation at                    1\n",
       "Sweerts, Pierre François, 1567-1629                               1\n",
       "Pas, Isaac Manassès de, 1590-1640                                  1\n",
       "Mesmes, Henri, d.1650                                              1\n",
       "Gardie, Magnus Gabriel de la, 1622-1686                            1\n",
       "Wertheim de Virneburg, Friedrich Ludwig Löwenstein, 1598-1657      1\n",
       "Bogislaw, Ernst, 1620-1684                                         1\n",
       "Hohenlohe-Langenburg, Philipp Ernst von, 1584-1628                 1\n",
       "Oldenbarnevelt, Willem van, 1590-1638                              1\n",
       "Beauharnais, François, d.1651                                      1\n",
       "Vair, Guillaume, 1566-1621                                         1\n",
       "Menasseh, Ben Israel, 1604-1657                                    1\n",
       "Name: addressee, dtype: int64"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_hugo['addressee'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some collections include letters between correspondents of the collection creator, while others only contains letters where the collection creator is the author or addressee of the letter. \n",
    "\n",
    "E.g. the collection of correspondence of Hugo de Groot includes letters between his brother and his brother-in-law.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Huygens, Christiaan, 1629-1695                         1345\n",
       "Huygens, Constantijn, 1628-1697                         175\n",
       "Oldenburg, Henry, 1619-1677                             105\n",
       "Huygens, Constantijn, 1596-1687                          75\n",
       "Sluse, René François de, 1622-1685                       72\n",
       "Chapelain, Jean, 1595-1674                               70\n",
       "Moray, Robert (Sir), 1608-1673                           68\n",
       "Schooten, Frans van, 1615-1660                           58\n",
       "Boulliau, Ismaël, 1605-1694                              54\n",
       "Leibniz, Gottfried Wilhelm, 1646-1716                    42\n",
       "Bruno, Henrick, 1617-1664                                40\n",
       "Huygens, Susanna, 1637-1725                              32\n",
       "Heinsius, Nicolaas, 1620-1681                            28\n",
       "Petit, Pierre, 1598 or before-1677                       24\n",
       "Medici, Leopoldo de', 1617-1675                          24\n",
       "Doublet, Philips, 1633-1707                              23\n",
       "Hudde, Johannes, 1628-1704                               21\n",
       "L'Hôpital, Guillaume François Antoine de, 1661-1704      21\n",
       "Wallis, John (Dr), 1616-1703                             20\n",
       "Mersenne, Marin, 1588-1648                               20\n",
       "Fatio de Duillier, Nicolas, 1664-1753                    20\n",
       "Saint-Vincent, Grégoire de, 1584-1667                    19\n",
       "Kinner von Löwenthurn, Gottfried Alois, b.1610           18\n",
       "Mylon, Claude, 1618-1660                                 18\n",
       "Fermat, Pierre, 1601-1665                                17\n",
       "Hire, Philippe de la, 1640-1718                          16\n",
       "Gent, Pieter, b.1640                                     16\n",
       "Graaf, Jan, 1673-1697                                    16\n",
       "Hevelius, Johannes, 1611-1687                            15\n",
       "Thévenot, Melchisédech, 1620-1692                        13\n",
       "                                                       ... \n",
       "Hobbes, Thomas, 1588-1679                                 1\n",
       "Mathion, Oded Louis, 1620-1700                            1\n",
       "Louise Hollandine, Countess Palatine, 1622-1709           1\n",
       "Regnauld, André, d.1702                                   1\n",
       "Benoit, Antoine, 1632-1717                                1\n",
       "Limojon, Alexandre-Toussaint de, 1630-1689                1\n",
       "Leeuwenhoek, Antoni van, 1632-1723                        1\n",
       "Gillet, Pierre François, 1648-1720                        1\n",
       "Nassau-Siegen, Hendrik of, 1611-1652                      1\n",
       "Boecler, Johann Heinrich, 1611-1672                       1\n",
       "Lely, Peter, 1618-1680                                    1\n",
       "Christina, Queen of Sweden, 1626-1689                     1\n",
       "Kircher, Athanasius, 1601-1680                            1\n",
       "Vossius, Isaac (Dr), 1618-1689                            1\n",
       "Molyneux, Thomas (Sir), 1661-1733                         1\n",
       "Briou, fl. 1675                                           1\n",
       "Douw, Simon, 1620-1663                                    1\n",
       "Court of Holland, Zeeland and West Friesland,             1\n",
       "Alberghetti, Sigismondo, d.1702                           1\n",
       "Placentius, Johann, d.1683                                1\n",
       "Wijk, Johan van der, 1625-1679                            1\n",
       "Magalotti, Lorenzo, 1637-1712                             1\n",
       "Varignon, Pierre, 1654-1722                               1\n",
       "Cock, Christopher, fl. 1684                               1\n",
       "Dodart, Denis, 1634-1707                                  1\n",
       "Holmes, Robert (Sir), 1622-1692                           1\n",
       "Vallot, Antoine, 1594-1671                                1\n",
       "Bilberg, Johann, 1650-1717                                1\n",
       "Loménie, Henri-Auguste, 1595-1666                         1\n",
       "Smethwick, Francis, d.1682                                1\n",
       "Name: author, dtype: int64"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_christiaan = df[(df['author'] == 'Huygens, Christiaan, 1629-1695') | (df['addressee'] == 'Huygens, Christiaan, 1629-1695')]\n",
    "\n",
    "df_christiaan = df[df['collection'] == 'Huygens, Christiaan']\n",
    "df_christiaan['author'].value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Huygens, Constantijn, 1596-1687                                          4252\n",
       "Solms-Braunfels, Amalia von, 1602-1675                                    768\n",
       "Huygens, Christiaan, 1551-1624                                             94\n",
       "Barlaeus, Caspar, 1584-1648                                                81\n",
       "Sauzin, Jean                                                               68\n",
       "Heinsius, Daniel, 1580-1655                                                63\n",
       "Lionne, Hugues de, 1611-1671                                               44\n",
       "Rivet, André, 1572-1651                                                    44\n",
       "Hooft, Pieter Cornelius, 1581-1647                                         40\n",
       "William III and II, King of England, Scotland, and Ireland, 1650-1702      39\n",
       "Beringhen, Henri, 1603-1692                                                33\n",
       "Chièze, Sebastien, 1625-1679                                               29\n",
       "Unknown                                                                    28\n",
       "Langes de Montmirail, Frédéric de, 1630-1697                               26\n",
       "Leu de Wilhem, David le, 1588-1658                                         25\n",
       "Chambrun, Jacques Pineton, 1635-1689                                       23\n",
       "Dohna, Frederick von, 1621-1688                                            21\n",
       "Council of the Prince, fl. 1656-1664                                       20\n",
       "Loménie, Henri Louis, 1635-1698                                            20\n",
       "Swann-Ogle, Utricia, 1611-1674                                             19\n",
       "Ban, Jan Albert, 1598-1644                                                 19\n",
       "Jermyn, Henry, 1605-1684                                                   18\n",
       "Puteanus, Erycius, 1574-1646                                               18\n",
       "Nassau-Siegen, Hendrik of, 1611-1652                                       18\n",
       "Westerbaen, Jacob, 1599-1670                                               17\n",
       "Cusance, Béatrix, 1614-1663                                               17\n",
       "Schurman, Anna Maria van, 1607-1678                                        17\n",
       "Boxhorn, Marcus Zuerius van, 1612-1653                                     15\n",
       "Bennet, Henry, 1618-1685                                                   15\n",
       "Le Tellier, François Michel, 1641-1691                                     14\n",
       "                                                                         ... \n",
       "Wierts, Joan, fl. 1650-1692                                                 1\n",
       "Muelen, Andries, 1591-1654                                                  1\n",
       "Molino, Domenico, 1573-1635                                                 1\n",
       "Burg, fl. 1662                                                              1\n",
       "Frédéric-Armand, comte de Schomberg, 1615-1690                              1\n",
       "Fürstenberg, Ferdinand von, 1626-1683                                       1\n",
       "Colvius, Andreas, 1594-1671                                                 1\n",
       "Sprecher von Bernegg, Fortunatus, 1585-1647                                 1\n",
       "Sipenesse, Cornelis, d.1635                                                 1\n",
       "Santen, Jan, fl. 1635-1649                                                  1\n",
       "Caron, Suzette, fl. 1669-1689                                               1\n",
       "Amat, Angélique, fl. 1666                                                   1\n",
       "Brancas, Marie, fl. 1613-1662                                               1\n",
       "Nicholas, Edward (Sir), 1593-1669                                           1\n",
       "Beauvais, Charles, b.1590                                                   1\n",
       "Sohier, Nicolaas, fl. 1638                                                  1\n",
       "Zuerius (Miss), fl. 1667                                                    1\n",
       "Sylvius, Jean, fl. 1662-1666                                                1\n",
       "Does, Jacob, 1641-1680                                                      1\n",
       "Huygen, Johan, fl. 1632-1640                                                1\n",
       "Coesmans, Jan, fl. 1643-1656                                                1\n",
       "Brederode, Juliana, 1622-1678                                               1\n",
       "Zuylen van Nyevelt, Mechtelt, fl. 1666                                      1\n",
       "Berckel, Clemens, fl. 1613                                                  1\n",
       "Cotton, John (Sir), 1621-1702                                               1\n",
       "Schagen, Diederik, fl. 1660                                                 1\n",
       "Magerus, Petrus, 1609-1653                                                  1\n",
       "Petit, Pierre, 1598 or before-1677                                          1\n",
       "Nanteuil, Robert, 1623-1678                                                 1\n",
       "Enclos, Anne de l', 1620-1705                                               1\n",
       "Name: addressee, dtype: int64"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_constantijn = df[df['collection'] == 'Huygens, Constantijn']\n",
    "df_constantijn['addressee'].value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['author_freq'] = df.groupby(['author'])['id'].transform('count')\n",
    "df['addressee_freq'] = df.groupby(['addressee'])['id'].transform('count')\n",
    "df['correspondents_freq'] = df.author_freq + df.addressee_freq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                                    author                                             addressee                              \n",
       "0000ab7c-f54a-493a-a066-ee929eedd1e3  Johnson, John, 1662-1725                           Charlett, Arthur (Reverend), 1655-1722     1\n",
       "0000bd85-9139-4fec-b362-78b3f4a6b9c2  Boywer, F., fl. 1737                               Rawlinson, Richard (Dr), 1690-1755         1\n",
       "0000d067-74f1-46ed-b4bc-e9d2765091e6  Alciatus, Francesco (Cardinal), 1522-1580          Aytta, Viglius Zuichemius ab, 1507-1577    1\n",
       "0002ce4c-db77-4863-a1ec-e8be9b9d121b  Buffon, George Louis Leclerc de, 1707-1788         Jurin, James, 1684-1750                    1\n",
       "0002dbb4-2785-4b05-aa59-2bbb3f654802  Sandford, Daniel (Reverend), 1729-1770             Ballard, George, 1705-1755                 1\n",
       "                                                                                                                                   ..\n",
       "fffdf05b-cba6-442e-a563-18598314a021  Howel, John, fl. 1760-1781                         Gough, Richard, 1735-1809                  1\n",
       "fffe4dd7-c074-4de0-b422-dc1c635eca0f  Villiers, Christophe de, fl. 1633-1639             Mersenne, Marin, 1588-1648                 1\n",
       "fffe9c50-5132-476e-9173-9b680d0f916e  August II of Braunschweig-Wolfenbüttel, 1579-1666  Andreae, Johann Valentin, 1586-1654        1\n",
       "fffefd34-9784-4673-920e-1696883ef2a2  Zapata, Rodrigo, fl. 1574                          Agustín, Antonio, 1517-1586                1\n",
       "ffffcc85-6dbc-4a57-8942-40323c4d18fc  Willis, Browne, 1682-1760                          Rawlinson, Richard (Dr), 1690-1755         1\n",
       "Length: 127418, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby(['id', 'author', 'addressee']).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "ids = list(df.sort_values('correspondents_freq').id)\n",
    "auths = list(df.sort_values('correspondents_freq').author)\n",
    "addrs = list(df.sort_values('correspondents_freq').addressee)\n",
    "auth_freqs = list(df.sort_values('correspondents_freq').author_freq)\n",
    "addr_freqs = list(df.sort_values('correspondents_freq').addressee_freq)\n",
    "\n",
    "auths = [auth if isinstance(auth, str) else None for auth in auths]\n",
    "addrs = [addr if isinstance(addr, str) else None for addr in addrs]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'addr': 'Jones, Robert (Reverend), fl. 1698',\n",
       " 'addr_freq': 1.0,\n",
       " 'auth': 'Meare, John, 1649-1710',\n",
       " 'author_freq': 1.0,\n",
       " 'id': 'aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a'}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corrs = [{'id': id, 'auth': auth, 'addr': addr, 'author_freq': auth_freq, 'addr_freq': addr_freq} for id, auth, addr, auth_freq, addr_freq in zip(ids, auths, addrs, auth_freqs, addr_freqs)]\n",
    "\n",
    "corrs[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1942\n",
      "3873\n"
     ]
    }
   ],
   "source": [
    "from collections import OrderedDict\n",
    "queued = {}\n",
    "fetch = OrderedDict()\n",
    "seen = {}\n",
    "for corr in corrs:\n",
    "    if corr['auth'] not in queued and corr['addr'] not in queued:\n",
    "        queued[corr['auth']] = corr['id']\n",
    "        queued[corr['addr']] = corr['id']\n",
    "        fetch[corr['id']] = corr\n",
    "\n",
    "print(len(fetch.keys()))\n",
    "print(len(queued.keys()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13827\n",
      "15758\n"
     ]
    }
   ],
   "source": [
    "for corr in corrs:\n",
    "    if corr['auth'] not in queued:\n",
    "        queued[corr['auth']] = corr['id']\n",
    "        fetch[corr['id']] = corr\n",
    "    elif corr['auth'] not in queued:\n",
    "        queued[corr['auth']] = corr['id']\n",
    "        fetch[corr['id']] = corr\n",
    "\n",
    "print(len(fetch.keys()))\n",
    "print(len(queued.keys()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "http://emlo.bodleian.ox.ac.uk/profile/work/aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a\n"
     ]
    }
   ],
   "source": [
    "for corr_id in fetch:\n",
    "    url = f'http://emlo.bodleian.ox.ac.uk/profile/work/{corr_id}'\n",
    "    print(url)\n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>id</th>\n",
       "      <th>type</th>\n",
       "      <th>collection</th>\n",
       "      <th>date</th>\n",
       "      <th>author</th>\n",
       "      <th>addressee</th>\n",
       "      <th>origin</th>\n",
       "      <th>destination</th>\n",
       "      <th>repository</th>\n",
       "      <th>author_freq</th>\n",
       "      <th>addressee_freq</th>\n",
       "      <th>correspondents_freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>73349</th>\n",
       "      <td>17320</td>\n",
       "      <td>aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a</td>\n",
       "      <td>Letter</td>\n",
       "      <td>Bodleian card catalogue</td>\n",
       "      <td>30 August 1698</td>\n",
       "      <td>Meare, John, 1649-1710</td>\n",
       "      <td>Jones, Robert (Reverend), fl. 1698</td>\n",
       "      <td>Oxfordshire, England</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Bodleian Library, University of Oxford: MS Bal...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Unnamed: 0                                    id    type  \\\n",
       "73349       17320  aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a  Letter   \n",
       "\n",
       "                    collection            date                  author  \\\n",
       "73349  Bodleian card catalogue  30 August 1698  Meare, John, 1649-1710   \n",
       "\n",
       "                                addressee                origin destination  \\\n",
       "73349  Jones, Robert (Reverend), fl. 1698  Oxfordshire, England         NaN   \n",
       "\n",
       "                                              repository  author_freq  \\\n",
       "73349  Bodleian Library, University of Oxford: MS Bal...          1.0   \n",
       "\n",
       "       addressee_freq  correspondents_freq  \n",
       "73349             1.0                  2.0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import requests\n",
    "from bs4 import BeautifulSoup as bsoup\n",
    "\n",
    "df[df.id == 'aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a']\n",
    "\n",
    "#response = requests.get(url)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_relation_info(rel_type, detail_soup):\n",
    "    rel_type_soup = detail_soup.find_all(class_=rel_type)\n",
    "    if len(rel_type_soup) == 0:\n",
    "        return None\n",
    "    relation_soup = rel_type_soup[0].find_all(class_='relations')[0]\n",
    "    return {\n",
    "        'relation_type': rel_type.split(' '),\n",
    "        'relation_text': [string for string in relation_soup.stripped_strings]\n",
    "    }\n",
    "\n",
    "def get_provenance(page_soup):\n",
    "    prov_soup = page_soup.find_all(class_='provenance')[0]\n",
    "    prov = prov_soup.text\n",
    "    return prov.replace('Source of data: ','')\n",
    "\n",
    "def get_page_details(corr_id, page_soup):\n",
    "    page_details = {\n",
    "        'correspondence_id': corr_id,\n",
    "        'relations': [],\n",
    "        'provenance': get_provenance(page_soup)\n",
    "    }\n",
    "    detail_soup = page_soup.find(id='details')\n",
    "    if detail_soup:\n",
    "        rel_types = ['people authors', 'people recipients', 'locations origin', 'locations destination']\n",
    "        relation_info = [get_relation_info(rel_type, detail_soup) for rel_type in rel_types]\n",
    "        page_details['relations'] = [relation for relation in relation_info if relation != None]\n",
    "    return page_details\n",
    "\n",
    "def get_correspondence_page(corr_id):\n",
    "    url = f'http://emlo.bodleian.ox.ac.uk/profile/work/{corr_id}'\n",
    "    response = requests.get(url)\n",
    "    page_soup = bsoup(response.content)\n",
    "    return get_page_details(corr_id, page_soup)\n",
    "\n",
    "corr_id = 'aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a'\n",
    "#detail_doc = get_page_details(corr_id, page_soup)\n",
    "detail_index = 'emlo_page_details'\n",
    "from elasticsearch import Elasticsearch\n",
    "\n",
    "es = Elasticsearch()\n",
    "\n",
    "#es.index(index=detail_index, doc_type='page_detail', id=detail_doc['correspondence_id'], body=detail_doc)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'addr': 'Jones, Robert (Reverend), fl. 1698',\n",
       " 'addr_freq': 1.0,\n",
       " 'auth': 'Meare, John, 1649-1710',\n",
       " 'author_freq': 1.0,\n",
       " 'id': 'aabeebf7-4c5b-4bc2-a2ec-8ace326cfa7a'}"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "headers = {\n",
    "    'user-agent': 'DataScopesAnalyzer (https://marijnkoolen.github.io/Data-Scopes-Developers-2018/)',\n",
    "    'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',\n",
    "    'Accept-Language': 'en-gb',\n",
    "}\n",
    "\n",
    "fetch[corr_id]\n",
    "#time.sleep(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "skipped 1000\n",
      "skipped 2000\n",
      "skipped 3000\n",
      "skipped 4000\n",
      "skipped 5000\n",
      "skipped 6000\n",
      "skipped 7000\n",
      "skipped 8000\n",
      "skipped 9000\n",
      "skipped 10000\n",
      "skipped 11000\n",
      "skipped 12000\n",
      "12100 correspondence pages fetched\n",
      "12200 correspondence pages fetched\n",
      "12300 correspondence pages fetched\n",
      "12400 correspondence pages fetched\n",
      "12500 correspondence pages fetched\n",
      "12600 correspondence pages fetched\n",
      "12700 correspondence pages fetched\n",
      "12800 correspondence pages fetched\n",
      "12900 correspondence pages fetched\n",
      "13000 correspondence pages fetched\n",
      "13100 correspondence pages fetched\n",
      "13200 correspondence pages fetched\n",
      "13300 correspondence pages fetched\n",
      "13400 correspondence pages fetched\n",
      "13500 correspondence pages fetched\n",
      "13600 correspondence pages fetched\n",
      "13700 correspondence pages fetched\n",
      "13800 correspondence pages fetched\n"
     ]
    }
   ],
   "source": [
    "from elasticsearch import exceptions\n",
    "\n",
    "skip = 0\n",
    "\n",
    "for ci, corr_id in enumerate(fetch):\n",
    "    if es.exists(index=detail_index, id=corr_id):\n",
    "        #print('skip', corr_id)\n",
    "        skip += 1\n",
    "        if skip % 1000 == 0:\n",
    "            print('skipped', skip)\n",
    "        continue\n",
    "    #print('fetching page for', corr_id)\n",
    "    detail_doc = get_correspondence_page(corr_id)\n",
    "    try:\n",
    "        detail_doc['author'] = fetch[corr_id]['auth']\n",
    "        detail_doc['addressee'] = fetch[corr_id]['addr']\n",
    "    except TypeError:\n",
    "        print(fetch[corr_id])\n",
    "        raise\n",
    "    try:\n",
    "        es.index(index=detail_index, doc_type='page_detail', id=detail_doc['correspondence_id'], body=detail_doc)\n",
    "    except exceptions.RequestError:\n",
    "        print(detail_doc)\n",
    "        raise\n",
    "    time.sleep(10)\n",
    "    if (ci+1) % 100 == 0:\n",
    "        print(ci+1, 'correspondence pages fetched')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}