{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "# Analyse des dialogues dans l'Avare de Molière"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "## Choix du fichier source\n",
    "\n",
    "*L'Avare* de Molière est disponible dans plusieurs formats différents. \n",
    "Cependant, tous ne se prêtent pas à une analyse sémantique d'une pièce de théâtre. \n",
    "Les formats reposant sur du texte brut (tels que Markdown ou iramuteq, et par extension, les fichiers ne contenant que les prises de parole) compliquent le triage des [didascalies](https://fr.wikipedia.org/wiki/Didascalie_(théâtre)) et autres blocs de texte insérés dans les scènes, et ne relevant pas directement du dialogue. \n",
    "Notre choix devra donc se porter sur un format plus structuré.\n",
    "Nous pourrions tenter l'analyse des fichiers epub ou kindle, mais ce sont des formats destinés à la présentation, ce qui rendrait leur analyse inutilement complexe et coûteuse, alors que de meilleurs formats sont disponibles.\n",
    "\n",
    "Les formats de fichier constituant de meilleurs candidats pour une analyse sémantique sont basés sur XML, qui permet la structuration du contenu : TEI (conçu par le *Text Encoding Initiative Consortium*), TXM (co-développé par l'École normale supérieure de Lyon et l'université de Franche-Comté), et HTML (le langage de balisage du web). \n",
    "\n",
    "Ces trois formats sont basés sur XML, et peuvent donc théoriquement être exploités avec une même API ([XPath](https://fr.wikipedia.org/wiki/XPath)), sans nécessiter de bibliothèque tierce. \n",
    "\n",
    "HTML présente toutefois des avantages considérables : son exploitation par, au minimum, quelques centaines de millions de sites web à travers le monde, et sa gouvernance par un consortium d'entreprises comme Apple, Google ou Mozilla outre-Atlantique, ou encore l'Inria en France. \n",
    "C'est le format qui a créé internet, et il est réutilisé dans des contextes très différents.\n",
    "De plus, en tant que développeur web depuis 30 ans, l'auteur de la présente analyse ne cache pas son intérêt particulier pour ce format, avec lequel il est bien plus familier qu'avec les autres.\n",
    "\n",
    "Nous poursuivrons donc cette étude avec le fichier `moliere_avare.html` [mis à disposition](http://dramacode.github.io/html/moliere_avare.html) par [dramacode](https://dramacode.github.io)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ouverture du fichier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Commençons par regrouper les importations, afin d'en avoir une vue d'ensemble."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Parseur XML, requêtes XPath\n",
    "import xml.etree.ElementTree as ET\n",
    "\n",
    "# Analyse\n",
    "import pandas as pd\n",
    "\n",
    "# Utile pour la gestion des caractères accentués\n",
    "import locale\n",
    "\n",
    "# Nous permet de combiner deux listes de longueur indéterminée\n",
    "from itertools import zip_longest\n",
    "\n",
    "# Utilisation d'expressions régulières (regex)\n",
    "import re\n",
    "\n",
    "# Traçage de graphiques\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Permet de définir et d'afficher la colormap des personnages\n",
    "from itertools import cycle\n",
    "import matplotlib.patches as mpatches\n",
    "\n",
    "# Utiles aux calculs effectués pour les graphiques\n",
    "import math\n",
    "import numpy as np\n",
    "\n",
    "# Dépendances pour le graphe final\n",
    "from pathlib import Path\n",
    "from jinja2 import Environment, FileSystemLoader\n",
    "\n",
    "# Pose un certain nombre de problèmes pour l'exportation,\n",
    "# en raison de sa nature dynamique (et écrit en javascript)\n",
    "import pyvis\n",
    "from pyvis.network import Network\n",
    "import matplotlib.colors as mcolors\n",
    "from IPython.display import HTML"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# La définition de la locale nous permettra de gérer correctement les majuscules\n",
    "# accentuées\n",
    "locale.setlocale(locale.LC_COLLATE, \"fr_FR.UTF-8\")  # ou \"fr_FR.UTF-8\", \"fr_FR\" selon le système\n",
    "\n",
    "ns = {\"x\": \"http://www.w3.org/1999/xhtml\"}\n",
    "root = ET.parse(\"moliere_avare.html\").getroot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Outils préliminaires\n",
    "\n",
    "Nous travaillons sur une pièce de théâtre, par définition divisée en actes et en scènes.\n",
    "Nous allons donc nous créer quelques outils pour accéder facilement à ces éléments structurés, que nous complèterons de diverses fonctions utilisées à plusieurs reprises dans notre étude."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Nettoyage des titres\n",
    "def clean_title(el, tag):\n",
    "    return \"\".join(el.find(tag, ns).itertext()).replace(\"§\", \"\").strip()\n",
    "\n",
    "# Actes avec ordre explicite\n",
    "def list_acts():\n",
    "    acts = []\n",
    "\n",
    "    for idx, act in enumerate(root.findall(\".//x:section[@class='div1 act level2']\", ns)):\n",
    "        acts.append({\n",
    "            \"id\": act.get(\"id\"),\n",
    "            \"title\": clean_title(act, \"x:h2\"),\n",
    "            \"node\": act,\n",
    "            \"order\": idx,\n",
    "        })\n",
    "\n",
    "    return acts\n",
    "\n",
    "# Scènes d’un acte donné, avec ordre explicite\n",
    "def list_scenes(act=None, act_id=None):\n",
    "    if act is None:\n",
    "        if act_id is None:\n",
    "            raise ValueError(\"Un élément `act` ou un identifiant doit être spécifié\")\n",
    "\n",
    "        act = root.find(f\".//x:section[@class='div1 act level2'][@id='{act_id}']\", ns)\n",
    "\n",
    "        if act is None:\n",
    "            raise ValueError(f\"Acte introuvable: {act_id}\")\n",
    "\n",
    "    scenes = []\n",
    "\n",
    "    for idx, scene in enumerate(act.findall(\"x:section[@class='div2 scene level3']\", ns)):\n",
    "        scenes.append({\n",
    "            \"id\": scene.get(\"id\"),\n",
    "            \"title\": clean_title(scene, \"x:h3\"),\n",
    "            \"node\": scene,\n",
    "            \"order\": idx,\n",
    "        })\n",
    "\n",
    "    return scenes\n",
    "\n",
    "# On compacte les espaces pour calquer le comptage sur l'OBVIL\n",
    "# Supprime les balises <i> de l'élément HTML soumis.\n",
    "# Cela permet de ne conserver que les noms de personnages extraits des blocs de dialogue,\n",
    "# sans les didascalies.\n",
    "def text_without_i(el):\n",
    "    parts = []\n",
    "\n",
    "    if el.tag != f\"{{{ns['x']}}}i\" and el.text and el.text.strip():\n",
    "        parts.append(el.text.strip())\n",
    "\n",
    "    for child in el:\n",
    "        if child.tag != f\"{{{ns['x']}}}i\":\n",
    "            parts.extend(text_without_i(child))\n",
    "\n",
    "        # on garde toujours le texte suivant, même si le noeud enfant est une balise <i>\n",
    "        if child.tail and child.tail.strip():\n",
    "            parts.append(child.tail.strip())\n",
    "\n",
    "    return parts\n",
    "\n",
    "# Formate le nom d'un acteur extrait d'un dialogue\n",
    "def speaker_name(sp):\n",
    "    name = \" \".join(text_without_i(sp)).strip()\n",
    "\n",
    "    # nettoyage simple de la ponctuation finale\n",
    "    name = name.rstrip(\",;:\").strip()\n",
    "\n",
    "    return name\n",
    "\n",
    "# Résolution d'un nom d'acteur à partir de notre table de correspondance\n",
    "alias_index = {}\n",
    "def resolve_name(name):\n",
    "    return alias_index.get(name, name)\n",
    "\n",
    "# Extrait le texte brut d'un dialogue soumis sous la forme d'un élément HTML\n",
    "def speech_text(sp):\n",
    "    parts = []\n",
    "\n",
    "    for p in sp.findall(\".//x:p[@class='p autofirst']\", ns):\n",
    "        parts.extend(text_without_i(p))\n",
    "\n",
    "    raw = \" \".join(parts)\n",
    "    return \" \".join(raw.split()).strip()\n",
    "\n",
    "# Compte le nombre de mots d'un texte brut.\n",
    "# On utilise ici une regex simple dédiée à cet usage.\n",
    "def word_count(txt):\n",
    "    return len(re.findall(r\"\\b\\w+\\b\", txt, flags=re.UNICODE))\n",
    "\n",
    "# Conversion d'un texte en nombre de lignes (60 caractères par ligne)\n",
    "def line_count(txt, line_length=60):\n",
    "    return len(txt) / line_length if txt else 0\n",
    "\n",
    "# Retourne l'acteur associé à une réplique (<div class=\"sp\">)\n",
    "def speech_actor(sp):\n",
    "    speaker_el = sp.find(\"x:p[@class='speaker']\", ns)\n",
    "\n",
    "    if speaker_el is None:\n",
    "        return \"\"\n",
    "\n",
    "    return resolve_name(speaker_name(speaker_el))\n",
    "\n",
    "# Liste les répliques d'une scène (par noeud ou identifiant) avec texte et nombre de mots\n",
    "def scene_speeches(scene=None, scene_id=None):\n",
    "    if scene is None:\n",
    "        if scene_id is None:\n",
    "            raise ValueError(\"Un élément `scene` ou un identifiant doit être spécifié\")\n",
    "\n",
    "        scene = root.find(f\".//x:section[@class='div2 scene level3'][@id='{scene_id}']\", ns)\n",
    "\n",
    "        if scene is None:\n",
    "            raise ValueError(f\"Scène introuvable: {scene_id}\")\n",
    "\n",
    "    speeches = []\n",
    "\n",
    "    for sp_div in scene.findall(\".//x:div[@class='sp']\", ns):\n",
    "        speaker = speech_actor(sp_div)\n",
    "\n",
    "        if not speaker:\n",
    "            continue\n",
    "\n",
    "        txt = speech_text(sp_div)\n",
    "\n",
    "        speeches.append({\n",
    "            \"speaker\": speaker,\n",
    "            \"text\": txt,\n",
    "            \"word_count\": word_count(txt),\n",
    "            \"node\": sp_div,\n",
    "        })\n",
    "\n",
    "    return speeches\n",
    "\n",
    "# Création d'une colormap associant une couleur à un personnage\n",
    "def create_actors_colormap(personnages):\n",
    "    # Définition d'une palette de couleurs pour les personnages\n",
    "    palette = cycle(plt.cm.tab20.colors)\n",
    "    color_map = {}\n",
    "\n",
    "    for p in personnages:\n",
    "        color_map[p] = next(palette)\n",
    "\n",
    "    return color_map"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vérifions que nous obtenons bien la liste des actes et des scènes :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Acte I - Scène I01\n",
      "Acte I - Scène I02\n",
      "Acte I - Scène I03\n",
      "Acte I - Scène I04\n",
      "Acte I - Scène I05\n",
      "Acte II - Scène II01\n",
      "Acte II - Scène II02\n",
      "Acte II - Scène II03\n",
      "Acte II - Scène II04\n",
      "Acte II - Scène II05\n",
      "Acte III - Scène III01\n",
      "Acte III - Scène III02\n",
      "Acte III - Scène III03\n",
      "Acte III - Scène III04\n",
      "Acte III - Scène III05\n",
      "Acte III - Scène III06\n",
      "Acte III - Scène III07\n",
      "Acte III - Scène III08\n",
      "Acte III - Scène III09\n",
      "Acte IV - Scène IV01\n",
      "Acte IV - Scène IV02\n",
      "Acte IV - Scène IV03\n",
      "Acte IV - Scène IV04\n",
      "Acte IV - Scène IV05\n",
      "Acte IV - Scène IV06\n",
      "Acte IV - Scène IV07\n",
      "Acte V - Scène V01\n",
      "Acte V - Scène V02\n",
      "Acte V - Scène V03\n",
      "Acte V - Scène V04\n",
      "Acte V - Scène V05\n",
      "Acte V - Scène V06\n"
     ]
    }
   ],
   "source": [
    "for act in list_acts():\n",
    "    for scene in list_scenes(act=act[\"node\"]):\n",
    "        print (\"Acte \" + act[\"id\"] + \" - Scène \" + scene[\"id\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Obtention de la liste des acteurs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "La requête xpath suivante permet d'extraire la liste des acteurs donnée au début du fichier, autrement appelée [_dramatis personae_](https://fr.wikipedia.org/wiki/Dramatis_personæ_(théâtre))."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>Personnage</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Harpagon</td>\n",
       "      <td>Père de Cléante et d'Élise, et Amoureux de Mar...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Cléante</td>\n",
       "      <td>Fils d'Harpagon, Amant de Mariane.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Élise</td>\n",
       "      <td>Fille d'Harpagon, Amante de Valère.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Valère</td>\n",
       "      <td>Fils d'Anselme, et Amant d'Élise.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Mariane</td>\n",
       "      <td>Amante de Cléante, et aimée d'Harpagon.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Anselme</td>\n",
       "      <td>Père de Valère et de Mariane.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Frosine</td>\n",
       "      <td>Femme d'Intrigue.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Maitre Simon</td>\n",
       "      <td>Courtier.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Maitre Jacques</td>\n",
       "      <td>Cuisinier et Cocher d'Harpagon.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>La Flèche</td>\n",
       "      <td>Valet de Cléante.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Dame Claude</td>\n",
       "      <td>Servante d'Harpagon.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Brindavoine</td>\n",
       "      <td>laquais d'Harpagon.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>La Merluche</td>\n",
       "      <td>laquais d'Harpagon.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Le commissaire</td>\n",
       "      <td>et son clerc.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Personnage                                        Description\n",
       "0         Harpagon  Père de Cléante et d'Élise, et Amoureux de Mar...\n",
       "1          Cléante                 Fils d'Harpagon, Amant de Mariane.\n",
       "2            Élise                Fille d'Harpagon, Amante de Valère.\n",
       "3           Valère                  Fils d'Anselme, et Amant d'Élise.\n",
       "4          Mariane            Amante de Cléante, et aimée d'Harpagon.\n",
       "5          Anselme                      Père de Valère et de Mariane.\n",
       "6          Frosine                                  Femme d'Intrigue.\n",
       "7     Maitre Simon                                          Courtier.\n",
       "8   Maitre Jacques                    Cuisinier et Cocher d'Harpagon.\n",
       "9        La Flèche                                  Valet de Cléante.\n",
       "10     Dame Claude                               Servante d'Harpagon.\n",
       "11     Brindavoine                                laquais d'Harpagon.\n",
       "12     La Merluche                                laquais d'Harpagon.\n",
       "13  Le commissaire                                      et son clerc."
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Retourne une liste formatée des personnages définis dans la dramatis personae.\n",
    "# Prend en compte les majuscules accentuées.\n",
    "def dramatis_personae():\n",
    "    rows = []\n",
    "\n",
    "    # Requête xpath permettant d'obtenir la liste des balises <li> listant les acteurs\n",
    "    for li in root.findall(\".//x:div[@id='castList']//x:li\", ns):\n",
    "        # L'acteur se trouve dans une balise <span>\n",
    "        span = li.find(\"x:span\", ns)\n",
    "        name = span.text.strip()\n",
    "\n",
    "        # description = texte qui suit le <span> dans la même balise <li>\n",
    "        desc = (span.tail or \"\").strip()\n",
    "\n",
    "        if desc.startswith(\",\"):\n",
    "            desc = desc[1:].strip()\n",
    "\n",
    "        rows.append({\"Personnage\": name, \"Description\": desc})\n",
    "\n",
    "    return pd.DataFrame(rows)\n",
    "\n",
    "dramatis_personae = dramatis_personae()\n",
    "\n",
    "# Affichage de la liste\n",
    "dramatis_personae"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nous pouvons déjà constater que le commissaire et son clerc sont considérés comme un acteur unique.\n",
    "Nous verrons plus tard si cette information est importante (par exemple, si le clerc s'exprime en son nom propre)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nous allons confronter cette liste avec la liste des protagonistes mentionnés en introduction de chaque scène, puis avec ceux qui interviennent \"réellement\", c'est-à-dire ceux qui ont une ligne de dialogue.\n",
    "Cette étape devra nous permettre d'identifier des différences d'orthographe subtiles qu'il sera utile de gérer."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Noms des personnages par scène"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nous allons itérer sur chaque acte, puis chaque scène, afin de consulter la liste des protagonistes. \n",
    "N'oublions pas que ces listes sont facultatives, et ne désignent pas les acteurs dotés d'une réplique.\n",
    "Néanmoins, nous pourrions identifier des éléments potentiellement intéressants, tels que des orthographes différentes ou une anomalie quelconque."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "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>Acte</th>\n",
       "      <th>Scène</th>\n",
       "      <th>Protagonistes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Valère, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Cléante, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Harpagon, La Flèche]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Élise, Cléante, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Valère, Harpagon, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Cléante, La Flèche]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Maître Simon, Harpagon, Cléante, La Flèche]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Frosine, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[La Flèche, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Harpagon, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Harpagon, Cléante, Élise, Valère, Dame Claude...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Maître Jacques, Valère]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Frosine, Mariane, Maître Jacques]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Mariane, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Harpagon, Frosine, Mariane]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[Élise, Harpagon, Mariane, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VII</td>\n",
       "      <td>[Cléante, Harpagon, Élise, Mariane, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VIII</td>\n",
       "      <td>[Harpagon, Mariane, Frosine, Cléante, Brindavo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène IX</td>\n",
       "      <td>[Harpagon, Mariane, Cléante, Élise, Frosine, L...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Cléante, Mariane, Élise, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Harpagon, Cléante, Mariane, Élise, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Harpagon, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Maître Jacques, Harpagon, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Cléante, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[La Flèche, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène VII</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Harpagon, Le Commissaire, son Clerc]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Maître Jacques, Harpagon, Le Commissaire, son...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Valère, Harpagon, le Commissaire, son Clerc, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Élise, Mariane, Frosine, Harpagon, Valère, Ma...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Anselme, Harpagon, Élise, Mariane, Frosine, V...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[Cléante, Valère, Mariane, Élise, Frosine, Har...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Acte           Scène  \\\n",
       "0   Acte Premier  Scène Première   \n",
       "1   Acte Premier        Scène II   \n",
       "2   Acte Premier       Scène III   \n",
       "3   Acte Premier        Scène IV   \n",
       "4   Acte Premier         Scène V   \n",
       "5        Acte II  Scène Première   \n",
       "6        Acte II        Scène II   \n",
       "7        Acte II       Scène III   \n",
       "8        Acte II        Scène IV   \n",
       "9        Acte II         Scène V   \n",
       "10      Acte III  Scène Première   \n",
       "11      Acte III        Scène II   \n",
       "12      Acte III       Scène III   \n",
       "13      Acte III        Scène IV   \n",
       "14      Acte III         Scène V   \n",
       "15      Acte III        Scène VI   \n",
       "16      Acte III       Scène VII   \n",
       "17      Acte III      Scène VIII   \n",
       "18      Acte III        Scène IX   \n",
       "19       Acte IV  Scène Première   \n",
       "20       Acte IV        Scène II   \n",
       "21       Acte IV       Scène III   \n",
       "22       Acte IV        Scène IV   \n",
       "23       Acte IV         Scène V   \n",
       "24       Acte IV        Scène VI   \n",
       "25       Acte IV       Scène VII   \n",
       "26        Acte V  Scène Première   \n",
       "27        Acte V        Scène II   \n",
       "28        Acte V       Scène III   \n",
       "29        Acte V        Scène IV   \n",
       "30        Acte V         Scène V   \n",
       "31        Acte V        Scène VI   \n",
       "\n",
       "                                        Protagonistes  \n",
       "0                                     [Valère, Élise]  \n",
       "1                                    [Cléante, Élise]  \n",
       "2                               [Harpagon, La Flèche]  \n",
       "3                          [Élise, Cléante, Harpagon]  \n",
       "4                           [Valère, Harpagon, Élise]  \n",
       "5                                [Cléante, La Flèche]  \n",
       "6        [Maître Simon, Harpagon, Cléante, La Flèche]  \n",
       "7                                 [Frosine, Harpagon]  \n",
       "8                                [La Flèche, Frosine]  \n",
       "9                                 [Harpagon, Frosine]  \n",
       "10  [Harpagon, Cléante, Élise, Valère, Dame Claude...  \n",
       "11                           [Maître Jacques, Valère]  \n",
       "12                 [Frosine, Mariane, Maître Jacques]  \n",
       "13                                 [Mariane, Frosine]  \n",
       "14                       [Harpagon, Frosine, Mariane]  \n",
       "15                [Élise, Harpagon, Mariane, Frosine]  \n",
       "16       [Cléante, Harpagon, Élise, Mariane, Frosine]  \n",
       "17  [Harpagon, Mariane, Frosine, Cléante, Brindavo...  \n",
       "18  [Harpagon, Mariane, Cléante, Élise, Frosine, L...  \n",
       "19                 [Cléante, Mariane, Élise, Frosine]  \n",
       "20       [Harpagon, Cléante, Mariane, Élise, Frosine]  \n",
       "21                                [Harpagon, Cléante]  \n",
       "22                [Maître Jacques, Harpagon, Cléante]  \n",
       "23                                [Cléante, Harpagon]  \n",
       "24                               [La Flèche, Cléante]  \n",
       "25                                                 []  \n",
       "26              [Harpagon, Le Commissaire, son Clerc]  \n",
       "27  [Maître Jacques, Harpagon, Le Commissaire, son...  \n",
       "28  [Valère, Harpagon, le Commissaire, son Clerc, ...  \n",
       "29  [Élise, Mariane, Frosine, Harpagon, Valère, Ma...  \n",
       "30  [Anselme, Harpagon, Élise, Mariane, Frosine, V...  \n",
       "31  [Cléante, Valère, Mariane, Élise, Frosine, Har...  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def list_scene_protagonists():\n",
    "    rows = []\n",
    "\n",
    "    for act in list_acts():\n",
    "        for scene in list_scenes(act=act[\"node\"]):\n",
    "            stage = scene[\"node\"].find(\"x:div[@class='stage stage']\", ns)\n",
    "\n",
    "            # Si nous trouvons un noeud xpath pour cette requête, c'est un personnage\n",
    "            if stage is not None:\n",
    "                raw = \"\".join(stage.itertext()).strip()\n",
    "                people = [p.strip() for p in raw.split(\",\") if p.strip()]\n",
    "            else:\n",
    "                people = []\n",
    "\n",
    "            rows.append({\n",
    "                \"Acte\": act[\"title\"],\n",
    "                \"Scène\": scene[\"title\"],\n",
    "                \"Protagonistes\": people,\n",
    "            })\n",
    "\n",
    "    return pd.DataFrame(rows)\n",
    "\n",
    "df_scenes = list_scene_protagonists()\n",
    "df_scenes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "Nous voyons ici que la scène VII de l'acte IV ne contient aucun protagoniste déclaré dans la liste attenante[^1], mais nous allons de toute façon la compléter par l'extraction individuelle des interventions concrètes de chaque acteur.\n",
    "\n",
    "[^1]: Cette liste n'est pas obligatoire dans le contexte théâtral. Ici, on peut supposer que son absence est dûe à un monologue par exemple."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "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>Acte</th>\n",
       "      <th>Scène</th>\n",
       "      <th>Intervenants</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Valère, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Cléante, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Harpagon, La Flèche]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Harpagon, Cléante, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Acte Premier</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Harpagon, Valère, Élise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Cléante, La Flèche]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Maître simon, Harpagon, La Flèche, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Frosine, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[La Flèche, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Harpagon, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Harpagon, Maître Jacques, La Merluche, Brinda...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Valère, Maître Jacques]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Frosine, Maître Jacques]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Mariane, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Harpagon, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[Mariane, Élise, Harpagon, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VII</td>\n",
       "      <td>[Cléante, Mariane, Harpagon, Frosine, Valère]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène VIII</td>\n",
       "      <td>[Brindavoine, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Acte III</td>\n",
       "      <td>Scène IX</td>\n",
       "      <td>[La Merluche, Harpagon, Cléante, Valère]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Cléante, Élise, Mariane, Frosine]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Harpagon, Élise, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Harpagon, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Maître Jacques, Cléante, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Cléante, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[La Flèche, Cléante]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Acte IV</td>\n",
       "      <td>Scène VII</td>\n",
       "      <td>[Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène Première</td>\n",
       "      <td>[Le Commissaire, Harpagon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>[Maître Jacques, Harpagon, Le Commissaire]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>[Harpagon, Valère, Maître Jacques]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène IV</td>\n",
       "      <td>[Harpagon, Valère, Élise, Maître Jacques, Fros...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène V</td>\n",
       "      <td>[Anselme, Harpagon, Valère, Mariane, Maître Ja...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>[Cléante, Harpagon, Mariane, Anselme, Le Commi...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Acte           Scène  \\\n",
       "0   Acte Premier  Scène Première   \n",
       "1   Acte Premier        Scène II   \n",
       "2   Acte Premier       Scène III   \n",
       "3   Acte Premier        Scène IV   \n",
       "4   Acte Premier         Scène V   \n",
       "5        Acte II  Scène Première   \n",
       "6        Acte II        Scène II   \n",
       "7        Acte II       Scène III   \n",
       "8        Acte II        Scène IV   \n",
       "9        Acte II         Scène V   \n",
       "10      Acte III  Scène Première   \n",
       "11      Acte III        Scène II   \n",
       "12      Acte III       Scène III   \n",
       "13      Acte III        Scène IV   \n",
       "14      Acte III         Scène V   \n",
       "15      Acte III        Scène VI   \n",
       "16      Acte III       Scène VII   \n",
       "17      Acte III      Scène VIII   \n",
       "18      Acte III        Scène IX   \n",
       "19       Acte IV  Scène Première   \n",
       "20       Acte IV        Scène II   \n",
       "21       Acte IV       Scène III   \n",
       "22       Acte IV        Scène IV   \n",
       "23       Acte IV         Scène V   \n",
       "24       Acte IV        Scène VI   \n",
       "25       Acte IV       Scène VII   \n",
       "26        Acte V  Scène Première   \n",
       "27        Acte V        Scène II   \n",
       "28        Acte V       Scène III   \n",
       "29        Acte V        Scène IV   \n",
       "30        Acte V         Scène V   \n",
       "31        Acte V        Scène VI   \n",
       "\n",
       "                                         Intervenants  \n",
       "0                                     [Valère, Élise]  \n",
       "1                                    [Cléante, Élise]  \n",
       "2                               [Harpagon, La Flèche]  \n",
       "3                          [Harpagon, Cléante, Élise]  \n",
       "4                           [Harpagon, Valère, Élise]  \n",
       "5                                [Cléante, La Flèche]  \n",
       "6        [Maître simon, Harpagon, La Flèche, Cléante]  \n",
       "7                                 [Frosine, Harpagon]  \n",
       "8                                [La Flèche, Frosine]  \n",
       "9                                 [Harpagon, Frosine]  \n",
       "10  [Harpagon, Maître Jacques, La Merluche, Brinda...  \n",
       "11                           [Valère, Maître Jacques]  \n",
       "12                          [Frosine, Maître Jacques]  \n",
       "13                                 [Mariane, Frosine]  \n",
       "14                                [Harpagon, Frosine]  \n",
       "15                [Mariane, Élise, Harpagon, Frosine]  \n",
       "16      [Cléante, Mariane, Harpagon, Frosine, Valère]  \n",
       "17                            [Brindavoine, Harpagon]  \n",
       "18           [La Merluche, Harpagon, Cléante, Valère]  \n",
       "19                 [Cléante, Élise, Mariane, Frosine]  \n",
       "20                         [Harpagon, Élise, Cléante]  \n",
       "21                                [Harpagon, Cléante]  \n",
       "22                [Maître Jacques, Cléante, Harpagon]  \n",
       "23                                [Cléante, Harpagon]  \n",
       "24                               [La Flèche, Cléante]  \n",
       "25                                         [Harpagon]  \n",
       "26                         [Le Commissaire, Harpagon]  \n",
       "27         [Maître Jacques, Harpagon, Le Commissaire]  \n",
       "28                 [Harpagon, Valère, Maître Jacques]  \n",
       "29  [Harpagon, Valère, Élise, Maître Jacques, Fros...  \n",
       "30  [Anselme, Harpagon, Valère, Mariane, Maître Ja...  \n",
       "31  [Cléante, Harpagon, Mariane, Anselme, Le Commi...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def list_scene_speakers():\n",
    "    rows = []\n",
    "\n",
    "    for act in list_acts():\n",
    "        for scene in list_scenes(act=act[\"node\"]):\n",
    "            speakers, seen = [], set()\n",
    "\n",
    "            for sp in scene[\"node\"].findall(\".//x:p[@class='speaker']\", ns):\n",
    "                name = speaker_name(sp)\n",
    "\n",
    "                # On évite d'ajouter à la liste un acteur que l'on a déjà vu passer\n",
    "                if name and name not in seen:\n",
    "                    seen.add(name)\n",
    "                    speakers.append(name)\n",
    "\n",
    "            rows.append({\n",
    "                \"Acte\": act[\"title\"],\n",
    "                \"Scène\": scene[\"title\"],\n",
    "                \"Intervenants\": speakers,\n",
    "            })\n",
    "\n",
    "    return pd.DataFrame(rows)\n",
    "\n",
    "df_speakers = list_scene_speakers()\n",
    "df_speakers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Intervenant</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Anselme</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Brindavoine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Cléante</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Élise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Frosine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Harpagon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>La Flèche</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>La Merluche</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Le Commissaire</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Maître Jacques</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Maître simon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Mariane</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Valère</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Intervenant\n",
       "0          Anselme\n",
       "1      Brindavoine\n",
       "2          Cléante\n",
       "3            Élise\n",
       "4          Frosine\n",
       "5         Harpagon\n",
       "6        La Flèche\n",
       "7      La Merluche\n",
       "8   Le Commissaire\n",
       "9   Maître Jacques\n",
       "10    Maître simon\n",
       "11         Mariane\n",
       "12          Valère"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Tri localisé pour les intervenants et déduplication\n",
    "intervenants_uniques = sorted(\n",
    "    {name for names in df_speakers[\"Intervenants\"] for name in names},\n",
    "    key=locale.strxfrm\n",
    ")\n",
    "\n",
    "intervenants_df = pd.DataFrame({\"Intervenant\": intervenants_uniques})\n",
    "intervenants_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On peut désormais identifier les différences avec la _dramatis personae_, afin de vérifier l'uniformité des orthographes.\n",
    "Par corollaire, on pourra, en même temps, identifier les acteurs sans réplique."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Personnage (en-tête)</th>\n",
       "      <th>Intervenant (répliques)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Dame Claude</td>\n",
       "      <td>Le Commissaire</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Le commissaire</td>\n",
       "      <td>Maître Jacques</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Maitre Jacques</td>\n",
       "      <td>Maître simon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Maitre Simon</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Personnage (en-tête) Intervenant (répliques)\n",
       "0          Dame Claude          Le Commissaire\n",
       "1       Le commissaire          Maître Jacques\n",
       "2       Maitre Jacques            Maître simon\n",
       "3         Maitre Simon                    None"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acteurs_set = set(dramatis_personae[\"Personnage\"])\n",
    "intervenants_set = set(intervenants_uniques)\n",
    "\n",
    "# On écarte les noms exactement identiques\n",
    "communs = acteurs_set & intervenants_set\n",
    "acteurs_only = sorted(acteurs_set - communs, key=locale.strxfrm)\n",
    "intervenants_only = sorted(intervenants_set - communs, key=locale.strxfrm)\n",
    "\n",
    "df_diff = pd.DataFrame(\n",
    "    list(zip_longest(acteurs_only, intervenants_only)),\n",
    "    columns=[\"Personnage (en-tête)\", \"Intervenant (répliques)\"]\n",
    ")\n",
    "\n",
    "df_diff"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`None` indique simplement un remplissage par `zip_longest` pour que les deux listes aient la même taille."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On identifie bien deux orthographes différentes pour trois acteurs. La liste des personnages initiale omet les accents circonflexes de \"Maître\", \"commissaire\" est écrit en minuscule, et \"Simon\" a perdu sa majuscule."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Enfin, il est clair que Dame Claude n'a aucune réplique (puisqu'on ne la retrouve pas dans la liste des intervenants)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On peut donc créer une table de correspondance, associant un nom correctement orthographié avec les variantes que l'on peut trouver dans le texte initial. Nous utiliserons comme référence la graphie française correcte de \"maître\", et \"Commissaire\" avec une majuscule."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "alias_map = {\n",
    "    \"Maître Jacques\": {\"Maître Jacques\", \"Maitre Jacques\"},\n",
    "    \"Maître Simon\": {\"Maitre Simon\", \"Maître simon\"},\n",
    "    \"Le Commissaire\": {\"Le Commissaire\", \"Le commissaire\"},\n",
    "}\n",
    "\n",
    "alias_index = {alias: canon for canon, aliases in alias_map.items() for alias in aliases}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Quantité de parole par acteur\n",
    "\n",
    "Maintenant que nous disposons d'une liste uniformisée des noms des différents acteurs, nous pouvons analyser l'ensemble de la pièce et quantifier le texte prononcé par chaque acteur."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>Acte</th>\n",
       "      <th>Scène</th>\n",
       "      <th>Personnage</th>\n",
       "      <th>Mots</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>Cléante</td>\n",
       "      <td>127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>Harpagon</td>\n",
       "      <td>171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>La Flèche</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène II</td>\n",
       "      <td>Maître Simon</td>\n",
       "      <td>197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Acte II</td>\n",
       "      <td>Scène III</td>\n",
       "      <td>Frosine</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>Cléante</td>\n",
       "      <td>130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>Harpagon</td>\n",
       "      <td>89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>Le Commissaire</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>Mariane</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>Acte V</td>\n",
       "      <td>Scène VI</td>\n",
       "      <td>Maître Jacques</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>95 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Acte      Scène      Personnage  Mots\n",
       "0   Acte II   Scène II         Cléante   127\n",
       "1   Acte II   Scène II        Harpagon   171\n",
       "2   Acte II   Scène II       La Flèche    12\n",
       "3   Acte II   Scène II    Maître Simon   197\n",
       "4   Acte II  Scène III         Frosine     1\n",
       "..      ...        ...             ...   ...\n",
       "90   Acte V   Scène VI         Cléante   130\n",
       "91   Acte V   Scène VI        Harpagon    89\n",
       "92   Acte V   Scène VI  Le Commissaire    26\n",
       "93   Acte V   Scène VI         Mariane    36\n",
       "94   Acte V   Scène VI  Maître Jacques    23\n",
       "\n",
       "[95 rows x 4 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# TODO : Est-ce que cette fonction est redondante avec\n",
    "# les fonctions utilitaires créées précédemment ?\n",
    "def count_words_by_actor():\n",
    "    rows = []\n",
    "\n",
    "    for act in list_acts():\n",
    "        for scene in list_scenes(act=act[\"node\"]):\n",
    "            for order, speech in enumerate(scene_speeches(scene=scene[\"node\"])):\n",
    "                txt = speech[\"text\"]\n",
    "                rows.append({\n",
    "                    \"Acte\": act[\"title\"],\n",
    "                    \"Scène\": scene[\"title\"],\n",
    "                    \"Ordre\": order,\n",
    "                    \"Personnage\": speech[\"speaker\"],\n",
    "                    \"Texte\": txt,\n",
    "                    \"Mots\": speech[\"word_count\"],\n",
    "                    \"Lignes\": line_count(txt),\n",
    "                })\n",
    "\n",
    "    return pd.DataFrame(rows)\n",
    "\n",
    "df_speeches = count_words_by_actor()\n",
    "df_counts = df_speeches.groupby([\"Acte\", \"Scène\", \"Personnage\"], as_index=False)[\"Mots\"].sum()\n",
    "df_counts"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "Le comptage semble s'effectuer correctement, mais un tel tableau n'est pas digeste. \n",
    "On peut noter par exemple que \"Acte Premier\" est dilué au centre du tableau, en raison de la clause `groupby`, qui trie implicitement le tableau, et ignore donc notre tri initial.\n",
    "Nous pouvons toutefois ignorer ce détail ici, et regrouper par personnage : de cette manière, nous aurons un aperçu global du temps de parole de chacun à travers l'oeuvre."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Personnage le plus taciturne : Brindavoine (43 mots)\n",
      "Personnage le plus locace  : Harpagon (6132 mots)\n"
     ]
    }
   ],
   "source": [
    "global_df = df_speeches.groupby([\"Personnage\"])[\"Mots\"].sum()\n",
    "\n",
    "moins_bavard_nom = global_df.idxmin()\n",
    "moins_bavard_mots = global_df.min()\n",
    "\n",
    "plus_bavard_nom = global_df.idxmax()\n",
    "plus_bavard_mots = global_df.max()\n",
    "\n",
    "print(f\"Personnage le plus taciturne : {moins_bavard_nom} ({moins_bavard_mots} mots)\")\n",
    "print(f\"Personnage le plus locace  : {plus_bavard_nom} ({plus_bavard_mots} mots)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Notons que nous avons déjà établi que Dame Claude n'avait aucune réplique, et bien que le Commissaire soit accompagné d'un clerc, ce dernier ne parle jamais non plus."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Montrons la proportion de dialogues par personnage à travers un diagramme circulaire :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_small = 5\n",
    "y0, dy = 1.3, 0.05   # Positionnement du bloc des acteurs les moins locaces\n",
    "x_col = -0.5\n",
    "\n",
    "totaux = df_speeches.groupby(\"Personnage\")[\"Mots\"].sum().sort_values(ascending=False)\n",
    "labels = totaux.index\n",
    "values = totaux.values\n",
    "color_map = create_actors_colormap(labels)\n",
    "colors = [color_map[p] for p in labels]\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(7, 7))\n",
    "wedges, _ = ax.pie(values, colors=colors, startangle=90, counterclock=False, normalize=True)\n",
    "\n",
    "small_roles = global_df.sort_values().head(n_small)  # les moins bavards, ordre croissant\n",
    "\n",
    "# Placement de l'étiquette pour les petits rôles\n",
    "for i, (name, val) in enumerate(small_roles.items()):\n",
    "    idx = labels.get_loc(name)\n",
    "    w = wedges[idx]\n",
    "\n",
    "    theta = np.deg2rad((w.theta1 + w.theta2) / 2)\n",
    "    xw, yw = np.cos(theta), np.sin(theta)\n",
    "    xpos, ypos = x_col, y0 - i * dy\n",
    "\n",
    "    ax.annotate(\n",
    "        f\"{name} ({val})\",\n",
    "        xy=(xw, yw), xytext=(xpos, ypos),\n",
    "        ha=\"right\", va=\"center\", fontsize=9,\n",
    "        arrowprops=dict(\n",
    "            arrowstyle=\"-\",\n",
    "            color=colors[idx],\n",
    "            lw=1,\n",
    "            connectionstyle=\"angle,angleA=0,angleB=90\",\n",
    "            shrinkA=0, shrinkB=0,\n",
    "        ),\n",
    "    )\n",
    "\n",
    "# Autres rôles : nom autour + valeur au centre\n",
    "for idx, name in enumerate(labels):\n",
    "    if name in small_roles.index:\n",
    "        continue\n",
    "\n",
    "    w = wedges[idx]\n",
    "    theta = np.deg2rad((w.theta1 + w.theta2) / 2)\n",
    "    r_label, r_value = 1.1, 0.7\n",
    "\n",
    "    ax.text(r_label * np.cos(theta), r_label * np.sin(theta), name, ha=\"center\", va=\"center\", fontsize=9)\n",
    "    ax.text(r_value * np.cos(theta), r_value * np.sin(theta), str(totaux[name]), ha=\"center\", va=\"center\", fontsize=9, color=\"black\")\n",
    "\n",
    "ax.set_title(\"Répartition des mots prononcés par personnage\", pad=50)\n",
    "ax.axis(\"equal\")\n",
    "plt.tight_layout()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Statistiques complémentaires\n",
    "\n",
    "Inspirées des tableaux de l'OBVIL, nous examinons la place de chaque personnage et les relations directes entre interlocuteurs (une ligne = 60 caractères).\n",
    "Commençons par la \"Table des rôles\" :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "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>Rôle</th>\n",
       "      <th>Scènes</th>\n",
       "      <th>Répl.</th>\n",
       "      <th>Répl. moy.</th>\n",
       "      <th>Présence</th>\n",
       "      <th>Texte</th>\n",
       "      <th>Texte % prés.</th>\n",
       "      <th>Texte × pers.</th>\n",
       "      <th>Interlocution</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[TOUS]</td>\n",
       "      <td>32 sc.</td>\n",
       "      <td>959 répl.</td>\n",
       "      <td>1,8 l.</td>\n",
       "      <td>1 769 l. (100 %)</td>\n",
       "      <td>1 769 l. (100 %)</td>\n",
       "      <td>100 %</td>\n",
       "      <td>5 847 l. (100 %)</td>\n",
       "      <td>3,3 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Harpagon</td>\n",
       "      <td>23 sc.</td>\n",
       "      <td>354 répl.</td>\n",
       "      <td>1,5 l.</td>\n",
       "      <td>1 296 l. (73 %)</td>\n",
       "      <td>514 l. (29 %)</td>\n",
       "      <td>40 %</td>\n",
       "      <td>4 729 l. (81 %)</td>\n",
       "      <td>3,7 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Cléante</td>\n",
       "      <td>14 sc.</td>\n",
       "      <td>161 répl.</td>\n",
       "      <td>1,8 l.</td>\n",
       "      <td>900 l. (51 %)</td>\n",
       "      <td>285 l. (16 %)</td>\n",
       "      <td>32 %</td>\n",
       "      <td>3 486 l. (60 %)</td>\n",
       "      <td>3,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Élise</td>\n",
       "      <td>9 sc.</td>\n",
       "      <td>51 répl.</td>\n",
       "      <td>1,8 l.</td>\n",
       "      <td>681 l. (39 %)</td>\n",
       "      <td>92 l. (5 %)</td>\n",
       "      <td>13 %</td>\n",
       "      <td>2 667 l. (46 %)</td>\n",
       "      <td>3,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Valère</td>\n",
       "      <td>9 sc.</td>\n",
       "      <td>101 répl.</td>\n",
       "      <td>2,3 l.</td>\n",
       "      <td>695 l. (39 %)</td>\n",
       "      <td>232 l. (13 %)</td>\n",
       "      <td>33 %</td>\n",
       "      <td>3 067 l. (52 %)</td>\n",
       "      <td>4,4 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Mariane</td>\n",
       "      <td>6 sc.</td>\n",
       "      <td>31 répl.</td>\n",
       "      <td>2,5 l.</td>\n",
       "      <td>359 l. (20 %)</td>\n",
       "      <td>79 l. (4 %)</td>\n",
       "      <td>22 %</td>\n",
       "      <td>1 638 l. (28 %)</td>\n",
       "      <td>4,6 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Anselme</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>20 répl.</td>\n",
       "      <td>2,3 l.</td>\n",
       "      <td>143 l. (8 %)</td>\n",
       "      <td>45 l. (3 %)</td>\n",
       "      <td>32 %</td>\n",
       "      <td>749 l. (13 %)</td>\n",
       "      <td>5,3 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Frosine</td>\n",
       "      <td>10 sc.</td>\n",
       "      <td>60 répl.</td>\n",
       "      <td>3,3 l.</td>\n",
       "      <td>466 l. (26 %)</td>\n",
       "      <td>201 l. (11 %)</td>\n",
       "      <td>43 %</td>\n",
       "      <td>1 465 l. (25 %)</td>\n",
       "      <td>3,1 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Maître Simon</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>5 répl.</td>\n",
       "      <td>3,2 l.</td>\n",
       "      <td>44 l. (2 %)</td>\n",
       "      <td>16 l. (1 %)</td>\n",
       "      <td>37 %</td>\n",
       "      <td>175 l. (3 %)</td>\n",
       "      <td>4,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Maître Jacques</td>\n",
       "      <td>9 sc.</td>\n",
       "      <td>85 répl.</td>\n",
       "      <td>1,6 l.</td>\n",
       "      <td>557 l. (32 %)</td>\n",
       "      <td>140 l. (8 %)</td>\n",
       "      <td>25 %</td>\n",
       "      <td>2 670 l. (46 %)</td>\n",
       "      <td>4,8 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>La Flèche</td>\n",
       "      <td>5 sc.</td>\n",
       "      <td>66 répl.</td>\n",
       "      <td>2,0 l.</td>\n",
       "      <td>255 l. (14 %)</td>\n",
       "      <td>132 l. (7 %)</td>\n",
       "      <td>52 %</td>\n",
       "      <td>598 l. (10 %)</td>\n",
       "      <td>2,3 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Dame Claude</td>\n",
       "      <td>0 sc.</td>\n",
       "      <td>0 répl.</td>\n",
       "      <td>0,0 l.</td>\n",
       "      <td>0 l.</td>\n",
       "      <td>0 l.</td>\n",
       "      <td>0 %</td>\n",
       "      <td>0 l. (0 %)</td>\n",
       "      <td>0,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Brindavoine</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>3 répl.</td>\n",
       "      <td>1,1 l.</td>\n",
       "      <td>166 l. (9 %)</td>\n",
       "      <td>3 l. (0 %)</td>\n",
       "      <td>2 %</td>\n",
       "      <td>1 146 l. (20 %)</td>\n",
       "      <td>6,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>La Merluche</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>5 répl.</td>\n",
       "      <td>0,9 l.</td>\n",
       "      <td>175 l. (10 %)</td>\n",
       "      <td>5 l. (0 %)</td>\n",
       "      <td>3 %</td>\n",
       "      <td>1 189 l. (20 %)</td>\n",
       "      <td>6,8 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Le Commissaire</td>\n",
       "      <td>3 sc.</td>\n",
       "      <td>17 répl.</td>\n",
       "      <td>1,5 l.</td>\n",
       "      <td>110 l. (6 %)</td>\n",
       "      <td>26 l. (1 %)</td>\n",
       "      <td>24 %</td>\n",
       "      <td>418 l. (7 %)</td>\n",
       "      <td>3,8 pers.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Rôle  Scènes      Répl. Répl. moy.          Présence  \\\n",
       "0           [TOUS]  32 sc.  959 répl.     1,8 l.  1 769 l. (100 %)   \n",
       "1         Harpagon  23 sc.  354 répl.     1,5 l.   1 296 l. (73 %)   \n",
       "2          Cléante  14 sc.  161 répl.     1,8 l.     900 l. (51 %)   \n",
       "3            Élise   9 sc.   51 répl.     1,8 l.     681 l. (39 %)   \n",
       "4           Valère   9 sc.  101 répl.     2,3 l.     695 l. (39 %)   \n",
       "5          Mariane   6 sc.   31 répl.     2,5 l.     359 l. (20 %)   \n",
       "6          Anselme   2 sc.   20 répl.     2,3 l.      143 l. (8 %)   \n",
       "7          Frosine  10 sc.   60 répl.     3,3 l.     466 l. (26 %)   \n",
       "8     Maître Simon   1 sc.    5 répl.     3,2 l.       44 l. (2 %)   \n",
       "9   Maître Jacques   9 sc.   85 répl.     1,6 l.     557 l. (32 %)   \n",
       "10       La Flèche   5 sc.   66 répl.     2,0 l.     255 l. (14 %)   \n",
       "11     Dame Claude   0 sc.    0 répl.     0,0 l.              0 l.   \n",
       "12     Brindavoine   2 sc.    3 répl.     1,1 l.      166 l. (9 %)   \n",
       "13     La Merluche   2 sc.    5 répl.     0,9 l.     175 l. (10 %)   \n",
       "14  Le Commissaire   3 sc.   17 répl.     1,5 l.      110 l. (6 %)   \n",
       "\n",
       "               Texte Texte % prés.     Texte × pers. Interlocution  \n",
       "0   1 769 l. (100 %)         100 %  5 847 l. (100 %)     3,3 pers.  \n",
       "1      514 l. (29 %)          40 %   4 729 l. (81 %)     3,7 pers.  \n",
       "2      285 l. (16 %)          32 %   3 486 l. (60 %)     3,9 pers.  \n",
       "3        92 l. (5 %)          13 %   2 667 l. (46 %)     3,9 pers.  \n",
       "4      232 l. (13 %)          33 %   3 067 l. (52 %)     4,4 pers.  \n",
       "5        79 l. (4 %)          22 %   1 638 l. (28 %)     4,6 pers.  \n",
       "6        45 l. (3 %)          32 %     749 l. (13 %)     5,3 pers.  \n",
       "7      201 l. (11 %)          43 %   1 465 l. (25 %)     3,1 pers.  \n",
       "8        16 l. (1 %)          37 %      175 l. (3 %)     4,0 pers.  \n",
       "9       140 l. (8 %)          25 %   2 670 l. (46 %)     4,8 pers.  \n",
       "10      132 l. (7 %)          52 %     598 l. (10 %)     2,3 pers.  \n",
       "11              0 l.           0 %        0 l. (0 %)     0,0 pers.  \n",
       "12        3 l. (0 %)           2 %   1 146 l. (20 %)     6,9 pers.  \n",
       "13        5 l. (0 %)           3 %   1 189 l. (20 %)     6,8 pers.  \n",
       "14       26 l. (1 %)          24 %      418 l. (7 %)     3,8 pers.  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Agrégats au niveau des scènes\n",
    "scene_totals = (df_speeches.groupby([\"Acte\", \"Scène\"], as_index=False)\n",
    "                .agg(scene_lines=(\"Lignes\", \"sum\"),\n",
    "                     participants=(\"Personnage\", \"nunique\")))\n",
    "scene_totals[\"textexpers\"] = scene_totals[\"scene_lines\"] * scene_totals[\"participants\"]\n",
    "scene_totals[\"SceneKey\"] = scene_totals[\"Acte\"] + \" | \" + scene_totals[\"Scène\"]\n",
    "\n",
    "total_lines_play = scene_totals[\"scene_lines\"].sum()\n",
    "textexpers_total = scene_totals[\"textexpers\"].sum()\n",
    "\n",
    "# Statistiques par personnage\n",
    "speech_stats = df_speeches.groupby(\"Personnage\").agg(\n",
    "    repl=(\"Texte\", \"count\"),\n",
    "    text_lines=(\"Lignes\", \"sum\"),\n",
    ")\n",
    "\n",
    "presence = (df_speeches[[\"Acte\", \"Scène\", \"Personnage\"]]\n",
    "            .drop_duplicates()\n",
    "            .merge(scene_totals[[\"Acte\", \"Scène\", \"scene_lines\", \"textexpers\"]],\n",
    "                   on=[\"Acte\", \"Scène\"], how=\"left\"))\n",
    "\n",
    "presence_stats = presence.groupby(\"Personnage\").agg(\n",
    "    scenes=(\"Scène\", \"count\"),\n",
    "    presence_lines=(\"scene_lines\", \"sum\"),\n",
    "    textexpers=(\"textexpers\", \"sum\"),\n",
    ")\n",
    "\n",
    "roles = speech_stats.join(presence_stats, how=\"outer\").fillna(0)\n",
    "\n",
    "# Ajout des rôles muets (ex : Dame Claude) pour qu'ils apparaissent dans le tableau\n",
    "for name in (resolve_name(p) for p in dramatis_personae[\"Personnage\"]):\n",
    "    if name not in roles.index:\n",
    "        roles.loc[name] = 0\n",
    "\n",
    "roles[\"repl_moy\"] = roles[\"text_lines\"] / roles[\"repl\"]\n",
    "roles[\"presence_pct\"] = roles[\"presence_lines\"] / total_lines_play\n",
    "roles[\"text_pct\"] = roles[\"text_lines\"] / total_lines_play\n",
    "roles[\"text_presence_pct\"] = roles[\"text_lines\"] / roles[\"presence_lines\"]\n",
    "roles[\"textexpers_pct\"] = roles[\"textexpers\"] / textexpers_total\n",
    "roles[\"interlocution\"] = roles[\"textexpers\"] / roles[\"presence_lines\"]\n",
    "\n",
    "roles = roles.replace([np.inf, -np.inf], 0).fillna(0)\n",
    "\n",
    "# Ordre basé sur la distribution initiale\n",
    "role_order = [name for name in (resolve_name(p) for p in dramatis_personae[\"Personnage\"]) if name in roles.index]\n",
    "role_order += [r for r in roles.index if r not in role_order]\n",
    "\n",
    "roles = roles.loc[role_order]\n",
    "\n",
    "# Ligne globale\n",
    "all_row = pd.Series({\n",
    "    \"scenes\": scene_totals.shape[0],\n",
    "    \"repl\": len(df_speeches),\n",
    "    \"repl_moy\": df_speeches[\"Lignes\"].sum() / len(df_speeches),\n",
    "    \"presence_lines\": total_lines_play,\n",
    "    \"presence_pct\": 1.0,\n",
    "    \"text_lines\": df_speeches[\"Lignes\"].sum(),\n",
    "    \"text_pct\": 1.0,\n",
    "    \"text_presence_pct\": df_speeches[\"Lignes\"].sum() / total_lines_play if total_lines_play else 0,\n",
    "    \"textexpers\": textexpers_total,\n",
    "    \"textexpers_pct\": 1.0,\n",
    "    \"interlocution\": textexpers_total / total_lines_play if total_lines_play else 0,\n",
    "})\n",
    "\n",
    "roles = pd.concat([\n",
    "    pd.DataFrame({\"Personnage\": [\"[TOUS]\"]}).set_index(\"Personnage\").assign(**all_row),\n",
    "    roles\n",
    "])\n",
    "\n",
    "roles.index.name = \"Rôle\"\n",
    "\n",
    "# Quelques utilitaires spécifiques\n",
    "\n",
    "def format_number(value, decimals=0):\n",
    "    fmt = f\"{value:,.{decimals}f}\"\n",
    "    return fmt.replace(\",\", \" \").replace(\".\", \",\")\n",
    "def format_lines(value, decimals=0):\n",
    "    return f\"{format_number(round(value, decimals), decimals)} l.\"\n",
    "def format_percent(value, decimals=0):\n",
    "    return f\"{format_number(value * 100, decimals)} %\"\n",
    "def format_people(value):\n",
    "    return f\"{format_number(value, 1)} pers.\"\n",
    "\n",
    "roles_display = roles.reset_index()\n",
    "roles_display[\"Scènes\"] = roles_display[\"scenes\"].fillna(0).astype(int).astype(str) + \" sc.\"\n",
    "roles_display[\"Répl.\"] = roles_display[\"repl\"].fillna(0).astype(int).astype(str) + \" répl.\"\n",
    "roles_display[\"Répl. moy.\"] = roles_display[\"repl_moy\"].apply(lambda v: format_lines(v, 1))\n",
    "roles_display[\"Présence\"] = roles_display.apply(\n",
    "    lambda r: f\"{format_lines(r['presence_lines'])}\" + (f\" ({format_percent(r['presence_pct'])})\" if r[\"presence_lines\"] else \"\"),\n",
    "    axis=1,\n",
    ")\n",
    "roles_display[\"Texte\"] = roles_display.apply(\n",
    "    lambda r: f\"{format_lines(r['text_lines'])}\" + (f\" ({format_percent(r['text_pct'])})\" if r[\"text_lines\"] else \"\"),\n",
    "    axis=1,\n",
    ")\n",
    "roles_display[\"Texte % prés.\"] = roles_display[\"text_presence_pct\"].apply(lambda v: format_percent(v, 0))\n",
    "roles_display[\"Texte × pers.\"] = roles_display.apply(\n",
    "    lambda r: f\"{format_lines(r['textexpers'])} ({format_percent(r['textexpers_pct'])})\",\n",
    "    axis=1,\n",
    ")\n",
    "roles_display[\"Interlocution\"] = roles_display[\"interlocution\"].apply(format_people)\n",
    "\n",
    "roles_table = roles_display[[\"Rôle\", \"Scènes\", \"Répl.\", \"Répl. moy.\", \"Présence\", \"Texte\", \"Texte % prés.\", \"Texte × pers.\", \"Interlocution\"]]\n",
    "roles_table\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Il existe des différences entre le tableau que nous avons généré et celui de l'OBVIL.\n",
    "Ces différences peuvent être attribuées à :\n",
    "\n",
    "- une méthode de nettoyage des lignes différente (nous avons opté pour un nettoyage agressif des espaces surnuméraires)\n",
    "- une gestion des décimales différentes (est-ce que l'OBVIL arrondi à l'entier supérieur ou inférieur, ou tronque les décimales ?)\n",
    "\n",
    "Ces différences affectent mathématiquement les statistiques qui découlent de ce comptage, notamment l'interlocution."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Statistiques par relation\n",
    "\n",
    "Chaque relation s'appuie sur l'enchaînement de répliques adjacentes entre deux personnages (monologues inclus), ce qui reflète les échanges directs plutôt que la simple coprésence sur scène.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Relation</th>\n",
       "      <th>Détail</th>\n",
       "      <th>Scènes</th>\n",
       "      <th>Texte</th>\n",
       "      <th>Interlocution</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Harpagon</td>\n",
       "      <td>33 l. (100 %) 1 répl. 32,7 l.</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>33 l. (2 %)</td>\n",
       "      <td>1,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Cléante / Harpagon</td>\n",
       "      <td>135 l. (49 %) 97 répl. 1,4 l. - 140 l. (51 %) ...</td>\n",
       "      <td>9 sc.</td>\n",
       "      <td>275 l. (16 %)</td>\n",
       "      <td>4,6 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Harpagon / Élise</td>\n",
       "      <td>41 l. (60 %) 28 répl. 1,5 l. - 28 l. (40 %) 28...</td>\n",
       "      <td>6 sc.</td>\n",
       "      <td>69 l. (4 %)</td>\n",
       "      <td>4,7 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Harpagon / Valère</td>\n",
       "      <td>96 l. (44 %) 65 répl. 1,5 l. - 121 l. (56 %) 5...</td>\n",
       "      <td>7 sc.</td>\n",
       "      <td>217 l. (12 %)</td>\n",
       "      <td>4,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Harpagon / Mariane</td>\n",
       "      <td>9 l. (39 %) 7 répl. 1,2 l. - 13 l. (61 %) 5 ré...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>22 l. (1 %)</td>\n",
       "      <td>4,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Anselme / Harpagon</td>\n",
       "      <td>25 l. (71 %) 11 répl. 2,2 l. - 10 l. (29 %) 8 ...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>35 l. (2 %)</td>\n",
       "      <td>5,3 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Frosine / Harpagon</td>\n",
       "      <td>128 l. (70 %) 39 répl. 3,3 l. - 56 l. (30 %) 3...</td>\n",
       "      <td>5 sc.</td>\n",
       "      <td>184 l. (10 %)</td>\n",
       "      <td>3,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Harpagon / Maître Simon</td>\n",
       "      <td>5 l. (22 %) 3 répl. 1,5 l. - 16 l. (78 %) 4 ré...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>20 l. (1 %)</td>\n",
       "      <td>4,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Harpagon / Maître Jacques</td>\n",
       "      <td>58 l. (38 %) 49 répl. 1,2 l. - 94 l. (62 %) 51...</td>\n",
       "      <td>7 sc.</td>\n",
       "      <td>153 l. (9 %)</td>\n",
       "      <td>4,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Harpagon / La Flèche</td>\n",
       "      <td>35 l. (62 %) 33 répl. 1,1 l. - 22 l. (38 %) 33...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>57 l. (3 %)</td>\n",
       "      <td>2,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Brindavoine / Harpagon</td>\n",
       "      <td>1 l. (38 %) 2 répl. 0,7 l. - 2 l. (62 %) 2 rép...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>4 l. (0 %)</td>\n",
       "      <td>6,9 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Harpagon / La Merluche</td>\n",
       "      <td>1 l. (11 %) 1 répl. 0,6 l. - 5 l. (89 %) 5 rép...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>5 l. (0 %)</td>\n",
       "      <td>6,8 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Harpagon / Le Commissaire</td>\n",
       "      <td>11 l. (53 %) 10 répl. 1,1 l. - 10 l. (47 %) 9 ...</td>\n",
       "      <td>3 sc.</td>\n",
       "      <td>21 l. (1 %)</td>\n",
       "      <td>3,8 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Cléante / Élise</td>\n",
       "      <td>66 l. (84 %) 10 répl. 6,6 l. - 13 l. (16 %) 9 ...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>78 l. (4 %)</td>\n",
       "      <td>3,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Cléante / Mariane</td>\n",
       "      <td>31 l. (60 %) 12 répl. 2,6 l. - 21 l. (40 %) 10...</td>\n",
       "      <td>3 sc.</td>\n",
       "      <td>52 l. (3 %)</td>\n",
       "      <td>4,8 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Cléante / Frosine</td>\n",
       "      <td>3 l. (8 %) 5 répl. 0,6 l. - 37 l. (92 %) 6 rép...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>40 l. (2 %)</td>\n",
       "      <td>4,5 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Cléante / Maître Jacques</td>\n",
       "      <td>14 l. (51 %) 8 répl. 1,8 l. - 14 l. (49 %) 8 r...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>28 l. (2 %)</td>\n",
       "      <td>3,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Cléante / La Flèche</td>\n",
       "      <td>31 l. (27 %) 25 répl. 1,2 l. - 85 l. (73 %) 26...</td>\n",
       "      <td>3 sc.</td>\n",
       "      <td>116 l. (7 %)</td>\n",
       "      <td>2,5 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Valère / Élise</td>\n",
       "      <td>65 l. (59 %) 11 répl. 5,9 l. - 45 l. (41 %) 11...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>110 l. (6 %)</td>\n",
       "      <td>2,5 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Mariane / Élise</td>\n",
       "      <td>1 l. (23 %) 2 répl. 0,6 l. - 4 l. (77 %) 1 rép...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>6 l. (0 %)</td>\n",
       "      <td>4,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Mariane / Valère</td>\n",
       "      <td>3 l. (41 %) 1 répl. 2,6 l. - 4 l. (59 %) 3 rép...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>6 l. (0 %)</td>\n",
       "      <td>5,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Anselme / Valère</td>\n",
       "      <td>19 l. (47 %) 8 répl. 2,4 l. - 22 l. (53 %) 7 r...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>41 l. (2 %)</td>\n",
       "      <td>5,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Maître Jacques / Valère</td>\n",
       "      <td>18 l. (49 %) 14 répl. 1,3 l. - 19 l. (51 %) 18...</td>\n",
       "      <td>4 sc.</td>\n",
       "      <td>37 l. (2 %)</td>\n",
       "      <td>5,2 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Frosine / Mariane</td>\n",
       "      <td>18 l. (48 %) 6 répl. 3,0 l. - 20 l. (52 %) 7 r...</td>\n",
       "      <td>4 sc.</td>\n",
       "      <td>38 l. (2 %)</td>\n",
       "      <td>4,1 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Frosine / Maître Jacques</td>\n",
       "      <td>1 l. (42 %) 1 répl. 1,0 l. - 1 l. (58 %) 2 rép...</td>\n",
       "      <td>2 sc.</td>\n",
       "      <td>2 l. (0 %)</td>\n",
       "      <td>4,7 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Frosine / La Flèche</td>\n",
       "      <td>11 l. (42 %) 5 répl. 2,3 l. - 16 l. (58 %) 5 r...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>28 l. (2 %)</td>\n",
       "      <td>2,0 pers.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Le Commissaire / Maître Jacques</td>\n",
       "      <td>13 l. (76 %) 7 répl. 1,8 l. - 4 l. (24 %) 5 ré...</td>\n",
       "      <td>1 sc.</td>\n",
       "      <td>17 l. (1 %)</td>\n",
       "      <td>3,0 pers.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Relation  \\\n",
       "0                          Harpagon   \n",
       "1                Cléante / Harpagon   \n",
       "2                  Harpagon / Élise   \n",
       "3                 Harpagon / Valère   \n",
       "4                Harpagon / Mariane   \n",
       "5                Anselme / Harpagon   \n",
       "6                Frosine / Harpagon   \n",
       "7           Harpagon / Maître Simon   \n",
       "8         Harpagon / Maître Jacques   \n",
       "9              Harpagon / La Flèche   \n",
       "10           Brindavoine / Harpagon   \n",
       "11           Harpagon / La Merluche   \n",
       "12        Harpagon / Le Commissaire   \n",
       "13                  Cléante / Élise   \n",
       "14                Cléante / Mariane   \n",
       "15                Cléante / Frosine   \n",
       "16         Cléante / Maître Jacques   \n",
       "17              Cléante / La Flèche   \n",
       "18                   Valère / Élise   \n",
       "19                  Mariane / Élise   \n",
       "20                 Mariane / Valère   \n",
       "21                 Anselme / Valère   \n",
       "22          Maître Jacques / Valère   \n",
       "23                Frosine / Mariane   \n",
       "24         Frosine / Maître Jacques   \n",
       "25              Frosine / La Flèche   \n",
       "26  Le Commissaire / Maître Jacques   \n",
       "\n",
       "                                               Détail Scènes          Texte  \\\n",
       "0                       33 l. (100 %) 1 répl. 32,7 l.  1 sc.    33 l. (2 %)   \n",
       "1   135 l. (49 %) 97 répl. 1,4 l. - 140 l. (51 %) ...  9 sc.  275 l. (16 %)   \n",
       "2   41 l. (60 %) 28 répl. 1,5 l. - 28 l. (40 %) 28...  6 sc.    69 l. (4 %)   \n",
       "3   96 l. (44 %) 65 répl. 1,5 l. - 121 l. (56 %) 5...  7 sc.  217 l. (12 %)   \n",
       "4   9 l. (39 %) 7 répl. 1,2 l. - 13 l. (61 %) 5 ré...  2 sc.    22 l. (1 %)   \n",
       "5   25 l. (71 %) 11 répl. 2,2 l. - 10 l. (29 %) 8 ...  2 sc.    35 l. (2 %)   \n",
       "6   128 l. (70 %) 39 répl. 3,3 l. - 56 l. (30 %) 3...  5 sc.  184 l. (10 %)   \n",
       "7   5 l. (22 %) 3 répl. 1,5 l. - 16 l. (78 %) 4 ré...  1 sc.    20 l. (1 %)   \n",
       "8   58 l. (38 %) 49 répl. 1,2 l. - 94 l. (62 %) 51...  7 sc.   153 l. (9 %)   \n",
       "9   35 l. (62 %) 33 répl. 1,1 l. - 22 l. (38 %) 33...  1 sc.    57 l. (3 %)   \n",
       "10  1 l. (38 %) 2 répl. 0,7 l. - 2 l. (62 %) 2 rép...  2 sc.     4 l. (0 %)   \n",
       "11  1 l. (11 %) 1 répl. 0,6 l. - 5 l. (89 %) 5 rép...  2 sc.     5 l. (0 %)   \n",
       "12  11 l. (53 %) 10 répl. 1,1 l. - 10 l. (47 %) 9 ...  3 sc.    21 l. (1 %)   \n",
       "13  66 l. (84 %) 10 répl. 6,6 l. - 13 l. (16 %) 9 ...  2 sc.    78 l. (4 %)   \n",
       "14  31 l. (60 %) 12 répl. 2,6 l. - 21 l. (40 %) 10...  3 sc.    52 l. (3 %)   \n",
       "15  3 l. (8 %) 5 répl. 0,6 l. - 37 l. (92 %) 6 rép...  2 sc.    40 l. (2 %)   \n",
       "16  14 l. (51 %) 8 répl. 1,8 l. - 14 l. (49 %) 8 r...  1 sc.    28 l. (2 %)   \n",
       "17  31 l. (27 %) 25 répl. 1,2 l. - 85 l. (73 %) 26...  3 sc.   116 l. (7 %)   \n",
       "18  65 l. (59 %) 11 répl. 5,9 l. - 45 l. (41 %) 11...  2 sc.   110 l. (6 %)   \n",
       "19  1 l. (23 %) 2 répl. 0,6 l. - 4 l. (77 %) 1 rép...  2 sc.     6 l. (0 %)   \n",
       "20  3 l. (41 %) 1 répl. 2,6 l. - 4 l. (59 %) 3 rép...  1 sc.     6 l. (0 %)   \n",
       "21  19 l. (47 %) 8 répl. 2,4 l. - 22 l. (53 %) 7 r...  1 sc.    41 l. (2 %)   \n",
       "22  18 l. (49 %) 14 répl. 1,3 l. - 19 l. (51 %) 18...  4 sc.    37 l. (2 %)   \n",
       "23  18 l. (48 %) 6 répl. 3,0 l. - 20 l. (52 %) 7 r...  4 sc.    38 l. (2 %)   \n",
       "24  1 l. (42 %) 1 répl. 1,0 l. - 1 l. (58 %) 2 rép...  2 sc.     2 l. (0 %)   \n",
       "25  11 l. (42 %) 5 répl. 2,3 l. - 16 l. (58 %) 5 r...  1 sc.    28 l. (2 %)   \n",
       "26  13 l. (76 %) 7 répl. 1,8 l. - 4 l. (24 %) 5 ré...  1 sc.    17 l. (1 %)   \n",
       "\n",
       "   Interlocution  \n",
       "0      1,0 pers.  \n",
       "1      4,6 pers.  \n",
       "2      4,7 pers.  \n",
       "3      4,9 pers.  \n",
       "4      4,9 pers.  \n",
       "5      5,3 pers.  \n",
       "6      3,0 pers.  \n",
       "7      4,0 pers.  \n",
       "8      4,9 pers.  \n",
       "9      2,0 pers.  \n",
       "10     6,9 pers.  \n",
       "11     6,8 pers.  \n",
       "12     3,8 pers.  \n",
       "13     3,0 pers.  \n",
       "14     4,8 pers.  \n",
       "15     4,5 pers.  \n",
       "16     3,0 pers.  \n",
       "17     2,5 pers.  \n",
       "18     2,5 pers.  \n",
       "19     4,0 pers.  \n",
       "20     5,0 pers.  \n",
       "21     5,0 pers.  \n",
       "22     5,2 pers.  \n",
       "23     4,1 pers.  \n",
       "24     4,7 pers.  \n",
       "25     2,0 pers.  \n",
       "26     3,0 pers.  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from collections import defaultdict\n",
    "\n",
    "# Accumulateur : pour chaque relation (acteur seul - monologues) ou (acteurA, acteurB),\n",
    "# on stocke les lignes, le nombre de répliques, les scènes concernées,\n",
    "# les lignes de présence cumulées et un poids d’interlocution (lignes × nb de participants).\n",
    "relation_stats = defaultdict(lambda: {\n",
    "    \"lines\": defaultdict(float),\n",
    "    \"counts\": defaultdict(int),\n",
    "    \"scenes\": set(),\n",
    "    \"presence_lines\": 0.0,\n",
    "    \"interlocution_weight\": 0.0,\n",
    "})\n",
    "\n",
    "scene_lookup = scene_totals.set_index([\"Acte\", \"Scène\"])[[\"scene_lines\", \"participants\"]]\n",
    "\n",
    "for (act, scene), scene_df in df_speeches.groupby([\"Acte\", \"Scène\"]):\n",
    "    scene_df = scene_df.sort_values(\"Ordre\")\n",
    "    speakers = scene_df[\"Personnage\"].tolist()\n",
    "    lines = scene_df[\"Lignes\"].tolist()\n",
    "    scene_lines = scene_lookup.loc[(act, scene), \"scene_lines\"]\n",
    "    participants = scene_lookup.loc[(act, scene), \"participants\"]\n",
    "    scene_key = f\"{act} | {scene}\"\n",
    "\n",
    "    # Scène à un seul intervenant : on enregistre un monologue\n",
    "    if len(set(speakers)) == 1:\n",
    "        actor = speakers[0]\n",
    "        stats = relation_stats[(actor,)]\n",
    "        stats[\"lines\"][actor] += scene_lines\n",
    "        stats[\"counts\"][actor] += len(speakers)\n",
    "        stats[\"scenes\"].add(scene_key)\n",
    "        stats[\"presence_lines\"] += scene_lines\n",
    "        stats[\"interlocution_weight\"] += scene_lines * participants\n",
    "        continue\n",
    "\n",
    "    relations_here = set()\n",
    "    \n",
    "    # Pour chaque changement d’intervenant, on attribue les lignes du locuteur au duo (ordre ignoré)\n",
    "    for speaker, next_speaker, speaker_lines in zip(speakers, speakers[1:], lines):\n",
    "        if speaker == next_speaker:\n",
    "            continue\n",
    "            \n",
    "        key = tuple(sorted((speaker, next_speaker)))\n",
    "        stats = relation_stats[key]\n",
    "        stats[\"lines\"][speaker] += speaker_lines\n",
    "        stats[\"counts\"][speaker] += 1\n",
    "        stats[\"scenes\"].add(scene_key)\n",
    "        relations_here.add(key)\n",
    "\n",
    "    # On ajoute la présence et l’interlocution une seule fois par scène et par relation\n",
    "    for key in relations_here:\n",
    "        stats = relation_stats[key]\n",
    "        stats[\"presence_lines\"] += scene_lines\n",
    "        stats[\"interlocution_weight\"] += scene_lines * participants\n",
    "\n",
    "role_order_index = {name: idx for idx, name in enumerate(role_order)}\n",
    "\n",
    "def relation_sort_key(rel):\n",
    "    if len(rel) == 1:\n",
    "        return (role_order_index.get(rel[0], len(role_order_index)), -1)\n",
    "    \n",
    "    a, b = rel\n",
    "    \n",
    "    return (\n",
    "        min(role_order_index.get(a, len(role_order_index)), role_order_index.get(b, len(role_order_index))),\n",
    "        max(role_order_index.get(a, len(role_order_index)), role_order_index.get(b, len(role_order_index))),\n",
    "    )\n",
    "\n",
    "relation_rows = []\n",
    "\n",
    "for rel in sorted(relation_stats, key=relation_sort_key):\n",
    "    data = relation_stats[rel]\n",
    "    total_lines = sum(data[\"lines\"].values())\n",
    "\n",
    "    # On ignore les relations sans matière (moins de 2 lignes au total)\n",
    "    if total_lines < 2:\n",
    "        continue\n",
    "        \n",
    "    # On ignore les relations où au moins un protagoniste n’a jamais prononcé de réplique dans ce duo\n",
    "    if len(rel) > 1 and any(data[\"counts\"].get(actor, 0) == 0 for actor in rel):\n",
    "        continue\n",
    "\n",
    "    scenes_count = len(data[\"scenes\"])\n",
    "    interlocution = data[\"interlocution_weight\"] / data[\"presence_lines\"] if data[\"presence_lines\"] else 0\n",
    "\n",
    "    parts = []\n",
    "    \n",
    "    for actor in rel:\n",
    "        actor_lines = data[\"lines\"].get(actor, 0)\n",
    "        actor_repl = data[\"counts\"].get(actor, 0)\n",
    "        avg_lines = actor_lines / actor_repl if actor_repl else 0\n",
    "        share = actor_lines / total_lines if total_lines else 0\n",
    "        \n",
    "        parts.append(\n",
    "            f\"{format_lines(actor_lines)} ({format_percent(share)}) {actor_repl} répl. {format_lines(avg_lines, 1)}\"\n",
    "        )\n",
    "\n",
    "    relation_rows.append({\n",
    "        \"Relation\": \" / \".join(rel),\n",
    "        \"Détail\": \" - \".join(parts),\n",
    "        \"Scènes\": f\"{scenes_count} sc.\",\n",
    "        \"Texte\": f\"{format_lines(total_lines)} ({format_percent(total_lines / total_lines_play)})\",\n",
    "        \"Interlocution\": format_people(interlocution),\n",
    "    })\n",
    "\n",
    "relations_table = pd.DataFrame(relation_rows)\n",
    "relations_table"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bien que le décompte des lignes diffère toujours de celui de l'OBVIL (comme attendu et pour les mêmes raisons que précédemment), l'interlocution est identique.\n",
    "En effet, les écarts de comptage de lignes n’affectent pas l’interlocution ; seule une différence de liste d’intervenants par scène la ferait varier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Voyons maintenant si nous parvenons à reproduire le graphique proposé par l'OBVIL :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x1600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Ordonnancement des actes et des scènes\n",
    "# On affichera les labels bruts (\"Acte Premier\") mais on utilisera un ordre \"naturel\"\n",
    "scene_order = []\n",
    "\n",
    "for act in list_acts():\n",
    "    for sc in list_scenes(act=act[\"node\"]):\n",
    "        scene_order.append(f\"{act['title']} | {sc['title']}\")\n",
    "\n",
    "# Préparation des données\n",
    "df = df_counts.copy()\n",
    "\n",
    "# Création d'une clé unique pour identifier un couple acte/scène\n",
    "# Cela permet de forcer l'ordre d'affichage, au lieu de suivre un ordre alphabétique\n",
    "# qui noierait \"Acte Premier\" au milieu de la liste, par exemple\n",
    "df[\"SceneKey\"] = pd.Categorical(df[\"Acte\"] + \" | \" + df[\"Scène\"], categories=scene_order, ordered=True)\n",
    "\n",
    "# Obtention du pourcentage que représente un dialogue particulier au sein d'une scène\n",
    "df[\"share\"] = df[\"Mots\"] / df.groupby(\"SceneKey\", observed=False)[\"Mots\"].transform(\"sum\")\n",
    "\n",
    "# Calcul du total de mots pour une scène donnée\n",
    "totals = df.groupby(\"SceneKey\", observed=False)[\"Mots\"].sum()\n",
    "\n",
    "# Paramétrage du graphique\n",
    "gap = 0 # Espace vertical ajouté entre deux scènes\n",
    "label_fs = 8 # Taille de la police des étiquettes\n",
    "min_target = 10  # Hauteur minimale souhaitée\n",
    "\n",
    "# Définition de la hauteur d'une scène.\n",
    "# Nous utilisons ici une fonction racine carré.\n",
    "# Le but de ce calcul est d'éviter que les scènes contenant le moins de mots\n",
    "# se trouvent compressées en une ligne si fine qu'il ne serait pas possible\n",
    "# de distinguer les différents protagonistes.\n",
    "# On sacrifie donc le rapport proportionnel strict au profit d'une meilleure\n",
    "# lisibilité.\n",
    "def scene_height(total):\n",
    "    return max(min_target, math.sqrt(total) * factor)\n",
    "\n",
    "create_actors_colormap(df[\"Personnage\"].unique())\n",
    "\n",
    "# Calcul de l'échelle des scènes : évite qu'une scène courte soit\n",
    "# représentée par une ligne trop fine pour être distinguée\n",
    "min_total = totals.min()\n",
    "factor = min_target / math.log1p(min_total)\n",
    "\n",
    "figure, axis = plt.subplots(figsize=(12, len(scene_order) * 0.5))\n",
    "\n",
    "# Affichage de la colormap des personnages\n",
    "handles = [mpatches.Patch(color=col, label=name) for name, col in color_map.items()]\n",
    "axis.legend(handles=handles, title=\"Personnage\", bbox_to_anchor=(1.25, 1), loc=\"upper left\")\n",
    "\n",
    "y = 0 # \"Curseur\" vertical permettant de positionner les scènes\n",
    "\n",
    "# Traçage des scènes\n",
    "for scene in scene_order:\n",
    "    scene_rows = df[df[\"SceneKey\"] == scene]\n",
    "    h = scene_height(totals.loc[scene])\n",
    "\n",
    "    left = 0\n",
    "\n",
    "    for _, row in scene_rows.iterrows():\n",
    "        # broken_barth est la méthode nous permettant de tracer des barres\n",
    "        # horizontales juxtaposés\n",
    "        axis.broken_barh([(left, row[\"share\"])], (y, h),\n",
    "                       facecolors=color_map[row[\"Personnage\"]],\n",
    "                       edgecolors=\"white\", linewidth=0.5)\n",
    "\n",
    "        # Décalage horizontal de la prochaine barre\n",
    "        left += row[\"share\"]\n",
    "\n",
    "    # Étiquette correspondant à la scène\n",
    "    axis.text(1.01, y + h/2, scene, va=\"center\", fontsize=label_fs)\n",
    "\n",
    "    # Décalage vertical de la prochaine scène\n",
    "    y += h + gap\n",
    "\n",
    "axis.set_xlim(0, 1)\n",
    "axis.set_ylim(0, y)\n",
    "axis.invert_yaxis()  # Acte I en haut\n",
    "axis.set_xlabel(\"% de la scène (largeur) - une ligne par scène\")\n",
    "axis.set_yticks([]) # On masque l'ordonnée à gauche\n",
    "axis.spines[[\"top\", \"right\", \"left\", \"bottom\"]].set_visible(False) # On masque les délimitations du graphique\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "L'ordre des dialogues est respecté sur ce graphique, contrairement au graphique de l'OBVIL.\n",
    "Par exemple, dans la deuxième scène du premier acte, Cléante est bien la première à prendre la parole, et non Élise."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Graphe réseau des dialogues"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "Ici, nous allons représenter les dialogues par un graphe réseau.\n",
    "Nous allons essayer de reproduire le [graphe](https://obtic.huma-num.fr/obvil-web/corpus/moliere/moliere_avare) proposé par l'OBVIL."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<html>\n",
       "    <head>\n",
       "        <meta charset=\"utf-8\">\n",
       "        \n",
       "            <script>function neighbourhoodHighlight(params) {\n",
       "  // console.log(\"in nieghbourhoodhighlight\");\n",
       "  allNodes = nodes.get({ returnType: \"Object\" });\n",
       "  // originalNodes = JSON.parse(JSON.stringify(allNodes));\n",
       "  // if something is selected:\n",
       "  if (params.nodes.length > 0) {\n",
       "    highlightActive = true;\n",
       "    var i, j;\n",
       "    var selectedNode = params.nodes[0];\n",
       "    var degrees = 2;\n",
       "\n",
       "    // mark all nodes as hard to read.\n",
       "    for (let nodeId in allNodes) {\n",
       "      // nodeColors[nodeId] = allNodes[nodeId].color;\n",
       "      allNodes[nodeId].color = \"rgba(200,200,200,0.5)\";\n",
       "      if (allNodes[nodeId].hiddenLabel === undefined) {\n",
       "        allNodes[nodeId].hiddenLabel = allNodes[nodeId].label;\n",
       "        allNodes[nodeId].label = undefined;\n",
       "      }\n",
       "    }\n",
       "    var connectedNodes = network.getConnectedNodes(selectedNode);\n",
       "    var allConnectedNodes = [];\n",
       "\n",
       "    // get the second degree nodes\n",
       "    for (i = 1; i < degrees; i++) {\n",
       "      for (j = 0; j < connectedNodes.length; j++) {\n",
       "        allConnectedNodes = allConnectedNodes.concat(\n",
       "          network.getConnectedNodes(connectedNodes[j])\n",
       "        );\n",
       "      }\n",
       "    }\n",
       "\n",
       "    // all second degree nodes get a different color and their label back\n",
       "    for (i = 0; i < allConnectedNodes.length; i++) {\n",
       "      // allNodes[allConnectedNodes[i]].color = \"pink\";\n",
       "      allNodes[allConnectedNodes[i]].color = \"rgba(150,150,150,0.75)\";\n",
       "      if (allNodes[allConnectedNodes[i]].hiddenLabel !== undefined) {\n",
       "        allNodes[allConnectedNodes[i]].label =\n",
       "          allNodes[allConnectedNodes[i]].hiddenLabel;\n",
       "        allNodes[allConnectedNodes[i]].hiddenLabel = undefined;\n",
       "      }\n",
       "    }\n",
       "\n",
       "    // all first degree nodes get their own color and their label back\n",
       "    for (i = 0; i < connectedNodes.length; i++) {\n",
       "      // allNodes[connectedNodes[i]].color = undefined;\n",
       "      allNodes[connectedNodes[i]].color = nodeColors[connectedNodes[i]];\n",
       "      if (allNodes[connectedNodes[i]].hiddenLabel !== undefined) {\n",
       "        allNodes[connectedNodes[i]].label =\n",
       "          allNodes[connectedNodes[i]].hiddenLabel;\n",
       "        allNodes[connectedNodes[i]].hiddenLabel = undefined;\n",
       "      }\n",
       "    }\n",
       "\n",
       "    // the main node gets its own color and its label back.\n",
       "    // allNodes[selectedNode].color = undefined;\n",
       "    allNodes[selectedNode].color = nodeColors[selectedNode];\n",
       "    if (allNodes[selectedNode].hiddenLabel !== undefined) {\n",
       "      allNodes[selectedNode].label = allNodes[selectedNode].hiddenLabel;\n",
       "      allNodes[selectedNode].hiddenLabel = undefined;\n",
       "    }\n",
       "  } else if (highlightActive === true) {\n",
       "    // console.log(\"highlightActive was true\");\n",
       "    // reset all nodes\n",
       "    for (let nodeId in allNodes) {\n",
       "      // allNodes[nodeId].color = \"purple\";\n",
       "      allNodes[nodeId].color = nodeColors[nodeId];\n",
       "      // delete allNodes[nodeId].color;\n",
       "      if (allNodes[nodeId].hiddenLabel !== undefined) {\n",
       "        allNodes[nodeId].label = allNodes[nodeId].hiddenLabel;\n",
       "        allNodes[nodeId].hiddenLabel = undefined;\n",
       "      }\n",
       "    }\n",
       "    highlightActive = false;\n",
       "  }\n",
       "\n",
       "  // transform the object into an array\n",
       "  var updateArray = [];\n",
       "  if (params.nodes.length > 0) {\n",
       "    for (let nodeId in allNodes) {\n",
       "      if (allNodes.hasOwnProperty(nodeId)) {\n",
       "        // console.log(allNodes[nodeId]);\n",
       "        updateArray.push(allNodes[nodeId]);\n",
       "      }\n",
       "    }\n",
       "    nodes.update(updateArray);\n",
       "  } else {\n",
       "    // console.log(\"Nothing was selected\");\n",
       "    for (let nodeId in allNodes) {\n",
       "      if (allNodes.hasOwnProperty(nodeId)) {\n",
       "        // console.log(allNodes[nodeId]);\n",
       "        // allNodes[nodeId].color = {};\n",
       "        updateArray.push(allNodes[nodeId]);\n",
       "      }\n",
       "    }\n",
       "    nodes.update(updateArray);\n",
       "  }\n",
       "}\n",
       "\n",
       "function filterHighlight(params) {\n",
       "  allNodes = nodes.get({ returnType: \"Object\" });\n",
       "  // if something is selected:\n",
       "  if (params.nodes.length > 0) {\n",
       "    filterActive = true;\n",
       "    let selectedNodes = params.nodes;\n",
       "\n",
       "    // hiding all nodes and saving the label\n",
       "    for (let nodeId in allNodes) {\n",
       "      allNodes[nodeId].hidden = true;\n",
       "      if (allNodes[nodeId].savedLabel === undefined) {\n",
       "        allNodes[nodeId].savedLabel = allNodes[nodeId].label;\n",
       "        allNodes[nodeId].label = undefined;\n",
       "      }\n",
       "    }\n",
       "\n",
       "    for (let i=0; i < selectedNodes.length; i++) {\n",
       "      allNodes[selectedNodes[i]].hidden = false;\n",
       "      if (allNodes[selectedNodes[i]].savedLabel !== undefined) {\n",
       "        allNodes[selectedNodes[i]].label = allNodes[selectedNodes[i]].savedLabel;\n",
       "        allNodes[selectedNodes[i]].savedLabel = undefined;\n",
       "      }\n",
       "    }\n",
       "\n",
       "  } else if (filterActive === true) {\n",
       "    // reset all nodes\n",
       "    for (let nodeId in allNodes) {\n",
       "      allNodes[nodeId].hidden = false;\n",
       "      if (allNodes[nodeId].savedLabel !== undefined) {\n",
       "        allNodes[nodeId].label = allNodes[nodeId].savedLabel;\n",
       "        allNodes[nodeId].savedLabel = undefined;\n",
       "      }\n",
       "    }\n",
       "    filterActive = false;\n",
       "  }\n",
       "\n",
       "  // transform the object into an array\n",
       "  var updateArray = [];\n",
       "  if (params.nodes.length > 0) {\n",
       "    for (let nodeId in allNodes) {\n",
       "      if (allNodes.hasOwnProperty(nodeId)) {\n",
       "        updateArray.push(allNodes[nodeId]);\n",
       "      }\n",
       "    }\n",
       "    nodes.update(updateArray);\n",
       "  } else {\n",
       "    for (let nodeId in allNodes) {\n",
       "      if (allNodes.hasOwnProperty(nodeId)) {\n",
       "        updateArray.push(allNodes[nodeId]);\n",
       "      }\n",
       "    }\n",
       "    nodes.update(updateArray);\n",
       "  }\n",
       "}\n",
       "\n",
       "function selectNode(nodes) {\n",
       "  network.selectNodes(nodes);\n",
       "  neighbourhoodHighlight({ nodes: nodes });\n",
       "  return nodes;\n",
       "}\n",
       "\n",
       "function selectNodes(nodes) {\n",
       "  network.selectNodes(nodes);\n",
       "  filterHighlight({nodes: nodes});\n",
       "  return nodes;\n",
       "}\n",
       "\n",
       "function highlightFilter(filter) {\n",
       "  let selectedNodes = []\n",
       "  let selectedProp = filter['property']\n",
       "  if (filter['item'] === 'node') {\n",
       "    let allNodes = nodes.get({ returnType: \"Object\" });\n",
       "    for (let nodeId in allNodes) {\n",
       "      if (allNodes[nodeId][selectedProp] && filter['value'].includes((allNodes[nodeId][selectedProp]).toString())) {\n",
       "        selectedNodes.push(nodeId)\n",
       "      }\n",
       "    }\n",
       "  }\n",
       "  else if (filter['item'] === 'edge'){\n",
       "    let allEdges = edges.get({returnType: 'object'});\n",
       "    // check if the selected property exists for selected edge and select the nodes connected to the edge\n",
       "    for (let edge in allEdges) {\n",
       "      if (allEdges[edge][selectedProp] && filter['value'].includes((allEdges[edge][selectedProp]).toString())) {\n",
       "        selectedNodes.push(allEdges[edge]['from'])\n",
       "        selectedNodes.push(allEdges[edge]['to'])\n",
       "      }\n",
       "    }\n",
       "  }\n",
       "  selectNodes(selectedNodes)\n",
       "}</script>\n",
       "            <link rel=\"stylesheet\" href=\"https://cdnjs.cloudflare.com/ajax/libs/vis-network/9.1.2/dist/dist/vis-network.min.css\" integrity=\"sha512-WgxfT5LWjfszlPHXRmBWHkV2eceiWTOBvrKCNbdgDYTHrT2AeLCGbF4sZlZw3UMN3WtL0tGUoIAKsu8mllg/XA==\" crossorigin=\"anonymous\" referrerpolicy=\"no-referrer\" />\n",
       "            <script src=\"https://cdnjs.cloudflare.com/ajax/libs/vis-network/9.1.2/dist/vis-network.min.js\" integrity=\"sha512-LnvoEWDFrqGHlHmDD2101OrLcbsfkrzoSpvtSQtxK3RMnRV0eOkhhBN2dXHKRrUU8p2DGRTk35n4O8nWSVe1mQ==\" crossorigin=\"anonymous\" referrerpolicy=\"no-referrer\"></script>\n",
       "            \n",
       "            \n",
       "            \n",
       "            \n",
       "            \n",
       "            \n",
       "\n",
       "        \n",
       "<center>\n",
       "<h1></h1>\n",
       "</center>\n",
       "\n",
       "<!-- <link rel=\"stylesheet\" href=\"../node_modules/vis/dist/vis.min.css\" type=\"text/css\" />\n",
       "<script type=\"text/javascript\" src=\"../node_modules/vis/dist/vis.js\"> </script>-->\n",
       "        <link\n",
       "          href=\"https://cdn.jsdelivr.net/npm/bootstrap@5.0.0-beta3/dist/css/bootstrap.min.css\"\n",
       "          rel=\"stylesheet\"\n",
       "          integrity=\"sha384-eOJMYsd53ii+scO/bJGFsiCZc+5NDVN2yr8+0RDqr0Ql0h+rP48ckxlpbzKgwra6\"\n",
       "          crossorigin=\"anonymous\"\n",
       "        />\n",
       "        <script\n",
       "          src=\"https://cdn.jsdelivr.net/npm/bootstrap@5.0.0-beta3/dist/js/bootstrap.bundle.min.js\"\n",
       "          integrity=\"sha384-JEW9xMcG8R+pH31jmWH6WWP0WintQrMb4s7ZOdauHnUtxwoG2vI5DkLtS3qm9Ekf\"\n",
       "          crossorigin=\"anonymous\"\n",
       "        ></script>\n",
       "\n",
       "\n",
       "        <center>\n",
       "          <h1></h1>\n",
       "        </center>\n",
       "        <style type=\"text/css\">\n",
       "\n",
       "             #mynetwork {\n",
       "                 width: 100%;\n",
       "                 height: 700px;\n",
       "                 background-color: #f8f8f8;\n",
       "                 border: 1px solid lightgray;\n",
       "                 position: relative;\n",
       "                 float: left;\n",
       "             }\n",
       "\n",
       "             \n",
       "\n",
       "             \n",
       "\n",
       "             \n",
       "        </style>\n",
       "    </head>\n",
       "\n",
       "\n",
       "    <body>\n",
       "        <div class=\"card\" style=\"width: 100%\">\n",
       "            \n",
       "            \n",
       "            <div id=\"mynetwork\" class=\"card-body\"></div>\n",
       "        </div>\n",
       "\n",
       "        \n",
       "        \n",
       "\n",
       "        <script type=\"text/javascript\">\n",
       "\n",
       "              // initialize global variables.\n",
       "              var edges;\n",
       "              var nodes;\n",
       "              var allNodes;\n",
       "              var allEdges;\n",
       "              var nodeColors;\n",
       "              var originalNodes;\n",
       "              var network;\n",
       "              var container;\n",
       "              var options, data;\n",
       "              var filter = {\n",
       "                  item : '',\n",
       "                  property : '',\n",
       "                  value : []\n",
       "              };\n",
       "\n",
       "              \n",
       "\n",
       "              \n",
       "\n",
       "              // This method is responsible for drawing the graph, returns the drawn network\n",
       "              function drawGraph() {\n",
       "                  var container = document.getElementById('mynetwork');\n",
       "\n",
       "                  \n",
       "\n",
       "                  // parsing and collecting nodes and edges from the python\n",
       "                  nodes = new vis.DataSet([{\"color\": \"#9467bd\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Anselme\", \"label\": \"Anselme\", \"shape\": \"dot\", \"size\": 21.642580401082487, \"title\": \"517 mots\"}, {\"color\": \"#e377c2\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Brindavoine\", \"label\": \"Brindavoine\", \"shape\": \"dot\", \"size\": 11.9344631145812, \"title\": \"43 mots\"}, {\"color\": \"#aec7e8\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Cl\\u00e9ante\", \"label\": \"Cl\\u00e9ante\", \"shape\": \"dot\", \"size\": 42.68083043988422, \"title\": \"3341 mots\"}, {\"color\": \"#ffbb78\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Frosine\", \"label\": \"Frosine\", \"shape\": \"dot\", \"size\": 37.01792549442499, \"title\": \"2339 mots\"}, {\"color\": \"#1f77b4\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Harpagon\", \"label\": \"Harpagon\", \"shape\": \"dot\", \"size\": 54.98425268108454, \"title\": \"6132 mots\"}, {\"color\": \"#98df8a\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"La Fl\\u00e8che\", \"label\": \"La Fl\\u00e8che\", \"shape\": \"dot\", \"size\": 31.330666514268298, \"title\": \"1512 mots\"}, {\"color\": \"#c49c94\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"La Merluche\", \"label\": \"La Merluche\", \"shape\": \"dot\", \"size\": 12.113392760240625, \"title\": \"47 mots\"}, {\"color\": \"#c5b0d5\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Le Commissaire\", \"label\": \"Le Commissaire\", \"shape\": \"dot\", \"size\": 18.287856919689347, \"title\": \"294 mots\"}, {\"color\": \"#ff9896\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Mariane\", \"label\": \"Mariane\", \"shape\": \"dot\", \"size\": 26.18900766946894, \"title\": \"919 mots\"}, {\"color\": \"#2ca02c\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Ma\\u00eetre Jacques\", \"label\": \"Ma\\u00eetre Jacques\", \"shape\": \"dot\", \"size\": 32.53405796031305, \"title\": \"1672 mots\"}, {\"color\": \"#8c564b\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Ma\\u00eetre Simon\", \"label\": \"Ma\\u00eetre Simon\", \"shape\": \"dot\", \"size\": 16.42140130857092, \"title\": \"197 mots\"}, {\"color\": \"#ff7f0e\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"Val\\u00e8re\", \"label\": \"Val\\u00e8re\", \"shape\": \"dot\", \"size\": 39.309423501559394, \"title\": \"2723 mots\"}, {\"color\": \"#d62728\", \"font\": {\"align\": \"center\", \"color\": \"#111\", \"size\": 18}, \"id\": \"\\u00c9lise\", \"label\": \"\\u00c9lise\", \"shape\": \"dot\", \"size\": 27.59897956527329, \"title\": \"1067 mots\"}]);\n",
       "                  edges = new vis.DataSet([{\"arrows\": \"to\", \"color\": \"#9467bd\", \"from\": \"Anselme\", \"title\": \"293 mots\", \"to\": \"Harpagon\", \"value\": 5.683579767338681}, {\"arrows\": \"to\", \"color\": \"#9467bd\", \"from\": \"Anselme\", \"title\": \"13 mots\", \"to\": \"Le Commissaire\", \"value\": 2.639057329615259}, {\"arrows\": \"to\", \"color\": \"#9467bd\", \"from\": \"Anselme\", \"title\": \"211 mots\", \"to\": \"Val\\u00e8re\", \"value\": 5.356586274672012}, {\"arrows\": \"to\", \"color\": \"#e377c2\", \"from\": \"Brindavoine\", \"title\": \"20 mots\", \"to\": \"Harpagon\", \"value\": 3.044522437723423}, {\"arrows\": \"to\", \"color\": \"#e377c2\", \"from\": \"Brindavoine\", \"title\": \"23 mots\", \"to\": \"La Merluche\", \"value\": 3.1780538303479458}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"32 mots\", \"to\": \"Frosine\", \"value\": 3.4965075614664802}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"1587 mots\", \"to\": \"Harpagon\", \"value\": 7.370230641807081}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"385 mots\", \"to\": \"La Fl\\u00e8che\", \"value\": 5.955837369464831}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"350 mots\", \"to\": \"Mariane\", \"value\": 5.860786223465865}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"170 mots\", \"to\": \"Ma\\u00eetre Jacques\", \"value\": 5.14166355650266}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"13 mots\", \"to\": \"Ma\\u00eetre Simon\", \"value\": 2.639057329615259}, {\"arrows\": \"to\", \"color\": \"#aec7e8\", \"from\": \"Cl\\u00e9ante\", \"title\": \"758 mots\", \"to\": \"\\u00c9lise\", \"value\": 6.6320017773956295}, {\"arrows\": \"to\", \"color\": \"#ffbb78\", \"from\": \"Frosine\", \"title\": \"436 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 6.07993319509559}, {\"arrows\": \"to\", \"color\": \"#ffbb78\", \"from\": \"Frosine\", \"title\": \"1498 mots\", \"to\": \"Harpagon\", \"value\": 7.312553498102598}, {\"arrows\": \"to\", \"color\": \"#ffbb78\", \"from\": \"Frosine\", \"title\": \"130 mots\", \"to\": \"La Fl\\u00e8che\", \"value\": 4.875197323201151}, {\"arrows\": \"to\", \"color\": \"#ffbb78\", \"from\": \"Frosine\", \"title\": \"203 mots\", \"to\": \"Mariane\", \"value\": 5.318119993844216}, {\"arrows\": \"to\", \"color\": \"#ffbb78\", \"from\": \"Frosine\", \"title\": \"10 mots\", \"to\": \"Ma\\u00eetre Jacques\", \"value\": 2.3978952727983707}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"120 mots\", \"to\": \"Anselme\", \"value\": 4.795790545596741}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"27 mots\", \"to\": \"Brindavoine\", \"value\": 3.332204510175204}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"1661 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 7.415776975415394}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"678 mots\", \"to\": \"Frosine\", \"value\": 6.520621127558696}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"434 mots\", \"to\": \"La Fl\\u00e8che\", \"value\": 6.075346031088684}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"6 mots\", \"to\": \"La Merluche\", \"value\": 1.9459101490553132}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"135 mots\", \"to\": \"Le Commissaire\", \"value\": 4.912654885736052}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"94 mots\", \"to\": \"Mariane\", \"value\": 4.553876891600541}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"677 mots\", \"to\": \"Ma\\u00eetre Jacques\", \"value\": 6.519147287940395}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"51 mots\", \"to\": \"Ma\\u00eetre Simon\", \"value\": 3.9512437185814275}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"1129 mots\", \"to\": \"Val\\u00e8re\", \"value\": 7.029972911706386}, {\"arrows\": \"to\", \"color\": \"#1f77b4\", \"from\": \"Harpagon\", \"title\": \"506 mots\", \"to\": \"\\u00c9lise\", \"value\": 6.2285110035911835}, {\"arrows\": \"to\", \"color\": \"#98df8a\", \"from\": \"La Fl\\u00e8che\", \"title\": \"953 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 6.860663671448287}, {\"arrows\": \"to\", \"color\": \"#98df8a\", \"from\": \"La Fl\\u00e8che\", \"title\": \"196 mots\", \"to\": \"Frosine\", \"value\": 5.2832037287379885}, {\"arrows\": \"to\", \"color\": \"#98df8a\", \"from\": \"La Fl\\u00e8che\", \"title\": \"258 mots\", \"to\": \"Harpagon\", \"value\": 5.556828061699537}, {\"arrows\": \"to\", \"color\": \"#c49c94\", \"from\": \"La Merluche\", \"title\": \"47 mots\", \"to\": \"Harpagon\", \"value\": 3.8712010109078907}, {\"arrows\": \"to\", \"color\": \"#c5b0d5\", \"from\": \"Le Commissaire\", \"title\": \"113 mots\", \"to\": \"Harpagon\", \"value\": 4.736198448394496}, {\"arrows\": \"to\", \"color\": \"#c5b0d5\", \"from\": \"Le Commissaire\", \"title\": \"148 mots\", \"to\": \"Ma\\u00eetre Jacques\", \"value\": 5.003946305945459}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"199 mots\", \"to\": \"Anselme\", \"value\": 5.298317366548036}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"245 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 5.5053315359323625}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"231 mots\", \"to\": \"Frosine\", \"value\": 5.44673737166631}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"163 mots\", \"to\": \"Harpagon\", \"value\": 5.099866427824199}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"29 mots\", \"to\": \"Val\\u00e8re\", \"value\": 3.4011973816621555}, {\"arrows\": \"to\", \"color\": \"#ff9896\", \"from\": \"Mariane\", \"title\": \"14 mots\", \"to\": \"\\u00c9lise\", \"value\": 2.70805020110221}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"23 mots\", \"to\": \"Anselme\", \"value\": 3.1780538303479458}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"154 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 5.043425116919247}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"18 mots\", \"to\": \"Frosine\", \"value\": 2.9444389791664403}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"1132 mots\", \"to\": \"Harpagon\", \"value\": 7.0326242610280065}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"8 mots\", \"to\": \"La Merluche\", \"value\": 2.1972245773362196}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"48 mots\", \"to\": \"Le Commissaire\", \"value\": 3.8918202981106265}, {\"arrows\": \"to\", \"color\": \"#2ca02c\", \"from\": \"Ma\\u00eetre Jacques\", \"title\": \"218 mots\", \"to\": \"Val\\u00e8re\", \"value\": 5.389071729816501}, {\"arrows\": \"to\", \"color\": \"#8c564b\", \"from\": \"Ma\\u00eetre Simon\", \"title\": \"194 mots\", \"to\": \"Harpagon\", \"value\": 5.272999558563747}, {\"arrows\": \"to\", \"color\": \"#8c564b\", \"from\": \"Ma\\u00eetre Simon\", \"title\": \"3 mots\", \"to\": \"La Fl\\u00e8che\", \"value\": 1.3862943611198906}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"262 mots\", \"to\": \"Anselme\", \"value\": 5.572154032177765}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"5 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 1.791759469228055}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"1432 mots\", \"to\": \"Harpagon\", \"value\": 7.267525427828172}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"4 mots\", \"to\": \"La Merluche\", \"value\": 1.6094379124341003}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"43 mots\", \"to\": \"Mariane\", \"value\": 3.784189633918261}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"218 mots\", \"to\": \"Ma\\u00eetre Jacques\", \"value\": 5.389071729816501}, {\"arrows\": \"to\", \"color\": \"#ff7f0e\", \"from\": \"Val\\u00e8re\", \"title\": \"742 mots\", \"to\": \"\\u00c9lise\", \"value\": 6.610696044717759}, {\"arrows\": \"to\", \"color\": \"#d62728\", \"from\": \"\\u00c9lise\", \"title\": \"154 mots\", \"to\": \"Cl\\u00e9ante\", \"value\": 5.043425116919247}, {\"arrows\": \"to\", \"color\": \"#d62728\", \"from\": \"\\u00c9lise\", \"title\": \"10 mots\", \"to\": \"Frosine\", \"value\": 2.3978952727983707}, {\"arrows\": \"to\", \"color\": \"#d62728\", \"from\": \"\\u00c9lise\", \"title\": \"328 mots\", \"to\": \"Harpagon\", \"value\": 5.796057750765372}, {\"arrows\": \"to\", \"color\": \"#d62728\", \"from\": \"\\u00c9lise\", \"title\": \"48 mots\", \"to\": \"Mariane\", \"value\": 3.8918202981106265}, {\"arrows\": \"to\", \"color\": \"#d62728\", \"from\": \"\\u00c9lise\", \"title\": \"514 mots\", \"to\": \"Val\\u00e8re\", \"value\": 6.244166900663736}]);\n",
       "\n",
       "                  nodeColors = {};\n",
       "                  allNodes = nodes.get({ returnType: \"Object\" });\n",
       "                  for (nodeId in allNodes) {\n",
       "                    nodeColors[nodeId] = allNodes[nodeId].color;\n",
       "                  }\n",
       "                  allEdges = edges.get({ returnType: \"Object\" });\n",
       "                  // adding nodes and edges to the graph\n",
       "                  data = {nodes: nodes, edges: edges};\n",
       "\n",
       "                  var options = {\"physics\": {\"solver\": \"repulsion\", \"repulsion\": {\"nodeDistance\": 240, \"springLength\": 180, \"springConstant\": 0.05, \"damping\": 0.12}}, \"edges\": {\"arrows\": {\"to\": {\"enabled\": true}}, \"smooth\": {\"type\": \"dynamic\"}}, \"nodes\": {\"shape\": \"dot\", \"borderWidth\": 1.2}};\n",
       "\n",
       "                  \n",
       "\n",
       "\n",
       "                  \n",
       "\n",
       "                  network = new vis.Network(container, data, options);\n",
       "\n",
       "                  \n",
       "\n",
       "                  \n",
       "\n",
       "                  \n",
       "\n",
       "\n",
       "                  \n",
       "\n",
       "                  return network;\n",
       "\n",
       "              }\n",
       "              drawGraph();\n",
       "        </script>\n",
       "    </body>\n",
       "</html>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Arêtes directionnelles source -> cible : succession des locuteurs par scène\n",
    "def interlocution_edges(df_speeches):\n",
    "    df = df_speeches.copy()\n",
    "    df[\"SceneKey\"] = df[\"Acte\"] + \" | \" + df[\"Scène\"]\n",
    "    df[\"order\"] = range(len(df))\n",
    "    df = df.sort_values([\"SceneKey\", \"order\"])\n",
    "    df[\"next\"] = df.groupby(\"SceneKey\")[\"Personnage\"].shift(-1)\n",
    "\n",
    "    return (df.dropna(subset=[\"next\"])\n",
    "              .groupby([\"Personnage\", \"next\"], as_index=False)[\"Mots\"]\n",
    "              .sum()\n",
    "              .rename(columns={\"Personnage\": \"source\", \"next\": \"target\", \"Mots\": \"weight\"}))\n",
    "\n",
    "# Données\n",
    "totaux = df_speeches.groupby(\"Personnage\")[\"Mots\"].sum()\n",
    "edges_dir = interlocution_edges(df_speeches)\n",
    "\n",
    "# Template + ressources inline (vis.js embarqué)\n",
    "tpl_dir = Path(pyvis.__file__).with_name(\"templates\")\n",
    "env = Environment(loader=FileSystemLoader(tpl_dir))\n",
    "\n",
    "net = Network(\n",
    "    height=\"700px\",\n",
    "    width=\"100%\",\n",
    "    bgcolor=\"#f8f8f8\",\n",
    "    notebook=True,\n",
    "    directed=True,\n",
    "    cdn_resources=\"remote\",\n",
    ")\n",
    "net.template = env.get_template(\"template.html\")\n",
    "\n",
    "# Réglage vis.js : on « écarte » les nœuds avec le solver repulsion\n",
    "# On importe exceptionnellement json ici pour pouvoir travailler les options\n",
    "# directement dans ce format\n",
    "import json\n",
    "\n",
    "options = {\n",
    "    \"physics\": {\n",
    "        \"solver\": \"repulsion\",\n",
    "        \"repulsion\": {\n",
    "            \"nodeDistance\": 240,\n",
    "            \"springLength\": 180,\n",
    "            \"springConstant\": 0.05,\n",
    "            \"damping\": 0.12,\n",
    "        },\n",
    "    },\n",
    "    \"edges\": {\n",
    "        \"arrows\": {\"to\": {\"enabled\": True}},\n",
    "        \"smooth\": {\"type\": \"dynamic\"},\n",
    "    },\n",
    "    \"nodes\": {\n",
    "        \"shape\": \"dot\",\n",
    "        \"borderWidth\": 1.2,\n",
    "    },\n",
    "}\n",
    "\n",
    "net.set_options(json.dumps(options))\n",
    "\n",
    "# Noeuds : label centré dans la bulle (font align center)\n",
    "for n, mots in totaux.items():\n",
    "    net.add_node(\n",
    "        n,\n",
    "        label=n,\n",
    "        title=f\"{mots} mots\",\n",
    "        shape=\"dot\",\n",
    "        size=8 + 0.6 * np.sqrt(mots),\n",
    "        color=mcolors.to_hex(color_map.get(n)) if color_map.get(n) else None,\n",
    "        font={\"size\": 18, \"align\": \"center\", \"color\": \"#111\"},\n",
    "    )\n",
    "\n",
    "# Arêtes : poids log1p, couleur héritée de la source\n",
    "for _, r in edges_dir.iterrows():\n",
    "    if r[\"source\"] == r[\"target\"]:\n",
    "        continue\n",
    "    net.add_edge(\n",
    "        r[\"source\"],\n",
    "        r[\"target\"],\n",
    "        value=np.log1p(r[\"weight\"]),\n",
    "        title=f\"{r['weight']} mots\",\n",
    "        color=net.get_node(r[\"source\"])[\"color\"],\n",
    "    )\n",
    "\n",
    "# On évite l'iframe qui poserait un problème de sécurité (Content Security Policy)\n",
    "html = net.generate_html(notebook=True)\n",
    "HTML(html)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "Ce graphique 3D interactif permet de cliquer sur un nom et de le déplacer afin de faire apparaître toutes les relations. En outre, le survol d'un nom ou de la flèche représentant une relation donnera le nombre de mots associés.\n",
    "\n",
    "La pile logicielle employée par l'OBVIL pour son propre graphique repose sur Sigma et ForceAtlas2, des modules `nodejs` que je ne souhaitais pas exploiter dans ce notebook."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "## Conclusion\n",
    "\n",
    "Cette étude a révélé que malgré tout le soin que l'on peut apporter au traitement d'un sujet spécifique, il est possible d'introduire involontairement des \"artefacts\" dans les données sur lesquelles on travaille. \n",
    "Ici, il s'agit de différences d'orthographes mineures, invisibles lors d'une lecture par un humain, mais qui peuvent rendre une étude assistée par l'informatique plus complexe, voire sujette aux erreurs.\n",
    "On le répète assez souvent en informatique (et particulièrement en développement web) : on ne doit jamais avoir confiance dans les entrées...\n",
    "Cela a nécessité un travail assez important en amont, et il faut préciser que ce n'est pas infaillible.\n",
    "\n",
    "Une autre source potentielle d'erreur, que ce soit au niveau de l'analyse ou de l'interprétation, consiste en des définitions ou des méthodologies différentes.\n",
    "Ici, nous avons calculé des statistiques divergentes de celles de l'OBVIL, probablement en raison d'une définition différente de ce qu'est une \"ligne\".\n",
    "Pour notre étude, nous considérons qu'une ligne est une suite de 60 caractères, dont les espaces surnuméraires ont été supprimés (incluant les sauts de lignes).\n",
    "\n",
    "Or, ce n'est qu'après une inspection plus poussée du dépôt de l'outil [dramagraph](https://github.com/dramacode/dramagraph/tree/gh-pages) de l'OBLIV révèle la méthode : après \"nettoyage\" des fichiers TEI source par l'emploi d'une [feuille de style XSL](https://github.com/dramacode/dramagraph/blob/gh-pages/naked.xsl), [un autre fichier XSL](https://github.com/dramacode/dramagraph/blob/gh-pages/drama2csv.xsl#L517) est en charge, notamment, du formatage des paragraphes (séparés par des retours de ligne), de la [gestion des accents](https://github.com/dramacode/dramagraph/blob/gh-pages/drama2csv.xsl#L14) et de [la casse](https://github.com/dramacode/dramagraph/blob/gh-pages/drama2csv.xsl#L524), et enfin de la production [des compteurs](https://github.com/dramacode/dramagraph/blob/gh-pages/drama2csv.xsl#L490).\n",
    "\n",
    "Par conséquent, pour retrouver des statistiques identiques à l'OBVIL, il aurait fallut passer par le même _pipeline_.\n",
    "On aurait du choisir d'utiliser le TEI comme fichier source et lui appliquer les mêmes fichiers XSL, ce qui nous aurait donné directement accès aux statistiques, sans avoir besoin de les recalculer nous-même, réduisant considérablement la taille de ce notebook.\n",
    "\n",
    "On en déduit finalement que :\n",
    "\n",
    "- l'enthousiasme est parfois un ennemi ! J'aurai du prendre davantage de temps pour examiner comment l'OBVIL a produit ses statistiques, avant de me lancer dans une étude personnelle\n",
    "- bien que structuré, HTML est un \"produit transformé\" : les fichiers TEI ont manifestement servi comme base à tous les autres formats proposés par l'OBVIL ; j'aurai du l'identifier comme source idéale"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}