{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8e67c736",
   "metadata": {},
   "source": [
    "# 12 — Annotation en aveugle : l'écart de registre ne survit pas à l'étiquetage\n",
    "\n",
    "Le [notebook 11](11_corpus_etendu.ipynb) a laissé une question ouverte, et une seule.\n",
    "L'appartenance à une catégorie de Wikipédia s'était révélée un **indicateur bruité** du\n",
    "registre émotionnel — « Lil Tay » figure dans une catégorie de canulars, mais son audience est\n",
    "celle d'une célébrité. Or un bruit d'étiquetage attire tout écart vers zéro. Le résultat nul\n",
    "du corpus étendu était donc compatible avec deux lectures que ces données ne séparaient pas :\n",
    "absence d'effet, ou effet dilué.\n",
    "\n",
    "Ce notebook les sépare. Les 440 sujets ont été **annotés à la main**, à partir du seul couple\n",
    "titre + chapeau d'article, sous une grille [pré-enregistrée](../docs/annotation.md) et\n",
    "antérieure à toute annotation.\n",
    "\n",
    "**Ce que l'annotation établit :**\n",
    "\n",
    "* le bruit d'étiquetage est **massif** — 40 % des sujets ne relèvent d'aucun des deux\n",
    "  registres, et l'accord entre catégorie et annotation n'est que de 59,5 % ;\n",
    "* l'écart de taux de basculement **disparaît complètement** : 8,6 % contre 2,7 % devient\n",
    "  **4,8 % contre 5,1 %**, rapport de cotes 0,93, $p = 1{,}00$ ;\n",
    "* l'écart de persistance reste nul, ×3,04 contre ×2,90, $p = 0{,}90$, et le test conserve la\n",
    "  puissance de détecter un écart de l'ampleur annoncée par le corpus pilote ;\n",
    "* et l'annotation expose un **défaut de conception** que ni le pilote ni le corpus étendu\n",
    "  n'avaient pu voir : correctement étiquetés, les deux registres ne portent presque pas sur\n",
    "  les mêmes types de sujets."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2eb4ef9d",
   "metadata": {},
   "source": [
    "## 1. Ce que l'annotation ajoute, et ce qu'elle ne change pas\n",
    "\n",
    "L'annotation **ne modifie pas le corpus** : elle lui ajoute une colonne. Le pool reste celui\n",
    "que les dix-sept catégories déclarées ont produit, et les 440 sujets sont annotés sans\n",
    "exception — annoter un sous-ensemble choisi rouvrirait le biais de sélection que le corpus\n",
    "dérivé de catégories avait fermé.\n",
    "\n",
    "La grille pose une seule question : **qu'est-ce qui mobiliserait l'attention du public sur cet\n",
    "article ?** Non pas de quoi l'article parle, mais quelle émotion porterait sa consultation.\n",
    "\n",
    "| Registre | Définition |\n",
    "|---|---|\n",
    "| `accusation` | une faute, une menace ou une tromperie attribuée à quelqu'un |\n",
    "| `discovery` | une découverte, une exploration ou une réussite |\n",
    "| `neither` | ni l'un ni l'autre — divertissement, célébrité, institution ordinaire, entrée de catalogue |\n",
    "\n",
    "C'est la troisième étiquette qui fait le travail, et le taux auquel elle est employée **mesure**\n",
    "le bruit d'étiquetage que le notebook 11 n'avait pu que diagnostiquer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8a722b68",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:00.582363Z",
     "iopub.status.busy": "2026-08-23T13:39:00.582131Z",
     "iopub.status.idle": "2026-08-23T13:39:01.134015Z",
     "shell.execute_reply": "2026-08-23T13:39:01.133496Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grille           : version 1.0\n",
      "manifeste        : 440 sujets\n",
      "annotations      : 440  (annotations.json)\n",
      "\n",
      "registre annoté\n",
      "  accusation  147  (33.4 %)\n",
      "  discovery   118  (26.8 %)\n",
      "  neither     175  (39.8 %)\n",
      "\n",
      "annotations incertaines : 79 sur 440  (toutes justifiées par écrit)\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import stats\n",
    "\n",
    "from ide.annotation import (\n",
    "    ANNOTATIONS_PATH,\n",
    "    CONTAMINATED,\n",
    "    KINDS,\n",
    "    REGISTERS,\n",
    "    RUBRIC_VERSION,\n",
    "    confusion_matrix,\n",
    "    load_annotations,\n",
    ")\n",
    "from ide.catalogue import load_catalogue\n",
    "from ide.pageviews import load_cached\n",
    "from ide.plotting import PALETTE, save_figure, use_project_style\n",
    "from ide.regime import scan_regime_shifts\n",
    "\n",
    "use_project_style()\n",
    "\n",
    "entries, report = load_catalogue()\n",
    "annotations = load_annotations()\n",
    "\n",
    "print(f\"grille           : version {RUBRIC_VERSION}\")\n",
    "print(f\"manifeste        : {len(entries)} sujets\")\n",
    "print(f\"annotations      : {len(annotations)}  ({ANNOTATIONS_PATH.name})\")\n",
    "\n",
    "counts = Counter(item.register for item in annotations.values())\n",
    "print(\"\\nregistre annoté\")\n",
    "for register in REGISTERS:\n",
    "    print(f\"  {register:11s} {counts[register]:3d}  ({100 * counts[register] / len(annotations):4.1f} %)\")\n",
    "\n",
    "confidences = Counter(item.confidence for item in annotations.values())\n",
    "print(f\"\\nannotations incertaines : {confidences['unsure']} sur {len(annotations)}\"\n",
    "      f\"  (toutes justifiées par écrit)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "feedcfb6",
   "metadata": {},
   "source": [
    "## 2. Le bruit d'étiquetage, mesuré\n",
    "\n",
    "Première mesure, et elle est un résultat en soi : **à quel point l'appartenance à une catégorie\n",
    "prédit-elle le registre ?**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "95d5a4fd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:01.134888Z",
     "iopub.status.busy": "2026-08-23T13:39:01.134787Z",
     "iopub.status.idle": "2026-08-23T13:39:01.138110Z",
     "shell.execute_reply": "2026-08-23T13:39:01.137681Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "catégorie        n     accusation     découverte   ni l'un ni l'autre\n",
      "----------------------------------------------------------------------\n",
      "accusation     220   147 (66.8 %)     3 ( 1.4 %)          70 (31.8 %)\n",
      "discovery      220     0 ( 0.0 %)   115 (52.3 %)         105 (47.7 %)\n",
      "\n",
      "accord catégorie / annotation : 262/440 = 59.5 %\n",
      "registre franchement inversé  : 3\n"
     ]
    }
   ],
   "source": [
    "matrix = confusion_matrix(entries, annotations)\n",
    "\n",
    "other = \"ni l'un ni l'autre\"\n",
    "print(f\"{'catégorie':12s} {'n':>5s} {'accusation':>14s} {'découverte':>14s} {other:>20s}\")\n",
    "print(\"-\" * 70)\n",
    "for category in (\"accusation\", \"discovery\"):\n",
    "    row = matrix[category]\n",
    "    total = sum(row.values())\n",
    "    cells = [f\"{row[r]:4d} ({100 * row[r] / total:4.1f} %)\" for r in REGISTERS]\n",
    "    print(f\"{category:12s} {total:5d} {cells[0]:>14s} {cells[1]:>14s} {cells[2]:>20s}\")\n",
    "\n",
    "agreement = sum(1 for e in entries if annotations[e.label].register == e.category)\n",
    "crossed = sum(1 for e in entries\n",
    "              if annotations[e.label].register not in (e.category, \"neither\"))\n",
    "\n",
    "print(f\"\\naccord catégorie / annotation : {agreement}/{len(entries)}\"\n",
    "      f\" = {100 * agreement / len(entries):.1f} %\")\n",
    "print(f\"registre franchement inversé  : {crossed}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb9afbf0",
   "metadata": {},
   "source": [
    "### Lecture\n",
    "\n",
    "**Deux sujets sur cinq ne relèvent d'aucun des deux registres.** Le bruit n'est pas marginal,\n",
    "il est majoritaire dans un cas sur trois côté accusation et un cas sur deux côté découverte.\n",
    "\n",
    "Il est aussi **asymétrique**, et pour une raison de construction : le registre « découverte »\n",
    "tirait ses effectifs de catalogues d'objets célestes, dont l'immense majorité sont des entrées\n",
    "techniques sans public. Le filtre de substance de dix mille octets n'y suffisait pas.\n",
    "\n",
    "En revanche, le registre franchement inversé est **quasi nul** : la catégorie se trompe en\n",
    "capturant des sujets hors registre, presque jamais en attribuant le mauvais registre. C'est\n",
    "exactement le profil d'un bruit qui dilue sans biaiser."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d4c05052",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:01.138838Z",
     "iopub.status.busy": "2026-08-23T13:39:01.138774Z",
     "iopub.status.idle": "2026-08-23T13:39:56.164917Z",
     "shell.execute_reply": "2026-08-23T13:39:56.164432Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sujets analysés      : 440\n",
      "changements détectés : 28\n",
      "dont identifiés      : 1\n"
     ]
    }
   ],
   "source": [
    "profiles, shifts = [], []\n",
    "for entry in entries:\n",
    "    series = load_cached(entry.project, entry.article)\n",
    "    if series is None:\n",
    "        continue\n",
    "    values = np.clip(series.filled(), 1.0, None)\n",
    "    try:\n",
    "        outcome = scan_regime_shifts(values, label=entry.label)\n",
    "    except ValueError:\n",
    "        continue\n",
    "\n",
    "    annotation = annotations[entry.label]\n",
    "    record = {\n",
    "        \"category\": entry.category,\n",
    "        \"register\": annotation.register,\n",
    "        \"kind\": annotation.kind,\n",
    "        \"confidence\": annotation.confidence,\n",
    "        \"subject\": entry.label,\n",
    "        \"traffic\": float(np.median(values)),\n",
    "        \"shifted\": bool(outcome.shifts),\n",
    "    }\n",
    "    profiles.append(record)\n",
    "    for shift in outcome.shifts:\n",
    "        shifts.append({**record,\n",
    "                       \"date\": series.day(shift.index),\n",
    "                       \"lift\": shift.lift,\n",
    "                       \"before\": shift.level_before,\n",
    "                       \"after\": shift.level_after,\n",
    "                       \"identified\": shift.has_identified_parameters})\n",
    "\n",
    "print(f\"sujets analysés      : {len(profiles)}\")\n",
    "print(f\"changements détectés : {len(shifts)}\")\n",
    "print(f\"dont identifiés      : {sum(s['identified'] for s in shifts)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f1183357",
   "metadata": {},
   "source": [
    "## 3. Le taux de basculement : l'écart s'évanouit\n",
    "\n",
    "Le corpus étendu avait trouvé ici son seul écart significatif — avant de montrer qu'il suivait\n",
    "le trafic. L'annotation permet de trancher autrement : en corrigeant l'étiquette elle-même."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fde62949",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.165784Z",
     "iopub.status.busy": "2026-08-23T13:39:56.165707Z",
     "iopub.status.idle": "2026-08-23T13:39:56.169745Z",
     "shell.execute_reply": "2026-08-23T13:39:56.169354Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "étiquette de catégorie   accusation  19/220 =  8.6 %   découverte   6/220 =  2.7 %   RC =  3.37   p = 0.012\n",
      "annotation manuelle      accusation   7/147 =  4.8 %   découverte   6/118 =  5.1 %   RC =  0.93   p = 1.000\n",
      "(ni l'un ni l'autre)      12/175 =  6.9 %\n"
     ]
    }
   ],
   "source": [
    "def switching_rate(group):\n",
    "    hits = sum(item[\"shifted\"] for item in group)\n",
    "    return hits, len(group), (100.0 * hits / len(group) if group else 0.0)\n",
    "\n",
    "\n",
    "def compare(group_a, group_d):\n",
    "    ha, na, pa = switching_rate(group_a)\n",
    "    hd, nd, pd = switching_rate(group_d)\n",
    "    test = stats.fisher_exact([[ha, na - ha], [hd, nd - hd]])\n",
    "    return (f\"accusation {ha:3d}/{na:3d} = {pa:4.1f} %   découverte {hd:3d}/{nd:3d} = {pd:4.1f} %\"\n",
    "            f\"   RC = {test.statistic:5.2f}   p = {test.pvalue:.3f}\"), test\n",
    "\n",
    "\n",
    "for key, label in ((\"category\", \"étiquette de catégorie\"), (\"register\", \"annotation manuelle\")):\n",
    "    line, _ = compare([p for p in profiles if p[key] == \"accusation\"],\n",
    "                      [p for p in profiles if p[key] == \"discovery\"])\n",
    "    print(f\"{label:24s} {line}\")\n",
    "\n",
    "hits, total, percent = switching_rate([p for p in profiles if p[\"register\"] == \"neither\"])\n",
    "print(f\"{\"(ni l'un ni l'autre)\":24s} {hits:3d}/{total:3d} = {percent:4.1f} %\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b34464e9",
   "metadata": {},
   "source": [
    "### Lecture\n",
    "\n",
    "**Le rapport de cotes passe de 3,4 à 0,93.** L'écart ne s'atténue pas, il disparaît — et\n",
    "change même très légèrement de sens.\n",
    "\n",
    "Les sujets écartés comme « ni l'un ni l'autre » basculent à 6,9 %, c'est-à-dire **plus souvent\n",
    "que les deux registres**. Ils n'étaient pas du bruit inerte : ce sont eux qui portaient l'écart\n",
    "attribué au registre d'accusation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "346099e0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.170533Z",
     "iopub.status.busy": "2026-08-23T13:39:56.170463Z",
     "iopub.status.idle": "2026-08-23T13:39:56.177923Z",
     "shell.execute_reply": "2026-08-23T13:39:56.177572Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "catégorie   trafic médian : accusation   39.0   découverte   11.0   p = 4.8e-14\n",
      "annotation  trafic médian : accusation   36.0   découverte   26.5   p = 2.6e-03\n",
      "\n",
      "Le déséquilibre d'audience était lui-même un effet de l'étiquetage :\n",
      "les entrées de catalogue sans public gonflaient le registre « découverte ».\n",
      "\n",
      "strate ≥ 47 vues/jour   accusation   5/ 68 =  7.4 %   découverte   6/ 40 = 15.0 %   RC =  0.45   p = 0.323\n",
      "appariement sur le trafic : 116 paires, discordantes 5/4, McNemar p = 1.000\n"
     ]
    }
   ],
   "source": [
    "accusation = [p for p in profiles if p[\"register\"] == \"accusation\"]\n",
    "discovery = [p for p in profiles if p[\"register\"] == \"discovery\"]\n",
    "\n",
    "traffic = {\n",
    "    \"catégorie\": ([p[\"traffic\"] for p in profiles if p[\"category\"] == \"accusation\"],\n",
    "                  [p[\"traffic\"] for p in profiles if p[\"category\"] == \"discovery\"]),\n",
    "    \"annotation\": ([p[\"traffic\"] for p in accusation], [p[\"traffic\"] for p in discovery]),\n",
    "}\n",
    "for label, (values_a, values_d) in traffic.items():\n",
    "    test = stats.mannwhitneyu(values_a, values_d, alternative=\"two-sided\")\n",
    "    print(f\"{label:11s} trafic médian : accusation {np.median(values_a):6.1f}   \"\n",
    "          f\"découverte {np.median(values_d):6.1f}   p = {test.pvalue:.1e}\")\n",
    "\n",
    "print(\"\\nLe déséquilibre d'audience était lui-même un effet de l'étiquetage :\")\n",
    "print(\"les entrées de catalogue sans public gonflaient le registre « découverte ».\")\n",
    "\n",
    "THRESHOLD = 47.0\n",
    "line, _ = compare([p for p in accusation if p[\"traffic\"] >= THRESHOLD],\n",
    "                  [p for p in discovery if p[\"traffic\"] >= THRESHOLD])\n",
    "print(f\"\\nstrate ≥ {THRESHOLD:.0f} vues/jour   {line}\")\n",
    "\n",
    "pool = sorted(discovery, key=lambda item: item[\"traffic\"])\n",
    "pool_logs = np.log([item[\"traffic\"] for item in pool])\n",
    "used, pairs = set(), []\n",
    "for probe in sorted(accusation, key=lambda item: -item[\"traffic\"]):\n",
    "    candidates = [(abs(np.log(probe[\"traffic\"]) - pool_logs[i]), i)\n",
    "                  for i in range(len(pool)) if i not in used]\n",
    "    if not candidates:\n",
    "        break\n",
    "    distance, index = min(candidates)\n",
    "    if distance <= np.log(2.0):\n",
    "        used.add(index)\n",
    "        pairs.append((probe, pool[index]))\n",
    "\n",
    "only_a = sum(1 for x, y in pairs if x[\"shifted\"] and not y[\"shifted\"])\n",
    "only_d = sum(1 for x, y in pairs if y[\"shifted\"] and not x[\"shifted\"])\n",
    "mcnemar = stats.binomtest(only_a, only_a + only_d, 0.5) if only_a + only_d else None\n",
    "print(f\"appariement sur le trafic : {len(pairs)} paires, discordantes {only_a}/{only_d},\"\n",
    "      f\" McNemar p = {mcnemar.pvalue:.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4de4676f",
   "metadata": {},
   "source": [
    "## 4. La persistance : toujours nulle, et cette fois avec la puissance de le dire\n",
    "\n",
    "C'est le test principal, celui que le corpus pilote avait cru remporter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "4a24ad24",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.178637Z",
     "iopub.status.busy": "2026-08-23T13:39:56.178572Z",
     "iopub.status.idle": "2026-08-23T13:39:56.183369Z",
     "shell.execute_reply": "2026-08-23T13:39:56.182985Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "corpus                                   accusation     découverte        p\n",
      "--------------------------------------------------------------------------\n",
      "  pilote, 24 sujets choisis à la main   ×9.20 (n= 8)   ×2.90 (n= 6)    0.081\n",
      "  étendu, étiquette de catégorie       ×3.04 (n=21)   ×2.90 (n= 7)    0.533\n",
      "  étendu, registre annoté              ×3.04 (n= 7)   ×2.90 (n= 7)    0.902\n",
      "\n",
      "contrôles de sensibilité\n",
      "  sans les sujets contaminés           ×2.87 (n= 6)   ×2.90 (n= 7)    0.945\n",
      "  sans les annotations incertaines     ×3.06 (n= 6)   ×2.90 (n= 5)    0.931\n",
      "  strate de trafic comparable          ×2.69 (n= 5)   ×2.90 (n= 7)    0.876\n",
      "\n",
      "  (ni l'un ni l'autre)                 ×3.20 (n=14)\n"
     ]
    }
   ],
   "source": [
    "def persistence(source, label):\n",
    "    lifts_a = np.array([s[\"lift\"] for s in source if s[\"register\"] == \"accusation\"])\n",
    "    lifts_d = np.array([s[\"lift\"] for s in source if s[\"register\"] == \"discovery\"])\n",
    "    test = stats.mannwhitneyu(lifts_a, lifts_d, alternative=\"two-sided\")\n",
    "    left = f\"×{np.median(lifts_a):.2f} (n={lifts_a.size:2d})\"\n",
    "    right = f\"×{np.median(lifts_d):.2f} (n={lifts_d.size:2d})\"\n",
    "    print(f\"  {label:34s} {left:>14s} {right:>14s} {test.pvalue:8.3f}\")\n",
    "\n",
    "\n",
    "print(f\"{'corpus':36s} {'accusation':>14s} {'découverte':>14s} {'p':>8s}\")\n",
    "print(\"-\" * 74)\n",
    "print(f\"  {'pilote, 24 sujets choisis à la main':34s} {'×9.20 (n= 8)':>14s}\"\n",
    "      f\" {'×2.90 (n= 6)':>14s} {0.081:8.3f}\")\n",
    "print(f\"  {'étendu, étiquette de catégorie':34s} {'×3.04 (n=21)':>14s}\"\n",
    "      f\" {'×2.90 (n= 7)':>14s} {0.533:8.3f}\")\n",
    "persistence(shifts, \"étendu, registre annoté\")\n",
    "\n",
    "print(\"\\ncontrôles de sensibilité\")\n",
    "persistence([s for s in shifts if s[\"subject\"] not in CONTAMINATED],\n",
    "            \"sans les sujets contaminés\")\n",
    "persistence([s for s in shifts if s[\"confidence\"] == \"sure\"],\n",
    "            \"sans les annotations incertaines\")\n",
    "persistence([s for s in shifts if s[\"traffic\"] >= THRESHOLD],\n",
    "            \"strate de trafic comparable\")\n",
    "\n",
    "neither = np.array([s[\"lift\"] for s in shifts if s[\"register\"] == \"neither\"])\n",
    "label_other = f\"×{np.median(neither):.2f} (n={neither.size:2d})\"\n",
    "print(f\"\\n  {\"(ni l'un ni l'autre)\":34s} {label_other:>14s}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b260bcbe",
   "metadata": {},
   "source": [
    "### Un résultat nul n'a de valeur que si le test pouvait détecter quelque chose\n",
    "\n",
    "Quatorze observations, sept de chaque côté : il faut le dire avant qu'on ne l'objecte. Mais un\n",
    "test de Mann-Whitney à sept contre sept n'est pas aveugle — il détecte une séparation nette, et\n",
    "c'est précisément ce que le corpus pilote annonçait, avec des intervalles interquartiles\n",
    "[4,0 ; 14,9] contre [2,7 ; 3,2] presque disjoints."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d549d5b0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.184079Z",
     "iopub.status.busy": "2026-08-23T13:39:56.184015Z",
     "iopub.status.idle": "2026-08-23T13:39:56.187840Z",
     "shell.execute_reply": "2026-08-23T13:39:56.187463Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "p minimal atteignable à n = 7 contre 7 : 5.8e-04  (séparation complète)\n",
      "\n",
      "rangs de chevauchement            p\n",
      "------------------------------------\n",
      "            0                0.0006  détecté\n",
      "            1                0.0026  détecté\n",
      "            2                0.0048  détecté\n",
      "            3                0.0124  détecté\n",
      "            4                0.0400  détecté\n",
      "            5                0.1395  \n",
      "\n",
      "Un écart de l'ampleur annoncée par le corpus pilote aurait été détecté ici.\n",
      "Le résultat nul n'est donc pas un simple manque de puissance.\n"
     ]
    }
   ],
   "source": [
    "from math import comb\n",
    "\n",
    "print(f\"p minimal atteignable à n = 7 contre 7 : {2 / comb(14, 7):.1e}  (séparation complète)\\n\")\n",
    "print(f\"{'rangs de chevauchement':26s} {'p':>8s}\")\n",
    "print(\"-\" * 36)\n",
    "for overlap in range(6):\n",
    "    lower = np.arange(7, dtype=float)\n",
    "    upper = np.arange(7, dtype=float) + 7 - overlap\n",
    "    pvalue = stats.mannwhitneyu(upper, lower, alternative=\"two-sided\").pvalue\n",
    "    verdict = \"détecté\" if pvalue < 0.05 else \"\"\n",
    "    print(f\"{overlap:^26d} {pvalue:8.4f}  {verdict}\")\n",
    "\n",
    "print(\"\\nUn écart de l'ampleur annoncée par le corpus pilote aurait été détecté ici.\")\n",
    "print(\"Le résultat nul n'est donc pas un simple manque de puissance.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a76c402",
   "metadata": {},
   "source": [
    "## 5. Ce que l'annotation révèle du plan d'expérience\n",
    "\n",
    "C'est la découverte que ni le pilote ni le corpus étendu ne pouvaient faire, faute d'étiquette\n",
    "fiable : **une fois correctement étiquetés, les deux registres ne portent presque pas sur les\n",
    "mêmes types de sujets.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e6d3c43e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.188556Z",
     "iopub.status.busy": "2026-08-23T13:39:56.188488Z",
     "iopub.status.idle": "2026-08-23T13:39:56.191789Z",
     "shell.execute_reply": "2026-08-23T13:39:56.191397Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type             accusation    discovery      neither\n",
      "-------------------------------------------------------\n",
      "event                    58            6            1\n",
      "person                   13           39           12\n",
      "organisation             11            0           13\n",
      "work                      2            0           19\n",
      "concept                  63           15           21\n",
      "object                    0           58          109\n",
      "\n",
      "taux de basculement à type de sujet constant\n",
      "  event          accusation n= 58  découverte n=  6   — effectifs insuffisants pour comparer\n",
      "  person         accusation   3/ 13 = 23.1 %   découverte   3/ 39 =  7.7 %   RC =  3.60   p = 0.157\n",
      "  organisation   accusation n= 11  découverte n=  0   — effectifs insuffisants pour comparer\n",
      "  work           accusation n=  2  découverte n=  0   — effectifs insuffisants pour comparer\n",
      "  concept        accusation   0/ 63 =  0.0 %   découverte   0/ 15 =  0.0 %   RC =   nan   p = 1.000\n",
      "  object         accusation n=  0  découverte n= 58   — effectifs insuffisants pour comparer\n"
     ]
    }
   ],
   "source": [
    "kinds_by_register = {\n",
    "    register: Counter(a.kind for a in annotations.values() if a.register == register)\n",
    "    for register in REGISTERS\n",
    "}\n",
    "\n",
    "print(f\"{'type':14s}\" + \"\".join(f\"{r:>13s}\" for r in REGISTERS))\n",
    "print(\"-\" * 55)\n",
    "for kind in KINDS:\n",
    "    print(f\"{kind:14s}\" + \"\".join(f\"{kinds_by_register[r][kind]:13d}\" for r in REGISTERS))\n",
    "\n",
    "print(\"\\ntaux de basculement à type de sujet constant\")\n",
    "for kind in KINDS:\n",
    "    group_a = [p for p in accusation if p[\"kind\"] == kind]\n",
    "    group_d = [p for p in discovery if p[\"kind\"] == kind]\n",
    "    if len(group_a) >= 10 and len(group_d) >= 10:\n",
    "        line, _ = compare(group_a, group_d)\n",
    "        print(f\"  {kind:14s} {line}\")\n",
    "    else:\n",
    "        print(f\"  {kind:14s} accusation n={len(group_a):3d}  découverte n={len(group_d):3d}\"\n",
    "              f\"   — effectifs insuffisants pour comparer\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ef15fd8",
   "metadata": {},
   "source": [
    "### Lecture\n",
    "\n",
    "Le registre d'accusation est fait de **concepts et d'événements** ; celui de découverte,\n",
    "d'**objets et de personnes**. Un seul type de sujet — les personnes — se trouve des deux côtés\n",
    "en nombre suffisant pour une comparaison, et il n'y donne rien de concluant.\n",
    "\n",
    "La conséquence est structurelle et dépasse ce corpus : **une comparaison entre registres bâtie\n",
    "sur des catégories thématiques compare aussi, et peut-être surtout, des natures d'objets.** Un\n",
    "concept encyclopédique — « Corruption au Mexique » — n'a pas la dynamique d'attention d'un\n",
    "événement daté, indépendamment de toute charge émotionnelle. Le notebook 11 avait nommé ce\n",
    "confondant sans pouvoir le mesurer ; il est ici mesuré, et il est sévère.\n",
    "\n",
    "Un troisième protocole devra donc apparier sur le **type de sujet** autant que sur le trafic."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4bf5cbf1",
   "metadata": {},
   "source": [
    "## 6. Les basculements retenus\n",
    "\n",
    "Vingt-huit changements de régime, et la moitié d'entre eux portent sur des sujets qu'aucun des\n",
    "deux registres ne revendique."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e1a3a11b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.192531Z",
     "iopub.status.busy": "2026-08-23T13:39:56.192470Z",
     "iopub.status.idle": "2026-08-23T13:39:56.195399Z",
     "shell.execute_reply": "2026-08-23T13:39:56.194999Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "annoté       catég.  sujet                                    date                avant→après        \n",
      "----------------------------------------------------------------------------------------------------\n",
      "neither      accusa  Lil Tay                                  2020-06-27       290 →    2290   ×  7.9\n",
      "neither      accusa  Watch Dogs (video game)                  2020-12-06        57 →     358   ×  6.3  ← identifié\n",
      "neither      accusa  Million Dollar Extreme                   2016-07-06       111 →     631   ×  5.7\n",
      "neither      accusa  The Capture (TV series)                  2022-07-30       388 →    2133   ×  5.5\n",
      "discovery    discov  David Baker (biochemist)                 2024-09-24        61 →     330   ×  5.4\n",
      "accusation   accusa  Mossack Fonseca                          2019-10-14       286 →    1378   ×  4.8\n",
      "neither      accusa  Illuminati (game)                        2020-01-16       208 →     737   ×  3.5\n",
      "neither      accusa  Herbert Kickl                            2024-08-24       144 →     494   ×  3.4\n",
      "neither      accusa  One America News Network                 2020-03-13      1315 →    4425   ×  3.4\n",
      "accusation   accusa  Joshua Schulte                           2025-08-31        83 →     280   ×  3.4\n",
      "discovery    discov  OSIRIS-REx                               2016-06-12       141 →     467   ×  3.3\n",
      "accusation   accusa  Lavon Affair                             2018-12-22       184 →     566   ×  3.1\n",
      "discovery    discov  Luc Montagnier                           2021-05-12       389 →    1191   ×  3.1\n",
      "accusation   accusa  Lawrence E. King Jr.                     2025-07-01        79 →     240   ×  3.0\n",
      "neither      accusa  The Capture (TV series)                  2026-02-13      1160 →    3530   ×  3.0\n",
      "discovery    discov  Double Asteroid Redirection Test         2021-09-27       123 →     358   ×  2.9\n",
      "accusation   accusa  Enfield poltergeist                      2023-07-01       916 →    2464   ×  2.7\n",
      "neither      accusa  République (video game)                  2015-11-19       160 →     429   ×  2.7\n",
      "neither      accusa  Confederation Liberty and Independence   2023-05-26       264 →     648   ×  2.5\n",
      "discovery    discov  Vera C. Rubin Observatory                2025-05-22       235 →     568   ×  2.4\n",
      "neither      accusa  Personal life of Clint Eastwood          2022-08-12       442 →    1060   ×  2.4\n",
      "discovery    discov  Luis Walter Alvarez                      2023-06-27       227 →     538   ×  2.4\n",
      "accusation   accusa  Stephen Glass                            2019-08-07       201 →     467   ×  2.3\n",
      "neither      accusa  Bidzina Ivanishvili                      2020-06-26       150 →     345   ×  2.3\n",
      "accusation   accusa  Coverage of the Hillsborough disaster by 2023-12-28       122 →     268   ×  2.2\n",
      "neither      accusa  One America News Network                 2017-04-05       924 →    1951   ×  2.1\n",
      "discovery    discov  Luc Montagnier                           2020-02-26       181 →     378   ×  2.1\n",
      "neither      accusa  Korean Central News Agency               2017-06-11       101 →     203   ×  2.0\n",
      "\n",
      "Sur les cinq plus fortes élévations, 4 portent sur des sujets codés « ni l'un ni l'autre ».\n",
      "Ce sont exactement les cas que le notebook 11 avait cités comme suspects.\n"
     ]
    }
   ],
   "source": [
    "print(f\"{'annoté':12s} {'catég.':7s} {'sujet':40s} {'date':12s} {'avant→après':>18s} {'':>7s}\")\n",
    "print(\"-\" * 100)\n",
    "for shift in sorted(shifts, key=lambda item: -item[\"lift\"]):\n",
    "    flag = \"  ← identifié\" if shift[\"identified\"] else \"\"\n",
    "    print(f\"{shift['register']:12s} {shift['category'][:6]:7s} {shift['subject'][:40]:40s} \"\n",
    "          f\"{str(shift['date']):12s} {shift['before']:7.0f} → {shift['after']:7.0f}\"\n",
    "          f\"   ×{shift['lift']:5.1f}{flag}\")\n",
    "\n",
    "top = sorted(shifts, key=lambda item: -item[\"lift\"])[:5]\n",
    "print(f\"\\nSur les cinq plus fortes élévations, {sum(1 for s in top if s['register'] == 'neither')}\"\n",
    "      f\" portent sur des sujets codés « ni l'un ni l'autre ».\")\n",
    "print(\"Ce sont exactement les cas que le notebook 11 avait cités comme suspects.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "05a111f6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.196103Z",
     "iopub.status.busy": "2026-08-23T13:39:56.196037Z",
     "iopub.status.idle": "2026-08-23T13:39:56.625027Z",
     "shell.execute_reply": "2026-08-23T13:39:56.624582Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1380x936 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figure, axes = plt.subplots(2, 2, figsize=(11.5, 7.8))\n",
    "noise, rates, persist, design = axes.ravel()\n",
    "\n",
    "# (a) Le bruit d'étiquetage : décomposition de chaque catégorie.\n",
    "categories = [\"accusation\", \"discovery\"]\n",
    "colours = {\"accusation\": PALETTE[\"field\"], \"discovery\": PALETTE[\"remedy\"],\n",
    "           \"neither\": PALETTE[\"neutral\"]}\n",
    "names = {\"accusation\": \"accusation\", \"discovery\": \"découverte\", \"neither\": \"ni l'un ni l'autre\"}\n",
    "bottom = np.zeros(2)\n",
    "for register in REGISTERS:\n",
    "    heights = np.array([matrix[c][register] for c in categories], dtype=float)\n",
    "    noise.bar(np.arange(2), heights, 0.55, bottom=bottom, color=colours[register],\n",
    "              alpha=0.85, label=f\"annoté {names[register]}\")\n",
    "    for index, (height, base) in enumerate(zip(heights, bottom, strict=True)):\n",
    "        if height >= 20:\n",
    "            noise.text(index, base + height / 2, f\"{height:.0f}\", ha=\"center\", va=\"center\",\n",
    "                       fontsize=8.5, color=\"white\", fontweight=\"bold\")\n",
    "    bottom += heights\n",
    "noise.set_xticks(np.arange(2))\n",
    "noise.set_xticklabels([\"catégorie\\naccusation\", \"catégorie\\ndécouverte\"])\n",
    "noise.set_ylabel(\"nombre de sujets\")\n",
    "noise.set_title(f\"Le bruit d'étiquetage : {100 * agreement / len(entries):.0f} % d'accord seulement\",\n",
    "                fontsize=10)\n",
    "noise.set_ylim(0, 292)\n",
    "noise.legend(fontsize=7.5, loc=\"upper center\", ncol=1)\n",
    "\n",
    "# (b) Le taux de basculement, avant et après annotation.\n",
    "labels = [\"étiquette\\nde catégorie\", \"registre\\nannoté\"]\n",
    "values_a, values_d, pvalues = [], [], []\n",
    "for key in (\"category\", \"register\"):\n",
    "    group_a = [p for p in profiles if p[key] == \"accusation\"]\n",
    "    group_d = [p for p in profiles if p[key] == \"discovery\"]\n",
    "    ha, na, pa = switching_rate(group_a)\n",
    "    hd, nd, pd = switching_rate(group_d)\n",
    "    values_a.append(pa)\n",
    "    values_d.append(pd)\n",
    "    pvalues.append(stats.fisher_exact([[ha, na - ha], [hd, nd - hd]]).pvalue)\n",
    "positions = np.arange(2)\n",
    "rates.bar(positions - 0.18, values_a, 0.34, color=PALETTE[\"field\"], label=\"accusation\")\n",
    "rates.bar(positions + 0.18, values_d, 0.34, color=PALETTE[\"remedy\"], label=\"découverte\")\n",
    "for index, pvalue in enumerate(pvalues):\n",
    "    height = max(values_a[index], values_d[index])\n",
    "    rates.text(index, height + 0.5, f\"p = {pvalue:.3f}\", ha=\"center\", fontsize=8,\n",
    "               color=PALETTE[\"neutral\"])\n",
    "rates.set_xticks(positions)\n",
    "rates.set_xticklabels(labels)\n",
    "rates.set_ylabel(\"sujets ayant basculé  [%]\")\n",
    "rates.set_ylim(0, max(values_a + values_d) * 1.35)\n",
    "rates.set_title(\"L'écart de taux disparaît avec l'étiquette\", fontsize=10)\n",
    "rates.legend(fontsize=8)\n",
    "\n",
    "# (c) La persistance, par registre annoté.\n",
    "groups = [np.array([s[\"lift\"] for s in shifts if s[\"register\"] == r]) for r in REGISTERS]\n",
    "parts = persist.boxplot(groups, widths=0.5, patch_artist=True, showfliers=False)\n",
    "for patch, register in zip(parts[\"boxes\"], REGISTERS, strict=True):\n",
    "    patch.set_facecolor(colours[register])\n",
    "    patch.set_alpha(0.35)\n",
    "for index, (values, register) in enumerate(zip(groups, REGISTERS, strict=True), start=1):\n",
    "    jitter = np.linspace(-0.12, 0.12, values.size)\n",
    "    persist.scatter(index + jitter, values, s=22, color=colours[register], zorder=3)\n",
    "persist.set_xticks([1, 2, 3])\n",
    "persist.set_xticklabels([f\"{names[r]}\\n(n={g.size})\" for r, g in zip(REGISTERS, groups, strict=True)])\n",
    "persist.set_ylabel(\"élévation durable du régime  [×]\")\n",
    "test = stats.mannwhitneyu(groups[0], groups[1], alternative=\"two-sided\")\n",
    "persist.set_title(f\"La persistance ne diffère pas (p = {test.pvalue:.2f})\", fontsize=10)\n",
    "\n",
    "# (d) Le défaut de conception : composition en types de sujet.\n",
    "shown = [\"event\", \"concept\", \"person\", \"object\", \"organisation\", \"work\"]\n",
    "width = 0.38\n",
    "offsets = {\"accusation\": -width / 2, \"discovery\": width / 2}\n",
    "for register in (\"accusation\", \"discovery\"):\n",
    "    heights = [kinds_by_register[register][k] for k in shown]\n",
    "    design.barh(np.arange(len(shown)) + offsets[register], heights, width,\n",
    "                color=colours[register], label=names[register])\n",
    "design.set_yticks(np.arange(len(shown)))\n",
    "design.set_yticklabels([\"événement\", \"concept\", \"personne\", \"objet\", \"organisation\", \"œuvre\"])\n",
    "design.invert_yaxis()\n",
    "design.set_xlabel(\"nombre de sujets\")\n",
    "design.set_title(\"Les deux registres ne portent pas sur les mêmes objets\", fontsize=10)\n",
    "design.legend(fontsize=8)\n",
    "\n",
    "figure.suptitle(\"Annotation en aveugle : le bruit d'étiquetage portait l'écart\", fontsize=12)\n",
    "figure.tight_layout(rect=(0, 0, 1, 0.96))\n",
    "save_figure(figure, \"fig12_annotation\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6aa4018",
   "metadata": {},
   "source": [
    "## 7. Ce que le notebook établit\n",
    "\n",
    "**Le bruit d'étiquetage était massif.** Deux sujets sur cinq ne relèvent d'aucun des deux\n",
    "registres, et l'accord entre catégorie et annotation n'atteint que 59,5 %. La conjecture du\n",
    "notebook 11 est vérifiée, et son ampleur dépasse ce qu'il supposait.\n",
    "\n",
    "**Il portait la totalité de l'écart de taux de basculement.** Le rapport de cotes de 3,4\n",
    "($p = 0{,}012$) devient 0,93 ($p = 1{,}00$) une fois l'étiquette corrigée. Les sujets écartés\n",
    "basculent à 6,9 %, plus souvent que les deux registres : ce n'étaient pas des observations\n",
    "inertes, c'étaient elles qui produisaient l'écart.\n",
    "\n",
    "**La persistance reste nulle, et cette fois le test avait la puissance de conclure.** ×3,04\n",
    "contre ×2,90, $p = 0{,}90$, robuste au retrait des sujets contaminés, des annotations\n",
    "incertaines et des sujets à faible trafic. Un écart de l'ampleur annoncée par le corpus pilote\n",
    "aurait été détecté.\n",
    "\n",
    "> **La question laissée ouverte par le corpus étendu est tranchée : l'écart n'était pas dilué\n",
    "> par l'étiquetage, il n'existe pas.**\n",
    "\n",
    "**Et l'annotation expose un défaut de conception.** Correctement étiquetés, les deux registres\n",
    "ne portent presque pas sur les mêmes types de sujets — concepts et événements d'un côté,\n",
    "objets et personnes de l'autre. Une comparaison bâtie sur des catégories thématiques compare\n",
    "donc aussi des natures d'objets. C'est une limite du plan d'expérience, pas du résultat : elle\n",
    "ne ressuscite pas l'écart, elle indique ce qu'un quatrième protocole devrait contrôler.\n",
    "\n",
    "**Ce que le dispositif ne couvre pas.** L'annotateur est unique : il n'y a pas d'accord\n",
    "inter-juges, donc pas de mesure de la fiabilité du codage. La grille écrite et publiée est ce\n",
    "qui rend le travail réplicable, non ce qui prouve qu'il serait reproduit à l'identique."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5d6ced2",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 8. La réplication : deux codeurs indépendants\n",
    "\n",
    "La section précédente laissait une réserve de méthode, et une seule : **l'annotateur était\n",
    "unique**, donc rien ne mesurait la fiabilité du codage. Le corpus a donc été recodé par deux\n",
    "lecteurs indépendants du contexte, sous la grille identique, à partir du même matériau —\n",
    "présenté dans un ordre différent et **sans l'étiquette de catégorie**, pour qu'aucun codage ne\n",
    "puisse recopier celui qu'il sert à vérifier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5334b6bc",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.625818Z",
     "iopub.status.busy": "2026-08-23T13:39:56.625747Z",
     "iopub.status.idle": "2026-08-23T13:39:56.630119Z",
     "shell.execute_reply": "2026-08-23T13:39:56.629732Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "codeur             accusation    discovery      neither\n",
      "-------------------------------------------------------\n",
      "C1 — initial              147          118          175\n",
      "C2-A                      140          114          186\n",
      "C2-B                      141          109          190\n"
     ]
    }
   ],
   "source": [
    "from ide.annotation import cohen_kappa, consensus_registers, fleiss_kappa, load_replication\n",
    "\n",
    "replication = load_replication()\n",
    "codings = {\"C1 — initial\": annotations, **replication}\n",
    "\n",
    "print(f\"{'codeur':16s}\" + \"\".join(f\"{r:>13s}\" for r in REGISTERS))\n",
    "print(\"-\" * 55)\n",
    "for name, coding in codings.items():\n",
    "    tally = Counter(item.register for item in coding.values())\n",
    "    print(f\"{name:16s}\" + \"\".join(f\"{tally[r]:13d}\" for r in REGISTERS))\n",
    "\n",
    "titles = sorted(annotations)\n",
    "labels = {name: [coding[t].register for t in titles] for name, coding in codings.items()}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f58d8158",
   "metadata": {},
   "source": [
    "Les distributions marginales sont déjà voisines. Reste à savoir si l'accord porte sur les\n",
    "**mêmes sujets** — ce que seul un accord corrigé du hasard peut dire, sur un corpus dont 40 %\n",
    "relèvent d'une seule étiquette."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "81e4f3ea",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.630839Z",
     "iopub.status.busy": "2026-08-23T13:39:56.630771Z",
     "iopub.status.idle": "2026-08-23T13:39:56.634333Z",
     "shell.execute_reply": "2026-08-23T13:39:56.633950Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "paire                             accord brut   κ de Cohen\n",
      "----------------------------------------------------------\n",
      "C1 — initial  vs  C2-A                  93.6 %        0.903\n",
      "C1 — initial  vs  C2-B                  94.5 %        0.917\n",
      "C2-A  vs  C2-B                          96.4 %        0.944\n",
      "\n",
      "κ de Fleiss, les trois codeurs ensemble : 0.921\n",
      "unanimité sur 406/440 = 92.3 % des sujets\n"
     ]
    }
   ],
   "source": [
    "names = list(codings)\n",
    "print(f\"{'paire':32s} {'accord brut':>12s} {'κ de Cohen':>12s}\")\n",
    "print(\"-\" * 58)\n",
    "for i, first in enumerate(names):\n",
    "    for second in names[i + 1:]:\n",
    "        left, right = labels[first], labels[second]\n",
    "        raw = sum(1 for a, b in zip(left, right, strict=True) if a == b) / len(left)\n",
    "        print(f\"{first + '  vs  ' + second:32s} {100 * raw:11.1f} % {cohen_kappa(left, right):12.3f}\")\n",
    "\n",
    "three_way = fleiss_kappa(list(labels.values()))\n",
    "print(f\"\\nκ de Fleiss, les trois codeurs ensemble : {three_way:.3f}\")\n",
    "\n",
    "unanimous = sum(1 for t in titles if len({coding[t].register for coding in codings.values()}) == 1)\n",
    "print(f\"unanimité sur {unanimous}/{len(titles)} = {100 * unanimous / len(titles):.1f} % des sujets\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "604e5194",
   "metadata": {},
   "source": [
    "### Lecture, et la réserve qui compte\n",
    "\n",
    "Sur l'échelle de Landis et Koch, $\\kappa > 0{,}80$ se lit « accord presque parfait ». La grille\n",
    "est donc **reproductible** : une lecture fraîche des mêmes consignes, sans accès au premier\n",
    "codage ni aux résultats, redonne les mêmes étiquettes.\n",
    "\n",
    "!!! danger \"Ce que cet accord ne mesure pas\"\n",
    "    Les trois codeurs sont des instances du **même modèle de langue**. L'accord obtenu mesure\n",
    "    la reproductibilité de la **grille**, pas l'accord entre juges humains indépendants — et il\n",
    "    le surestime nécessairement, des instances d'un même modèle partageant leurs a priori. La\n",
    "    réserve d'un codage humain multiple subsiste entière ; ce qui a changé, c'est qu'on sait\n",
    "    désormais que les consignes écrites suffisent à produire un codage stable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "6c01d958",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.634977Z",
     "iopub.status.busy": "2026-08-23T13:39:56.634918Z",
     "iopub.status.idle": "2026-08-23T13:39:56.637677Z",
     "shell.execute_reply": "2026-08-23T13:39:56.637304Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34 sujets non unanimes, par nature du désaccord :\n",
      "\n",
      "   17  discovery / neither\n",
      "   16  accusation / neither\n",
      "    1  accusation / discovery  ← inverse les deux registres comparés\n",
      "\n",
      "Désaccords portant sur l'appartenance à un registre : 33\n",
      "Désaccords inversant accusation et découverte        : 1\n"
     ]
    }
   ],
   "source": [
    "disagreements = [\n",
    "    (t, tuple(sorted({coding[t].register for coding in codings.values()})))\n",
    "    for t in titles\n",
    "    if len({coding[t].register for coding in codings.values()}) > 1\n",
    "]\n",
    "patterns = Counter(pattern for _, pattern in disagreements)\n",
    "\n",
    "print(f\"{len(disagreements)} sujets non unanimes, par nature du désaccord :\\n\")\n",
    "for pattern, count in patterns.most_common():\n",
    "    swaps = \"  ← inverse les deux registres comparés\" if \"neither\" not in pattern else \"\"\n",
    "    print(f\"  {count:3d}  {' / '.join(pattern)}{swaps}\")\n",
    "\n",
    "crossing = sum(count for pattern, count in patterns.items() if \"neither\" not in pattern)\n",
    "print(f\"\\nDésaccords portant sur l'appartenance à un registre : {len(disagreements) - crossing}\")\n",
    "print(f\"Désaccords inversant accusation et découverte        : {crossing}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d4e631e",
   "metadata": {},
   "source": [
    "### Le désaccord est structuré, et il tombe au bon endroit\n",
    "\n",
    "Presque tous les désaccords opposent un registre à « ni l'un ni l'autre » — c'est-à-dire qu'ils\n",
    "portent sur **l'appartenance** d'un sujet à la comparaison, non sur le côté où le ranger. Un\n",
    "seul sujet sur 440 voit un codeur dire « accusation » là où un autre dit « découverte ».\n",
    "\n",
    "La conséquence est directe : l'ambiguïté résiduelle de la grille fait varier les **effectifs**\n",
    "de la comparaison, pas son **sens**. C'est la forme d'imprécision la moins dommageable qu'on\n",
    "pouvait espérer.\n",
    "\n",
    "Les cas litigieux sont d'ailleurs interprétables : ce sont des étoiles dont le chapeau ne dit\n",
    "pas si leur notabilité tient à une découverte, des dispositifs de physique sans annonce\n",
    "associée, et quelques affaires dont le chapeau ne rapporte pas la mise en cause. La règle 4 de\n",
    "la grille — l'objet de catalogue — est celle qui laisse le plus de latitude."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33545a55",
   "metadata": {},
   "source": [
    "## 9. Le résultat tient-il sous le codage consensuel ?\n",
    "\n",
    "C'est la seule question qui compte vraiment. Reprenons les deux tests avec, pour chaque sujet,\n",
    "le registre **majoritaire des trois codeurs**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "16acfee1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.638346Z",
     "iopub.status.busy": "2026-08-23T13:39:56.638282Z",
     "iopub.status.idle": "2026-08-23T13:39:56.642876Z",
     "shell.execute_reply": "2026-08-23T13:39:56.642459Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "registre consensuel : {'neither': 188, 'accusation': 141, 'discovery': 111}\n",
      "\n",
      "étiquette                      accusation       découverte     RC        p\n",
      "--------------------------------------------------------------------------\n",
      "catégorie                 19/220 =  8.6 %   6/220 =  2.7 %   3.37    0.012\n",
      "annotation C1              7/147 =  4.8 %   6/118 =  5.1 %   0.93    1.000\n",
      "consensus 3 codeurs        6/141 =  4.3 %   6/111 =  5.4 %   0.78    0.769\n"
     ]
    }
   ],
   "source": [
    "consensus = consensus_registers(list(codings.values()))\n",
    "print(\"registre consensuel :\", dict(Counter(consensus.values())))\n",
    "\n",
    "for record in profiles:\n",
    "    record[\"consensus\"] = consensus[record[\"subject\"]]\n",
    "for record in shifts:\n",
    "    record[\"consensus\"] = consensus[record[\"subject\"]]\n",
    "\n",
    "print(f\"\\n{'étiquette':24s} {'accusation':>16s} {'découverte':>16s} {'RC':>6s} {'p':>8s}\")\n",
    "print(\"-\" * 74)\n",
    "for key, label in ((\"category\", \"catégorie\"),\n",
    "                   (\"register\", \"annotation C1\"),\n",
    "                   (\"consensus\", \"consensus 3 codeurs\")):\n",
    "    group_a = [p for p in profiles if p[key] == \"accusation\"]\n",
    "    group_d = [p for p in profiles if p[key] == \"discovery\"]\n",
    "    ha, na, pa = switching_rate(group_a)\n",
    "    hd, nd, pd = switching_rate(group_d)\n",
    "    test = stats.fisher_exact([[ha, na - ha], [hd, nd - hd]])\n",
    "    print(f\"{label:24s} {f'{ha:3d}/{na:3d} = {pa:4.1f} %':>16s}\"\n",
    "          f\" {f'{hd:3d}/{nd:3d} = {pd:4.1f} %':>16s}\"\n",
    "          f\" {test.statistic:6.2f} {test.pvalue:8.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "aa9225fb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.643565Z",
     "iopub.status.busy": "2026-08-23T13:39:56.643497Z",
     "iopub.status.idle": "2026-08-23T13:39:56.646812Z",
     "shell.execute_reply": "2026-08-23T13:39:56.646502Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "étiquette                    accusation     découverte        p\n",
      "--------------------------------------------------------------\n",
      "annotation C1              ×3.04 (n= 7)   ×2.90 (n= 7)    0.902\n",
      "consensus 3 codeurs        ×3.06 (n= 6)   ×2.90 (n= 7)    0.836\n",
      "\n",
      "Les deux résultats sont inchangés. Le codage initial n'était pas un cas particulier.\n"
     ]
    }
   ],
   "source": [
    "print(f\"{'étiquette':24s} {'accusation':>14s} {'découverte':>14s} {'p':>8s}\")\n",
    "print(\"-\" * 62)\n",
    "for key, label in ((\"register\", \"annotation C1\"), (\"consensus\", \"consensus 3 codeurs\")):\n",
    "    lifts_a = np.array([s[\"lift\"] for s in shifts if s[key] == \"accusation\"])\n",
    "    lifts_d = np.array([s[\"lift\"] for s in shifts if s[key] == \"discovery\"])\n",
    "    test = stats.mannwhitneyu(lifts_a, lifts_d, alternative=\"two-sided\")\n",
    "    left = f\"×{np.median(lifts_a):.2f} (n={lifts_a.size:2d})\"\n",
    "    right = f\"×{np.median(lifts_d):.2f} (n={lifts_d.size:2d})\"\n",
    "    print(f\"{label:24s} {left:>14s} {right:>14s} {test.pvalue:8.3f}\")\n",
    "\n",
    "print(\"\\nLes deux résultats sont inchangés. Le codage initial n'était pas un cas particulier.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "94d2b7e6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-08-23T13:39:56.647496Z",
     "iopub.status.busy": "2026-08-23T13:39:56.647433Z",
     "iopub.status.idle": "2026-08-23T13:39:56.882188Z",
     "shell.execute_reply": "2026-08-23T13:39:56.881647Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x468 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figure, axes = plt.subplots(1, 3, figsize=(12.0, 3.9))\n",
    "agreement_panel, structure, robustness = axes\n",
    "\n",
    "# (a) Les accords deux à deux, et l'accord à trois.\n",
    "pairs, values = [], []\n",
    "for i, first in enumerate(names):\n",
    "    for second in names[i + 1:]:\n",
    "        pairs.append(f\"{first.split(' —')[0].split(' ')[0]}\\n{second}\")\n",
    "        values.append(cohen_kappa(labels[first], labels[second]))\n",
    "pairs.append(\"trois codeurs\\n(Fleiss)\")\n",
    "values.append(three_way)\n",
    "colours = [PALETTE[\"remedy\"]] * (len(values) - 1) + [PALETTE[\"order\"]]\n",
    "agreement_panel.bar(np.arange(len(values)), values, 0.6, color=colours, alpha=0.85)\n",
    "agreement_panel.axhline(0.8, color=PALETTE[\"neutral\"], linestyle=\"--\", linewidth=1.2)\n",
    "agreement_panel.text(-0.45, 1.10, \"κ = 0,80 — seuil de l'accord presque parfait\",\n",
    "                     fontsize=7.5, color=PALETTE[\"neutral\"])\n",
    "for index, value in enumerate(values):\n",
    "    agreement_panel.text(index, value + 0.02, f\"{value:.3f}\", ha=\"center\", fontsize=8)\n",
    "agreement_panel.set_xticks(np.arange(len(values)))\n",
    "agreement_panel.set_xticklabels(pairs, fontsize=7)\n",
    "agreement_panel.set_ylim(0, 1.20)\n",
    "agreement_panel.set_ylabel(\"κ\")\n",
    "agreement_panel.set_title(\"La grille est reproductible\", fontsize=10)\n",
    "\n",
    "# (b) Où porte le désaccord.\n",
    "kinds = [\"appartenance\\nà un registre\", \"inversion\\naccusation / découverte\"]\n",
    "counts = [len(disagreements) - crossing, crossing]\n",
    "structure.bar(kinds, counts, 0.5, color=[PALETTE[\"neutral\"], PALETTE[\"disorder\"]], alpha=0.85)\n",
    "for index, count in enumerate(counts):\n",
    "    structure.text(index, count + 0.6, str(count), ha=\"center\", fontsize=9)\n",
    "structure.set_ylabel(\"sujets non unanimes\")\n",
    "structure.set_ylim(0, max(counts) * 1.25)\n",
    "structure.set_title(f\"Le désaccord change l'effectif,\\npas le sens ({len(titles)} sujets)\",\n",
    "                    fontsize=10)\n",
    "\n",
    "# (c) Le résultat sous les trois étiquetages.\n",
    "schemes = [(\"category\", \"catégorie\"), (\"register\", \"annotation C1\"),\n",
    "           (\"consensus\", \"consensus\")]\n",
    "rates_a, rates_d, pvalues = [], [], []\n",
    "for key, _ in schemes:\n",
    "    ha, na, pa = switching_rate([p for p in profiles if p[key] == \"accusation\"])\n",
    "    hd, nd, pd = switching_rate([p for p in profiles if p[key] == \"discovery\"])\n",
    "    rates_a.append(pa)\n",
    "    rates_d.append(pd)\n",
    "    pvalues.append(stats.fisher_exact([[ha, na - ha], [hd, nd - hd]]).pvalue)\n",
    "positions = np.arange(len(schemes))\n",
    "robustness.bar(positions - 0.18, rates_a, 0.34, color=PALETTE[\"field\"], label=\"accusation\")\n",
    "robustness.bar(positions + 0.18, rates_d, 0.34, color=PALETTE[\"remedy\"], label=\"découverte\")\n",
    "for index, pvalue in enumerate(pvalues):\n",
    "    robustness.text(index, max(rates_a[index], rates_d[index]) + 0.35,\n",
    "                    f\"p = {pvalue:.3f}\", ha=\"center\", fontsize=7.5, color=PALETTE[\"neutral\"])\n",
    "robustness.set_xticks(positions)\n",
    "robustness.set_xticklabels([label for _, label in schemes], fontsize=8)\n",
    "robustness.set_ylabel(\"sujets ayant basculé  [%]\")\n",
    "robustness.set_ylim(0, max(rates_a + rates_d) * 1.35)\n",
    "robustness.set_title(\"Le résultat ne dépend pas du codeur\", fontsize=10)\n",
    "robustness.legend(fontsize=8)\n",
    "\n",
    "figure.suptitle(\"Réplication : κ de Fleiss = %.3f, et un résultat inchangé\" % three_way,\n",
    "                fontsize=12)\n",
    "figure.tight_layout(rect=(0, 0, 1, 0.93))\n",
    "save_figure(figure, \"fig12b_replication\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37aab989",
   "metadata": {},
   "source": [
    "## 10. Ce que la réplication ajoute\n",
    "\n",
    "**La grille est reproductible** — $\\kappa$ de Fleiss de 0,92, unanimité sur 92 % des sujets.\n",
    "Les consignes écrites suffisent à produire un codage stable, ce qui rend le travail\n",
    "réplicable par un tiers plutôt que seulement consultable.\n",
    "\n",
    "**Le désaccord résiduel tombe au bon endroit.** Un seul sujet sur 440 voit deux codeurs\n",
    "inverser les registres comparés ; tout le reste porte sur l'appartenance à la comparaison.\n",
    "L'imprécision fait varier des effectifs, pas un sens.\n",
    "\n",
    "**Et le résultat est inchangé** sous le codage consensuel : le taux de basculement reste sans\n",
    "écart, la persistance aussi. Le codage initial n'était pas un cas particulier.\n",
    "\n",
    "> **La dernière réserve de méthode identifiable sans juges humains est levée.** Ce qui subsiste\n",
    "> — que trois instances d'un même modèle ne valent pas trois juges indépendants — ne se lèvera\n",
    "> qu'avec des annotateurs humains, et ce n'est plus une question de calcul."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a3700ec8",
   "metadata": {},
   "source": [
    "## Pistes ouvertes\n",
    "\n",
    "1. **Faire annoter le même corpus par un second codeur**, à l'aveugle également, et publier le\n",
    "   $\\kappa$ de Cohen. C'est le complément direct et peu coûteux de ce travail.\n",
    "2. **Apparier sur le type de sujet** autant que sur le trafic, dès la construction du corpus.\n",
    "   Comparer des événements à des événements et des personnes à des personnes est désormais\n",
    "   une exigence mesurée, non une précaution théorique.\n",
    "3. **Abaisser le seuil de détection en agrégeant par semaine.** Vingt-huit basculements sur\n",
    "   440 sujets laissent le test principal à sept observations par registre ; c'est le facteur\n",
    "   limitant qui reste.\n",
    "4. **Chercher l'effet ailleurs que dans la persistance.** Trois quantités ont été testées —\n",
    "   taux d'amplification, persistance, taux de basculement — et aucune ne distingue les\n",
    "   registres. Si le mécanisme de la charge émotionnelle existe, il ne se lit pas dans la\n",
    "   dynamique d'attention agrégée d'une encyclopédie."
   ]
  }
 ],
 "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": "3.12.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}