{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 🤖 Notebook 04 — Modelo Baseline\n",
    "\n",
    "Entrenamos el modelo más simple posible que funcione bien:\n",
    "**Logistic Regression** sobre TF-IDF.\n",
    "Este es nuestro punto de referencia — cualquier modelo más complejo\n",
    "debe superar estas métricas para justificar su complejidad.\n",
    "\n",
    "**Usamos Logistic Regression como baseline:**\n",
    "- Funciona excepcionalmente bien con matrices TF-IDF sparse\n",
    "- Coeficientes interpretables: podemos ver exactamente qué palabras usa\n",
    "- Rápido de entrenar y difícil de sobreajustar con regularización\n",
    "- En NLP, LR bate a modelos más complejos sorprendentemente seguido\n",
    "\n",
    "### Alerta del notebook anterior\n",
    "El análisis de features detectó nombres propios (`stefan`, `peggy`, `hubbard`)\n",
    "con alto peso en la clase no-tóxica. Analizaremos si el modelo los aprende\n",
    "y tomaremos acción si es necesario.\n",
    "\n",
    "### Output\n",
    "- Métricas baseline: F1, Precision, Recall, ROC-AUC\n",
    "- Análisis de coeficientes (qué aprendió el modelo)\n",
    "- Análisis de errores (qué falla y por qué)\n",
    "- Experimento en MLflow: `04_baseline`\n",
    "- Modelo guardado: `models/lr_baseline.joblib`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 0. Imports y configuración"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "kernel\n"
     ]
    }
   ],
   "source": [
    "print(\"kernel\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/miraekang/proyectos/ai-nlp/.venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PROJECT_ROOT: /Users/miraekang/proyectos/ai-nlp\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "import yaml\n",
    "import json\n",
    "import joblib\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import mlflow\n",
    "import mlflow.sklearn\n",
    "from pathlib import Path\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split, StratifiedKFold, cross_validate\n",
    "from sklearn.metrics import (\n",
    "    classification_report, confusion_matrix,\n",
    "    f1_score, precision_score, recall_score,\n",
    "    roc_auc_score, RocCurveDisplay\n",
    ")\n",
    "from sklearn.pipeline import Pipeline\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "PROJECT_ROOT = Path.cwd().parent\n",
    "sys.path.insert(0, str(PROJECT_ROOT))\n",
    "\n",
    "plt.rcParams['figure.figsize'] = (12, 5)\n",
    "plt.rcParams['axes.spines.top'] = False\n",
    "plt.rcParams['axes.spines.right'] = False\n",
    "\n",
    "print(f'PROJECT_ROOT: {PROJECT_ROOT}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR config: {'C': 0.4, 'max_iter': 1000, 'class_weight': 'balanced', 'solver': 'lbfgs'}\n",
      "CV folds : 5 | random_state: 42\n"
     ]
    }
   ],
   "source": [
    "CONFIG_FEAT = PROJECT_ROOT / 'configs' / 'features.yaml'\n",
    "CONFIG_PIPE = PROJECT_ROOT / 'configs' / 'pipeline.yaml'\n",
    "CONFIG_MOD  = PROJECT_ROOT / 'configs' / 'models.yaml'\n",
    "\n",
    "with open(CONFIG_FEAT) as f: feat_cfg = yaml.safe_load(f)\n",
    "with open(CONFIG_PIPE) as f: pipe_cfg = yaml.safe_load(f)\n",
    "with open(CONFIG_MOD)  as f: mod_cfg  = yaml.safe_load(f)\n",
    "\n",
    "tfidf_cfg = feat_cfg['vectorization']['tfidf']\n",
    "lr_cfg    = mod_cfg['models']['logistic_regression']\n",
    "TARGET    = pipe_cfg['data']['target_binary']\n",
    "SUBLABELS = pipe_cfg['data']['target_multilabel']\n",
    "RAND      = pipe_cfg['pipeline']['random_state']\n",
    "TEST_SIZE = pipe_cfg['pipeline']['test_size']\n",
    "CV_FOLDS  = pipe_cfg['pipeline']['cv_folds']\n",
    "\n",
    "print('LR config:', lr_cfg)\n",
    "print(f'CV folds : {CV_FOLDS} | random_state: {RAND}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Carga de datos y split\n",
    "\n",
    "Mismo split que en vectorización — `random_state` fijo desde YAML\n",
    "garantiza que train/test son idénticos en todos los notebooks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train: 797 | Test: 200\n",
      "Train positivos: 367 (46.0%)\n",
      "Test positivos : 92 (46.0%)\n"
     ]
    }
   ],
   "source": [
    "PROCESSED = PROJECT_ROOT / 'data' / 'processed' / 'v2' / 'comments_preprocessed.csv'\n",
    "\n",
    "df = pd.read_csv(PROCESSED)\n",
    "df['clean_text'] = df['clean_text'].fillna('').astype(str)\n",
    "\n",
    "X = df['clean_text']\n",
    "y = df[TARGET]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=TEST_SIZE, random_state=RAND, stratify=y\n",
    ")\n",
    "\n",
    "print(f'Train: {len(X_train)} | Test: {len(X_test)}')\n",
    "print(f'Train positivos: {y_train.sum()} ({y_train.mean()*100:.1f}%)')\n",
    "print(f'Test positivos : {y_test.sum()} ({y_test.mean()*100:.1f}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Pipeline sklearn\n",
    "\n",
    "Encadenamos TF-IDF + Logistic Regression en un `sklearn.Pipeline`.\n",
    "\n",
    "### ¿Por qué Pipeline y no pasos separados?\n",
    "- En cross-validation, el Pipeline hace `fit` del TF-IDF **dentro** de cada fold\n",
    "  usando solo los datos de ese fold de entrenamiento\n",
    "- Si vectorizáramos fuera del Pipeline, el TF-IDF vería todos los folds → data leakage\n",
    "- Además, el Pipeline es directamente serializable con `joblib` y deployable\n",
    "\n",
    "### Parámetros de Logistic Regression\n",
    "- `C=1.0` → regularización L2 por defecto (penaliza coeficientes grandes)\n",
    "- `class_weight='balanced'` → compensa si hay desbalance entre clases\n",
    "- `max_iter=1000` → suficiente para convergencia con datos sparse\n",
    "- `solver='lbfgs'` → eficiente para problemas medianos con L2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pipeline creado:\n",
      "Pipeline(steps=[('tfidf',\n",
      "                 TfidfVectorizer(max_features=5000, min_df=3,\n",
      "                                 ngram_range=(1, 2), strip_accents='unicode',\n",
      "                                 sublinear_tf=True)),\n",
      "                ('clf',\n",
      "                 LogisticRegression(C=0.4, class_weight='balanced',\n",
      "                                    max_iter=1000, random_state=42))])\n"
     ]
    }
   ],
   "source": [
    "pipeline = Pipeline([\n",
    "    ('tfidf', TfidfVectorizer(\n",
    "        max_features = tfidf_cfg['max_features'],\n",
    "        ngram_range  = tuple(tfidf_cfg['ngram_range']),\n",
    "        sublinear_tf = tfidf_cfg['sublinear_tf'],\n",
    "        min_df       = tfidf_cfg['min_df'],\n",
    "        analyzer     = 'word',\n",
    "        strip_accents= 'unicode',\n",
    "    )),\n",
    "    ('clf', LogisticRegression(\n",
    "        C            = lr_cfg['C'],\n",
    "        max_iter     = lr_cfg['max_iter'],\n",
    "        class_weight = lr_cfg['class_weight'],\n",
    "        solver       = lr_cfg['solver'],\n",
    "        random_state = RAND,\n",
    "    ))\n",
    "])\n",
    "\n",
    "print('Pipeline creado:')\n",
    "print(pipeline)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Cross-Validation — evaluación real\n",
    "\n",
    "Con 1000 muestras un solo split train/test no es fiable.\n",
    "StratifiedKFold k=5 divide en 5 partes manteniendo la proporción de clases,\n",
    "entrena en 4 y valida en 1, rotando las partes.\n",
    "\n",
    "**Resultado:** 5 métricas → tomamos la media y la desviación estándar.\n",
    "La std alta indica que el modelo es inestable con este dataset pequeño.\n",
    "\n",
    "> La rúbrica exige diferencia train/test < 5 puntos porcentuales.\n",
    "> Calculamos ambas para verificarlo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ejecutando 5-fold cross-validation...\n",
      "Completado.\n",
      "\n",
      "Métrica                   Train        Val         Gap(pp)    Std Val\n",
      "----------------------------------------------------------------------\n",
      "  f1                     0.8908     0.7143       17.65 pp     0.0325 ⚠️ Alto gap\n",
      "  precision              0.8909     0.7156       17.53 pp     0.0320 ⚠️ Alto gap\n",
      "  recall                 0.8908     0.7151       17.57 pp     0.0324 ⚠️ Alto gap\n",
      "  roc_auc                0.9589     0.7834       17.56 pp     0.0353 ⚠️ Alto gap\n"
     ]
    }
   ],
   "source": [
    "cv = StratifiedKFold(n_splits=CV_FOLDS, shuffle=True, random_state=RAND)\n",
    "\n",
    "# Métricas que queremos en cada fold\n",
    "scoring = {\n",
    "    'f1'       : 'f1_weighted',\n",
    "    'precision': 'precision_weighted',\n",
    "    'recall'   : 'recall_weighted',\n",
    "    'roc_auc'  : 'roc_auc',\n",
    "}\n",
    "\n",
    "print(f'Ejecutando {CV_FOLDS}-fold cross-validation...')\n",
    "cv_results = cross_validate(\n",
    "    pipeline, X_train, y_train,\n",
    "    cv=cv,\n",
    "    scoring=scoring,\n",
    "    return_train_score=True,   # para detectar overfitting\n",
    "    n_jobs=-1\n",
    ")\n",
    "\n",
    "print('Completado.\\n')\n",
    "print(f\"{'Métrica':20} {'Train':>10} {'Val':>10} {'Gap(pp)':>15} {'Std Val':>10}\")\n",
    "print('-' * 70)\n",
    "\n",
    "for metric in ['f1', 'precision', 'recall', 'roc_auc']:\n",
    "    train_mean = cv_results[f'train_{metric}'].mean()\n",
    "    val_mean   = cv_results[f'test_{metric}'].mean()\n",
    "    val_std    = cv_results[f'test_{metric}'].std()\n",
    "    gap_pp = abs(train_mean - val_mean) * 100\n",
    "\n",
    "    flag       = '⚠️ Alto gap' if gap_pp > 5 else '✅ Estable'\n",
    "    print(f'  {metric:18} {train_mean:>10.4f} {val_mean:>10.4f} {gap_pp:>11.2f} pp {val_std:>10.4f} {flag:>10}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Guardado en reports/v2/09_cv_metrics.png\n"
     ]
    }
   ],
   "source": [
    "# Visualizacion de métricas por fold\n",
    "fig, axes = plt.subplots(1, 2, figsize=(13, 5))\n",
    "\n",
    "metrics = ['f1', 'precision', 'recall', 'roc_auc']\n",
    "colors_train = '#7F77DD'\n",
    "colors_val   = '#5DCAA5'\n",
    "\n",
    "# F1 por fold\n",
    "folds = range(1, CV_FOLDS + 1)\n",
    "axes[0].plot(folds, cv_results['train_f1'], 'o-', color=colors_train,\n",
    "             label='Train F1', linewidth=2, markersize=8)\n",
    "axes[0].plot(folds, cv_results['test_f1'], 's-', color=colors_val,\n",
    "             label='Val F1', linewidth=2, markersize=8)\n",
    "axes[0].axhline(cv_results['test_f1'].mean(), color=colors_val,\n",
    "                linestyle='--', alpha=0.5, label=f'Media val: {cv_results[\"test_f1\"].mean():.3f}')\n",
    "axes[0].set_title('F1-weighted por fold', fontweight='bold')\n",
    "axes[0].set_xlabel('Fold')\n",
    "axes[0].set_ylabel('F1 score')\n",
    "axes[0].set_ylim(0.5, 1.0)\n",
    "axes[0].legend()\n",
    "axes[0].set_xticks(list(folds))\n",
    "\n",
    "# Comparativa train vs val todas las métricas\n",
    "x = np.arange(len(metrics))\n",
    "width = 0.35\n",
    "train_means = [cv_results[f'train_{m}'].mean() for m in metrics]\n",
    "val_means   = [cv_results[f'test_{m}'].mean()  for m in metrics]\n",
    "val_stds    = [cv_results[f'test_{m}'].std()   for m in metrics]\n",
    "\n",
    "axes[1].bar(x - width/2, train_means, width, label='Train', color=colors_train, alpha=0.8)\n",
    "axes[1].bar(x + width/2, val_means,   width, label='Validación', color=colors_val,\n",
    "            alpha=0.8, yerr=val_stds, capsize=4)\n",
    "axes[1].set_title('Train vs Validación (media 5 folds)', fontweight='bold')\n",
    "axes[1].set_xticks(x)\n",
    "axes[1].set_xticklabels(['F1', 'Precision', 'Recall', 'ROC-AUC'])\n",
    "axes[1].set_ylim(0.5, 1.0)\n",
    "axes[1].legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(PROJECT_ROOT / 'reports' / 'v2' / '09_cv_metrics.png', dpi=150, bbox_inches='tight')\n",
    "plt.show()\n",
    "print('Guardado en reports/v2/09_cv_metrics.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Evaluación final en test\n",
    "\n",
    "Entrenamos el modelo en todo el train y evaluamos **una sola vez** en test.\n",
    "Este es el número que reportamos como resultado final del baseline.\n",
    "\n",
    "> ⚠️ El test set no se toca hasta este momento.\n",
    "> Usarlo antes contamina la evaluación final."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=======================================================\n",
      "RESULTADOS FINALES — LOGISTIC REGRESSION BASELINE\n",
      "=======================================================\n",
      "F1-weighted Train : 0.8708\n",
      "F1-weighted Test  : 0.7086\n",
      "Train-Test Gap    : 16.22 pp\n",
      "\n",
      "CV F1 Mean        : 0.7143\n",
      "CV-Test Gap       : 0.57 pp\n",
      "Rubrica CV/Test   : ✅ Estable\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "   No toxico       0.72      0.77      0.74       108\n",
      "      Toxico       0.70      0.64      0.67        92\n",
      "\n",
      "    accuracy                           0.71       200\n",
      "   macro avg       0.71      0.70      0.71       200\n",
      "weighted avg       0.71      0.71      0.71       200\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Entrenar en todo el train\n",
    "pipeline.fit(X_train, y_train)\n",
    "\n",
    "# Predecir en test\n",
    "y_pred       = pipeline.predict(X_test)\n",
    "y_pred_proba = pipeline.predict_proba(X_test)[:, 1]\n",
    "y_pred_train = pipeline.predict(X_train)\n",
    "\n",
    "# Métricas train vs test (chequeo de overfitting para rúbrica)\n",
    "f1_train = f1_score(y_train, y_pred_train, average='weighted')\n",
    "f1_test  = f1_score(y_test,  y_pred,       average='weighted')\n",
    "\n",
    "\n",
    "train_test_gap_pp = abs(f1_train - f1_test) * 100\n",
    "\n",
    "cv_f1_mean = cv_results['test_f1'].mean()\n",
    "cv_test_gap_pp = abs(cv_f1_mean - f1_test) * 100\n",
    "\n",
    "print('=' * 55)\n",
    "print('RESULTADOS FINALES — LOGISTIC REGRESSION BASELINE')\n",
    "print('=' * 55)\n",
    "\n",
    "print(f'F1-weighted Train : {f1_train:.4f}')\n",
    "print(f'F1-weighted Test  : {f1_test:.4f}')\n",
    "print(f'Train-Test Gap    : {train_test_gap_pp:.2f} pp')\n",
    "\n",
    "\n",
    "print()\n",
    "\n",
    "print(f'CV F1 Mean        : {cv_f1_mean:.4f}')\n",
    "print(f'CV-Test Gap       : {cv_test_gap_pp:.2f} pp')\n",
    "\n",
    "status = '✅ Estable' if cv_test_gap_pp < 5 else '⚠️ Alto gap'\n",
    "print(f'Rubrica CV/Test   : {status}')\n",
    "\n",
    "print()\n",
    "print(classification_report(y_test, y_pred, target_names=['No toxico', 'Toxico']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x500 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Guardado en reports/v2/10_baseline_confusion_roc.png\n"
     ]
    }
   ],
   "source": [
    "# Matriz de confusion + Curva ROC\n",
    "fig, axes = plt.subplots(1, 2, figsize=(13, 5))\n",
    "\n",
    "# Matriz de confusion\n",
    "cm = confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', ax=axes[0],\n",
    "            xticklabels=['No toxico','Toxico'],\n",
    "            yticklabels=['No toxico','Toxico'],\n",
    "            linewidths=0.5)\n",
    "axes[0].set_title('Matriz de Confusión — Test Set', fontweight='bold')\n",
    "axes[0].set_ylabel('Real')\n",
    "axes[0].set_xlabel('Predicho')\n",
    "\n",
    "# Curva ROC\n",
    "RocCurveDisplay.from_predictions(y_test, y_pred_proba, ax=axes[1],\n",
    "                                  color='#7F77DD', name='LR Baseline')\n",
    "axes[1].plot([0,1],[0,1],'--', color='gray', alpha=0.5, label='Random')\n",
    "axes[1].set_title('Curva ROC — Test Set', fontweight='bold')\n",
    "axes[1].legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(PROJECT_ROOT / 'reports' / 'v2' / '10_baseline_confusion_roc.png',\n",
    "            dpi=150, bbox_inches='tight')\n",
    "plt.show()\n",
    "print('Guardado en reports/v2/10_baseline_confusion_roc.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Análisis de coeficientes — ¿qué aprendió el modelo?\n",
    "\n",
    "Los coeficientes de Logistic Regression indican el peso de cada feature.\n",
    "Coeficiente positivo alto → predice TÓXICO.\n",
    "Coeficiente negativo alto → predice NO TÓXICO.\n",
    "\n",
    "**Buscamos específicamente** si el modelo aprendió nombres propios\n",
    "(`stefan`, `peggy`, `hubbard`) en lugar de lenguaje tóxico real.\n",
    "Si los vemos en top features → añadirlos a custom stopwords y re-ejecutar."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TOP 20 — Palabras que MAS predicen TOXICO (coef positivo):\n",
      "feature  coeficiente\n",
      "   fuck     1.582224\n",
      "    run     1.215538\n",
      "  idiot     1.190952\n",
      "   shit     1.155516\n",
      " stupid     0.932082\n",
      "    ass     0.883411\n",
      "  shoot     0.814162\n",
      "   thug     0.737356\n",
      "  bitch     0.731911\n",
      "  white     0.682751\n",
      "   dumb     0.657736\n",
      "  black     0.593531\n",
      "fucking     0.593391\n",
      "   cunt     0.522571\n",
      "    job     0.499430\n",
      "    cnn     0.488097\n",
      "  every     0.474585\n",
      "   kill     0.472122\n",
      "  masri     0.467155\n",
      "  bunch     0.452629\n",
      "\n",
      "TOP 20 — Palabras que MAS predicen NO TOXICO (coef negativo):\n",
      " feature  coeficiente\n",
      "   peggy    -0.810510\n",
      "   truth    -0.660594\n",
      "   thank    -0.649197\n",
      "  stefan    -0.631774\n",
      "     say    -0.611535\n",
      "   woman    -0.521453\n",
      "     rap    -0.482543\n",
      "    move    -0.481717\n",
      "    tell    -0.476465\n",
      "     put    -0.463828\n",
      "    much    -0.460757\n",
      "    good    -0.446983\n",
      "  pretty    -0.442561\n",
      "ferguson    -0.441739\n",
      " hubbard    -0.411202\n",
      "   cigar    -0.410384\n",
      "    side    -0.369870\n",
      "   could    -0.352536\n",
      "    fair    -0.349018\n",
      "    well    -0.349007\n"
     ]
    }
   ],
   "source": [
    "# Extraer coeficientes del clasificador dentro del pipeline\n",
    "lr_model    = pipeline.named_steps['clf']\n",
    "tfidf_model = pipeline.named_steps['tfidf']\n",
    "feature_names = tfidf_model.get_feature_names_out()\n",
    "coef = lr_model.coef_[0]\n",
    "\n",
    "coef_df = pd.DataFrame({\n",
    "    'feature'    : feature_names,\n",
    "    'coeficiente': coef\n",
    "}).sort_values('coeficiente', ascending=False)\n",
    "\n",
    "print('TOP 20 — Palabras que MAS predicen TOXICO (coef positivo):')\n",
    "print(coef_df.head(20).to_string(index=False))\n",
    "print()\n",
    "print('TOP 20 — Palabras que MAS predicen NO TOXICO (coef negativo):')\n",
    "print(coef_df.tail(20).sort_values('coeficiente').to_string(index=False))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Guardado en reports/v2/11_lr_coeficientes.png\n",
      "\n",
      "⚠️  Nombres propios / términos sospechosos en el modelo:\n",
      " feature  coeficiente\n",
      "     cnn     0.488097\n",
      "   cigar    -0.410384\n",
      " hubbard    -0.411202\n",
      "ferguson    -0.441739\n",
      "  stefan    -0.631774\n",
      "   peggy    -0.810510\n",
      "\n",
      "Considerar: añadir a custom_stopwords en preprocessing y re-ejecutar\n"
     ]
    }
   ],
   "source": [
    "# Visualizacion de coeficientes\n",
    "fig, axes = plt.subplots(1, 2, figsize=(15, 8))\n",
    "\n",
    "top_pos = coef_df.head(25)\n",
    "top_neg = coef_df.tail(25).sort_values('coeficiente')\n",
    "\n",
    "axes[0].barh(top_pos['feature'], top_pos['coeficiente'],\n",
    "             color='#E8593C', edgecolor='white')\n",
    "axes[0].set_title('Top 25 — predicen TOXICO', fontweight='bold', color='#993C1D')\n",
    "axes[0].set_xlabel('Coeficiente LR')\n",
    "axes[0].invert_yaxis()\n",
    "\n",
    "axes[1].barh(top_neg['feature'], top_neg['coeficiente'].abs(),\n",
    "             color='#5DCAA5', edgecolor='white')\n",
    "axes[1].set_title('Top 25 — predicen NO TOXICO', fontweight='bold', color='#0F6E56')\n",
    "axes[1].set_xlabel('|Coeficiente LR|')\n",
    "axes[1].invert_yaxis()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(PROJECT_ROOT / 'reports' / 'v2' / '11_lr_coeficientes.png',\n",
    "            dpi=150, bbox_inches='tight')\n",
    "plt.show()\n",
    "print('Guardado en reports/v2/11_lr_coeficientes.png')\n",
    "\n",
    "# Alerta: ¿hay nombres propios en top features?\n",
    "suspicious = ['stefan','peggy','hubbard','ferguson','cigar','cnn']\n",
    "found = coef_df[coef_df['feature'].isin(suspicious)]\n",
    "if len(found) > 0:\n",
    "    print('\\n⚠️  Nombres propios / términos sospechosos en el modelo:')\n",
    "    print(found.to_string(index=False))\n",
    "    print('\\nConsiderar: añadir a custom_stopwords en preprocessing y re-ejecutar')\n",
    "else:\n",
    "    print('\\n✅ No se detectaron nombres propios en top features')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Análisis de errores\n",
    "\n",
    "Los errores del modelo son tan importantes como sus aciertos.\n",
    "Entender qué falla nos guía para mejorar el preprocesamiento,\n",
    "cambiar el vectorizador o ajustar hiperparámetros.\n",
    "\n",
    "**Falso Positivo (FP):** comentario NO tóxico clasificado como TÓXICO → censura incorrecta.\n",
    "**Falso Negativo (FN):** comentario TÓXICO clasificado como NO tóxico → hate speech no detectado.\n",
    "\n",
    "Para la app de YouTube: los FN son más dañinos (dejan pasar hate speech)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total test   : 200\n",
      "Correctos    : 142 (71.0%)\n",
      "Falsos Pos   : 25 — no toxicos clasificados como toxicos\n",
      "Falsos Neg   : 33 — toxicos NO detectados\n",
      "\n",
      "FALSOS POSITIVOS (el modelo censura incorrectamente):\n",
      "-----------------------------------------------------------------\n",
      "  Prob: 0.671 | Who the fuck has time to be on the freeway???\n",
      "  Limpio: fuck time freeway\n",
      "\n",
      "  Prob: 0.601 | If that cop shot a white guy, there would not be any riots. But since he shot a black guy, everyone \n",
      "  Limpio: cop shoot white guy would riot since shoot black guy everyone seem think black l\n",
      "\n",
      "  Prob: 0.591 | 7:28 Lost my shit to that GUY. lmao\n",
      "  Limpio: lose shit guy lmao\n",
      "\n",
      "  Prob: 0.587 | police shot 1 black guy , & then the blacks riot ,!! when a black shoot a black , what happens ?? ,,\n",
      "  Limpio: police shoot black guy black riot black shoot black happen\n",
      "\n",
      "  Prob: 0.584 | people get shot everyday, how come this is the only one thats getting attention\n",
      "  Limpio: people get shoot everyday come one get attention\n",
      "\n",
      "FALSOS NEGATIVOS (hate speech que se escapa):\n",
      "-----------------------------------------------------------------\n",
      "  Prob: 0.394 | AMEN!!!, Peggy Hubbard!! Color makes no difference, it's the person on the inside. The innocent nine\n",
      "  Limpio: amen peggy hubbard color make difference person inside innocent nine year old li\n",
      "\n",
      "  Prob: 0.406 | We all know with the chimpers there is no compassion in their mud pumping hearts.\n",
      "  Limpio: know chimper compassion mud pump heart\n",
      "\n",
      "  Prob: 0.416 | This is so far off base that you can't even call it propaganda! BS!\n",
      "  Limpio: far base even call propaganda bs\n",
      "\n",
      "  Prob: 0.417 | That video of him pushing that clerk around really made me lose any sympathy for him. You can defini\n",
      "  Limpio: push clerk around really make I lose sympathy definitely tell guy likely dick.bu\n",
      "\n",
      "  Prob: 0.430 | Let's get this straight and present the facts as the mass media should have from the beginning. 18 y\n",
      "  Limpio: let get straight present fact mass media beginning yr old legal aspect adult rob\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Reconstruir DataFrame de test para analisis\n",
    "test_df = df.iloc[X_test.index].copy()\n",
    "test_df['predicted'] = y_pred\n",
    "test_df['proba_toxic'] = y_pred_proba\n",
    "test_df['correct'] = test_df[TARGET] == test_df['predicted']\n",
    "\n",
    "# Falsos positivos y negativos\n",
    "fp = test_df[(test_df[TARGET] == False) & (test_df['predicted'] == True)]\n",
    "fn = test_df[(test_df[TARGET] == True)  & (test_df['predicted'] == False)]\n",
    "\n",
    "print(f'Total test   : {len(test_df)}')\n",
    "print(f'Correctos    : {test_df[\"correct\"].sum()} ({test_df[\"correct\"].mean()*100:.1f}%)')\n",
    "print(f'Falsos Pos   : {len(fp)} — no toxicos clasificados como toxicos')\n",
    "print(f'Falsos Neg   : {len(fn)} — toxicos NO detectados')\n",
    "print()\n",
    "\n",
    "print('FALSOS POSITIVOS (el modelo censura incorrectamente):')\n",
    "print('-' * 65)\n",
    "for _, row in fp.nlargest(5, 'proba_toxic').iterrows():\n",
    "    print(f'  Prob: {row[\"proba_toxic\"]:.3f} | {row[\"Text\"][:100]}')\n",
    "    print(f'  Limpio: {row[\"clean_text\"][:80]}')\n",
    "    print()\n",
    "\n",
    "print('FALSOS NEGATIVOS (hate speech que se escapa):')\n",
    "print('-' * 65)\n",
    "for _, row in fn.nsmallest(5, 'proba_toxic').iterrows():\n",
    "    print(f'  Prob: {row[\"proba_toxic\"]:.3f} | {row[\"Text\"][:100]}')\n",
    "    print(f'  Limpio: {row[\"clean_text\"][:80]}')\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Registro en MLflow\n",
    "\n",
    "Guardamos todo en MLflow para poder comparar este baseline\n",
    "contra los modelos ensemble del siguiente notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2026/05/25 02:03:53 INFO mlflow.tracking.fluent: Experiment with name 'Youtube_project_experiment' does not exist. Creating a new experiment.\n",
      "2026/05/25 02:03:53 WARNING mlflow.models.model: `artifact_path` is deprecated. Please use `name` instead.\n",
      "2026/05/25 02:03:53 WARNING mlflow.sklearn: Saving scikit-learn models in the pickle or cloudpickle format requires exercising caution because these formats rely on Python's object serialization mechanism, which can execute arbitrary code during deserialization. The recommended safe alternative is the 'skops' format. For more information, see: https://scikit-learn.org/stable/model_persistence.html\n",
      "2026/05/25 02:03:55 WARNING mlflow.utils.environment: Failed to resolve installed pip version. ``pip`` will be added to conda.yaml environment spec without a version specifier.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MLflow registrado\n",
      "  Run ID     : 690a2205048349e49d50abe0ebafb9fe\n",
      "  Experimento: Youtube_project_experiment\n",
      "  Run name   : lr_baseline\n"
     ]
    }
   ],
   "source": [
    "MLFLOW_DIR = PROJECT_ROOT / 'mlruns'\n",
    "mlflow.set_tracking_uri(f'file://{MLFLOW_DIR}')\n",
    "mlflow.set_experiment('Youtube_project_experiment')\n",
    "\n",
    "with mlflow.start_run(run_name='lr_baseline'):\n",
    "\n",
    "    # Params\n",
    "    mlflow.log_param('model',          'LogisticRegression')\n",
    "    mlflow.log_param('C',              lr_cfg['C'])\n",
    "    mlflow.log_param('max_iter',       lr_cfg['max_iter'])\n",
    "    mlflow.log_param('class_weight',   lr_cfg['class_weight'])\n",
    "    mlflow.log_param('solver',         lr_cfg['solver'])\n",
    "    mlflow.log_param('tfidf_features', tfidf_cfg['max_features'])\n",
    "    mlflow.log_param('ngram_range',    str(tuple(tfidf_cfg['ngram_range'])))\n",
    "    mlflow.log_param('cv_folds',       CV_FOLDS)\n",
    "    mlflow.log_param('vectorizer',     'tfidf')\n",
    "\n",
    "    # Metrics CV\n",
    "    mlflow.log_metric('cv_f1_mean',        cv_results['test_f1'].mean())\n",
    "    mlflow.log_metric('cv_f1_std',         cv_results['test_f1'].std())\n",
    "    mlflow.log_metric('cv_roc_auc_mean',   cv_results['test_roc_auc'].mean())\n",
    "    mlflow.log_metric('cv_precision_mean', cv_results['test_precision'].mean())\n",
    "    mlflow.log_metric('cv_recall_mean',    cv_results['test_recall'].mean())\n",
    "\n",
    "    # Metrics test final\n",
    "    mlflow.log_metric('test_f1',         f1_test)\n",
    "    mlflow.log_metric('train_f1',        f1_train)\n",
    "    # Gap clásico Train vs Test\n",
    "    mlflow.log_metric('train_test_gap_pp', round(train_test_gap_pp, 2))\n",
    "\n",
    "    # Consistencia CV vs Test\n",
    "    mlflow.log_metric('cv_f1_mean', round(cv_f1_mean, 4))\n",
    "    mlflow.log_metric('cv_test_gap_pp', round(cv_test_gap_pp, 2))\n",
    "    mlflow.log_metric('test_roc_auc',    roc_auc_score(y_test, y_pred_proba))\n",
    "    \n",
    "    mlflow.log_metric('fp_count',        len(fp))\n",
    "    mlflow.log_metric('fn_count',        len(fn))\n",
    "\n",
    "    # Modelo\n",
    "    mlflow.sklearn.log_model(pipeline, 'lr_baseline_pipeline')\n",
    "\n",
    "    # Artefactos\n",
    "    for png in ['09_cv_metrics','10_baseline_confusion_roc','11_lr_coeficientes']:\n",
    "        mlflow.log_artifact(str(PROJECT_ROOT / 'reports' / 'v2' / f'{png}.png'))\n",
    "\n",
    "    print(f'MLflow registrado')\n",
    "    print(f'  Run ID     : {mlflow.active_run().info.run_id}')\n",
    "    print(f'  Experimento: Youtube_project_experiment')\n",
    "    print(f'  Run name   : lr_baseline')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. Guardar modelo\n",
    "\n",
    "Guardamos el pipeline completo (TF-IDF + LR) en `models/`.\n",
    "El pipeline incluye el vectorizador entrenado, por lo que basta\n",
    "con cargar este archivo y llamar `.predict()` sobre texto limpio."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo guardado en: /Users/miraekang/proyectos/ai-nlp/models/lr_baseline.joblib\n",
      "\n",
      "Verificacion modelo cargado:\n",
      "  [TOXICO 0.74] you are such a stupid thug get out\n",
      "  [NO TOXICO 0.43] I think the police should be more transparent\n",
      "  [TOXICO 0.59] black people are criminal thugs\n"
     ]
    }
   ],
   "source": [
    "MODELS_DIR = PROJECT_ROOT / 'models'\n",
    "MODELS_DIR.mkdir(exist_ok=True)\n",
    "\n",
    "model_path = MODELS_DIR / 'lr_baseline.joblib'\n",
    "joblib.dump(pipeline, model_path)\n",
    "print(f'Modelo guardado en: {model_path}')\n",
    "\n",
    "# Verificacion: cargar y predecir\n",
    "loaded = joblib.load(model_path)\n",
    "test_comments = [\n",
    "    'you are such a stupid thug get out',\n",
    "    'I think the police should be more transparent',\n",
    "    'black people are criminal thugs',\n",
    "]\n",
    "preds = loaded.predict(test_comments)\n",
    "probas = loaded.predict_proba(test_comments)[:, 1]\n",
    "print('\\nVerificacion modelo cargado:')\n",
    "for comment, pred, prob in zip(test_comments, preds, probas):\n",
    "    label = 'TOXICO' if pred else 'NO TOXICO'\n",
    "    print(f'  [{label} {prob:.2f}] {comment[:60]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9. Conclusiones y decisiones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "CONCLUSIONES — BASELINE\n",
      "=======================================================\n",
      "Modelo : Logistic Regression + TF-IDF\n",
      "Dataset: 800 train / 200 test / 5-fold CV\n",
      "\n",
      "Metricas CV (media 5 folds):\n",
      "  F1-weighted  : 0.7143\n",
      "  ROC-AUC      : 0.7834\n",
      "\n",
      "Metricas test final:\n",
      "  F1-weighted  : 0.7086\n",
      "  Train F1     : 0.8708\n",
      "\n",
      "Gaps:\n",
      "  Train-Test Gap : 16.22 pp\n",
      "  CV-Test Gap    : 0.57 pp\n",
      "  Rubrica CV/Test: ✅ Estable\n",
      "\n",
      "Errores:\n",
      "  False Positives : 25\n",
      "  False Negatives : 33\n",
      "\n",
      "Coeficientes sospechosos detectados:\n",
      "  Revisar si stefan/peggy/hubbard aparecen en top coef\n",
      "  Si aparecen -> agregar a custom_stopwords\n",
      "  y re-ejecutar preprocessing\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "f1_cv = cv_results['test_f1'].mean()\n",
    "roc_cv = cv_results['test_roc_auc'].mean()\n",
    "\n",
    "train_test_gap_pp = abs(f1_train - f1_test) * 100\n",
    "cv_test_gap_pp = abs(f1_cv - f1_test) * 100\n",
    "\n",
    "cv_status = '✅ Estable' if cv_test_gap_pp < 5 else '⚠️ Alto gap'\n",
    "\n",
    "print(f\"\"\"\n",
    "CONCLUSIONES — BASELINE\n",
    "=======================================================\n",
    "Modelo : Logistic Regression + TF-IDF\n",
    "Dataset: 800 train / 200 test / {CV_FOLDS}-fold CV\n",
    "\n",
    "Metricas CV (media {CV_FOLDS} folds):\n",
    "  F1-weighted  : {f1_cv:.4f}\n",
    "  ROC-AUC      : {roc_cv:.4f}\n",
    "\n",
    "Metricas test final:\n",
    "  F1-weighted  : {f1_test:.4f}\n",
    "  Train F1     : {f1_train:.4f}\n",
    "\n",
    "Gaps:\n",
    "  Train-Test Gap : {train_test_gap_pp:.2f} pp\n",
    "  CV-Test Gap    : {cv_test_gap_pp:.2f} pp\n",
    "  Rubrica CV/Test: {cv_status}\n",
    "\n",
    "Errores:\n",
    "  False Positives : {len(fp)}\n",
    "  False Negatives : {len(fn)}\n",
    "\n",
    "Coeficientes sospechosos detectados:\n",
    "  Revisar si stefan/peggy/hubbard aparecen en top coef\n",
    "  Si aparecen -> agregar a custom_stopwords\n",
    "  y re-ejecutar preprocessing\n",
    "\n",
    "\"\"\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}