{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import os\n",
        "import json\n",
        "from pathlib import Path\n",
        "import matplotlib.pyplot as plt\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "BASE_DIR = Path('.')\n",
        "ART_DIR = BASE_DIR / 'artifacts' / 'py'\n",
        "FIG_DIR = ART_DIR / 'figures'\n",
        "TAB_DIR = ART_DIR / 'tables'\n",
        "\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "TAB_DIR.mkdir(parents=True, exist_ok=True)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "df_food = pd.read_csv('clean_food_products.csv')\n",
        "df_reviews = pd.read_csv('synthetic_food_reviews.csv')\n",
        "\n",
        "print(df_food.shape)\n",
        "print(df_reviews.shape)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# KPI calculation\n",
        "kpis = {\n",
        "    'n_products': int(len(df_food)),\n",
        "    'avg_calories_per_100g': round(float(df_food['energy-kcal_100g'].mean()), 2),\n",
        "    'healthy_count': int((df_food['health_label'] == 'healthy').sum()),\n",
        "    'unhealthy_count': int((df_food['health_label'] == 'unhealthy').sum())\n",
        "}\n",
        "\n",
        "with open(TAB_DIR / 'kpis.json', 'w') as f:\n",
        "    json.dump(kpis, f, indent=2)\n",
        "\n",
        "kpis\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Save main dashboard table\n",
        "df_food.to_csv(TAB_DIR / 'food_dashboard.csv', index=False)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Health label counts\n",
        "health_counts = df_food['health_label'].value_counts().reset_index()\n",
        "health_counts.columns = ['health_label', 'count']\n",
        "health_counts.to_csv(TAB_DIR / 'health_label_counts.csv', index=False)\n",
        "health_counts\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Nutrition by health label\n",
        "nutrition = df_food.groupby('health_label')[[\n",
        "    'energy-kcal_100g',\n",
        "    'sugars_100g',\n",
        "    'fat_100g',\n",
        "    'salt_100g',\n",
        "    'proteins_100g',\n",
        "    'fiber_100g'\n",
        "]].mean().reset_index()\n",
        "\n",
        "nutrition.to_csv(TAB_DIR / 'nutrition_by_health_label.csv', index=False)\n",
        "nutrition\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Nutri-score vs health label\n",
        "df_compare = df_food[~df_food['nutriscore_grade'].isin(['unknown','not-applicable'])]\n",
        "\n",
        "nutri = pd.crosstab(df_compare['nutriscore_grade'], df_compare['health_label']).reset_index()\n",
        "nutri_long = nutri.melt(id_vars='nutriscore_grade', var_name='health_label', value_name='count')\n",
        "\n",
        "nutri_long.to_csv(TAB_DIR / 'nutriscore_vs_health.csv', index=False)\n",
        "nutri_long\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Recommendations\n",
        "df_food[['product_name','health_label','nutriscore_grade']].to_csv(\n",
        "    TAB_DIR / 'recommendations.csv', index=False\n",
        ")\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "# Charts\n",
        "plt.figure(figsize=(6,4))\n",
        "df_food['health_label'].value_counts().plot(kind='bar')\n",
        "plt.title('Health Label Distribution')\n",
        "plt.tight_layout()\n",
        "plt.savefig(FIG_DIR / 'health_label_distribution.png')\n",
        "plt.close()\n",
        "\n",
        "plt.figure(figsize=(8,5))\n",
        "pd.crosstab(df_compare['nutriscore_grade'], df_compare['health_label']).plot(kind='bar', stacked=True)\n",
        "plt.title('Nutri-Score vs Health Label')\n",
        "plt.tight_layout()\n",
        "plt.savefig(FIG_DIR / 'nutriscore_vs_health.png')\n",
        "plt.close()\n"
      ]
    }
  ]
}