notebooks/01_CreateAdjacencyMatrices.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create Adjacency Matrices\n",
    "We create a series of adjacency matrices in which the nodes are the different ingredients found in the Recipes1M+ Database. We generate 3 different ingredient lists:\n",
    "1. Full ingredients found in the Recipes 1M+ database\n",
    "2. First words of ingredients in the Recipes 1M+ database\n",
    "3. Full ingredients in the Recipes 1M+ database that can be cross-referenced with ingredients in the USDA nutrition facts database\n",
    "\n",
    "We then use each of these ingredient lists to create a variety of different adjacency matrices which we subsequently use for exploration, processing, graph filtering, ingredient suggestion, ingredient replacement, and more!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 1: Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ujYVXo3iPrZW"
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import os, sys\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import json\n",
    "from pandas.io.json import json_normalize\n",
    "import itertools\n",
    "from scipy.spatial.distance import pdist, squareform\n",
    "from graphProcessingFunctions import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "wTXgrVDBRZ3h"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fsa_lights_per100g</th>\n",
       "      <th>id</th>\n",
       "      <th>ingredients</th>\n",
       "      <th>instructions</th>\n",
       "      <th>nutr_per_ingredient</th>\n",
       "      <th>nutr_values_per100g</th>\n",
       "      <th>partition</th>\n",
       "      <th>quantity</th>\n",
       "      <th>title</th>\n",
       "      <th>unit</th>\n",
       "      <th>url</th>\n",
       "      <th>weight_per_ingr</th>\n",
       "      <th>check</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>{'fat': 'green', 'salt': 'green', 'saturates':...</td>\n",
       "      <td>000095fc1d</td>\n",
       "      <td>[{'text': 'yogurt, greek, plain, nonfat'}, {'t...</td>\n",
       "      <td>[{'text': 'Layer all ingredients in a serving ...</td>\n",
       "      <td>[{'fat': 0.8845044000000001, 'nrg': 133.80964,...</td>\n",
       "      <td>{'energy': 81.12946131894766, 'fat': 2.1401392...</td>\n",
       "      <td>train</td>\n",
       "      <td>[{'text': '8'}, {'text': '1'}, {'text': '1/4'}]</td>\n",
       "      <td>Yogurt Parfaits</td>\n",
       "      <td>[{'text': 'ounce'}, {'text': 'cup'}, {'text': ...</td>\n",
       "      <td>http://tastykitchen.com/recipes/breakfastbrunc...</td>\n",
       "      <td>[226.796, 152.0, 30.5]</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>{'fat': 'red', 'salt': 'orange', 'saturates': ...</td>\n",
       "      <td>00051d5b9d</td>\n",
       "      <td>[{'text': 'sugars, granulated'}, {'text': 'oil...</td>\n",
       "      <td>[{'text': 'Cream sugar and butter together til...</td>\n",
       "      <td>[{'fat': 0.0, 'nrg': 384.0, 'pro': 0.0, 'sat':...</td>\n",
       "      <td>{'energy': 477.09640393594606, 'fat': 23.41248...</td>\n",
       "      <td>val</td>\n",
       "      <td>[{'text': '1/2'}, {'text': '3/4'}, {'text': '1...</td>\n",
       "      <td>Salt Free, Low Cholesterol Sugar Cookies Recipe</td>\n",
       "      <td>[{'text': 'cup'}, {'text': 'cup'}, {'text': 'c...</td>\n",
       "      <td>http://cookeatshare.com/recipes/salt-free-low-...</td>\n",
       "      <td>[100.80000000000001, 168.0, 21.25, 5.166666666...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>{'fat': 'orange', 'salt': 'orange', 'saturates...</td>\n",
       "      <td>00059b093b</td>\n",
       "      <td>[{'text': 'chicken, broilers or fryers, wing, ...</td>\n",
       "      <td>[{'text': 'Preheat oven to 400 degrees F.'}, {...</td>\n",
       "      <td>[{'fat': 174.859716, 'nrg': 2599.0821599999995...</td>\n",
       "      <td>{'energy': 208.05898280960727, 'fat': 14.29704...</td>\n",
       "      <td>train</td>\n",
       "      <td>[{'text': '3'}, {'text': '1'}, {'text': '1/2'}...</td>\n",
       "      <td>Honey Sriracha Chicken Wings</td>\n",
       "      <td>[{'text': 'pound'}, {'text': 'teaspoon'}, {'te...</td>\n",
       "      <td>http://tastykitchen.com/recipes/main-courses/h...</td>\n",
       "      <td>[1360.7759999999998, 6.0, 1.15, 13.5, 42.59999...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>{'fat': 'orange', 'salt': 'orange', 'saturates...</td>\n",
       "      <td>0005fc89f7</td>\n",
       "      <td>[{'text': 'crustaceans, shrimp, raw (not previ...</td>\n",
       "      <td>[{'text': 'In a large bowl, toss the shrimp, g...</td>\n",
       "      <td>[{'fat': 4.6266384, 'nrg': 771.1064, 'pro': 18...</td>\n",
       "      <td>{'energy': 194.7525956849092, 'fat': 15.980767...</td>\n",
       "      <td>train</td>\n",
       "      <td>[{'text': '2'}, {'text': '3/4'}, {'text': '1 1...</td>\n",
       "      <td>Shrimp and Caper Salad</td>\n",
       "      <td>[{'text': 'pound'}, {'text': 'cup'}, {'text': ...</td>\n",
       "      <td>http://allrecipes.com/recipe/shrimp-and-caper-...</td>\n",
       "      <td>[907.184, 75.0, 151.5, 25.799999999999997, 134...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>{'fat': 'red', 'salt': 'green', 'saturates': '...</td>\n",
       "      <td>0006ca31f4</td>\n",
       "      <td>[{'text': 'cocoa, dry powder, unsweetened'}, {...</td>\n",
       "      <td>[{'text': 'Measure out the cocoa powder into a...</td>\n",
       "      <td>[{'fat': 5.889999999999999, 'nrg': 98.0, 'pro'...</td>\n",
       "      <td>{'energy': 457.0971184631804, 'fat': 29.329775...</td>\n",
       "      <td>train</td>\n",
       "      <td>[{'text': '1/2'}, {'text': '1/2'}, {'text': '1'}]</td>\n",
       "      <td>Natural Peanut Butter Chocolate Bon Bons</td>\n",
       "      <td>[{'text': 'cup'}, {'text': 'cup'}, {'text': 'c...</td>\n",
       "      <td>http://www.food.com/recipe/natural-peanut-butt...</td>\n",
       "      <td>[43.0, 169.5, 256.0]</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  fsa_lights_per100g          id  \\\n",
       "0  {'fat': 'green', 'salt': 'green', 'saturates':...  000095fc1d   \n",
       "1  {'fat': 'red', 'salt': 'orange', 'saturates': ...  00051d5b9d   \n",
       "2  {'fat': 'orange', 'salt': 'orange', 'saturates...  00059b093b   \n",
       "3  {'fat': 'orange', 'salt': 'orange', 'saturates...  0005fc89f7   \n",
       "4  {'fat': 'red', 'salt': 'green', 'saturates': '...  0006ca31f4   \n",
       "\n",
       "                                         ingredients  \\\n",
       "0  [{'text': 'yogurt, greek, plain, nonfat'}, {'t...   \n",
       "1  [{'text': 'sugars, granulated'}, {'text': 'oil...   \n",
       "2  [{'text': 'chicken, broilers or fryers, wing, ...   \n",
       "3  [{'text': 'crustaceans, shrimp, raw (not previ...   \n",
       "4  [{'text': 'cocoa, dry powder, unsweetened'}, {...   \n",
       "\n",
       "                                        instructions  \\\n",
       "0  [{'text': 'Layer all ingredients in a serving ...   \n",
       "1  [{'text': 'Cream sugar and butter together til...   \n",
       "2  [{'text': 'Preheat oven to 400 degrees F.'}, {...   \n",
       "3  [{'text': 'In a large bowl, toss the shrimp, g...   \n",
       "4  [{'text': 'Measure out the cocoa powder into a...   \n",
       "\n",
       "                                 nutr_per_ingredient  \\\n",
       "0  [{'fat': 0.8845044000000001, 'nrg': 133.80964,...   \n",
       "1  [{'fat': 0.0, 'nrg': 384.0, 'pro': 0.0, 'sat':...   \n",
       "2  [{'fat': 174.859716, 'nrg': 2599.0821599999995...   \n",
       "3  [{'fat': 4.6266384, 'nrg': 771.1064, 'pro': 18...   \n",
       "4  [{'fat': 5.889999999999999, 'nrg': 98.0, 'pro'...   \n",
       "\n",
       "                                 nutr_values_per100g partition  \\\n",
       "0  {'energy': 81.12946131894766, 'fat': 2.1401392...     train   \n",
       "1  {'energy': 477.09640393594606, 'fat': 23.41248...       val   \n",
       "2  {'energy': 208.05898280960727, 'fat': 14.29704...     train   \n",
       "3  {'energy': 194.7525956849092, 'fat': 15.980767...     train   \n",
       "4  {'energy': 457.0971184631804, 'fat': 29.329775...     train   \n",
       "\n",
       "                                            quantity  \\\n",
       "0    [{'text': '8'}, {'text': '1'}, {'text': '1/4'}]   \n",
       "1  [{'text': '1/2'}, {'text': '3/4'}, {'text': '1...   \n",
       "2  [{'text': '3'}, {'text': '1'}, {'text': '1/2'}...   \n",
       "3  [{'text': '2'}, {'text': '3/4'}, {'text': '1 1...   \n",
       "4  [{'text': '1/2'}, {'text': '1/2'}, {'text': '1'}]   \n",
       "\n",
       "                                             title  \\\n",
       "0                                  Yogurt Parfaits   \n",
       "1  Salt Free, Low Cholesterol Sugar Cookies Recipe   \n",
       "2                     Honey Sriracha Chicken Wings   \n",
       "3                           Shrimp and Caper Salad   \n",
       "4         Natural Peanut Butter Chocolate Bon Bons   \n",
       "\n",
       "                                                unit  \\\n",
       "0  [{'text': 'ounce'}, {'text': 'cup'}, {'text': ...   \n",
       "1  [{'text': 'cup'}, {'text': 'cup'}, {'text': 'c...   \n",
       "2  [{'text': 'pound'}, {'text': 'teaspoon'}, {'te...   \n",
       "3  [{'text': 'pound'}, {'text': 'cup'}, {'text': ...   \n",
       "4  [{'text': 'cup'}, {'text': 'cup'}, {'text': 'c...   \n",
       "\n",
       "                                                 url  \\\n",
       "0  http://tastykitchen.com/recipes/breakfastbrunc...   \n",
       "1  http://cookeatshare.com/recipes/salt-free-low-...   \n",
       "2  http://tastykitchen.com/recipes/main-courses/h...   \n",
       "3  http://allrecipes.com/recipe/shrimp-and-caper-...   \n",
       "4  http://www.food.com/recipe/natural-peanut-butt...   \n",
       "\n",
       "                                     weight_per_ingr  check  \n",
       "0                             [226.796, 152.0, 30.5]    NaN  \n",
       "1  [100.80000000000001, 168.0, 21.25, 5.166666666...    NaN  \n",
       "2  [1360.7759999999998, 6.0, 1.15, 13.5, 42.59999...    NaN  \n",
       "3  [907.184, 75.0, 151.5, 25.799999999999997, 134...    NaN  \n",
       "4                               [43.0, 169.5, 256.0]  False  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Load Recipes 1M+ Database\n",
    "with open('../data/recipes_with_nutritional_info_fixed_qty.json') as json_file:\n",
    "    data = json.load(json_file)\n",
    "df = pd.DataFrame(data)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "vH2eEzjr_XFJ",
    "outputId": "9d79a644-91c2-428e-8b9a-d9d6b797387a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The number of recipes in the Recipes 1M+ Database is 51235.\n"
     ]
    }
   ],
   "source": [
    "print(\"The number of recipes in the Recipes 1M+ Database is {}.\".format(len(df)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SR description</th>\n",
       "      <th>Nutrient description</th>\n",
       "      <th>Nutrient value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>Butter, salted</td>\n",
       "      <td>Protein</td>\n",
       "      <td>0.85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Butter, salted</td>\n",
       "      <td>Total Fat</td>\n",
       "      <td>81.11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>Butter, salted</td>\n",
       "      <td>Carbohydrate</td>\n",
       "      <td>0.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Butter, salted</td>\n",
       "      <td>Energy</td>\n",
       "      <td>717.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>Butter, salted</td>\n",
       "      <td>Alcohol</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SR description Nutrient description  Nutrient value\n",
       "0  Butter, salted              Protein            0.85\n",
       "1  Butter, salted            Total Fat           81.11\n",
       "2  Butter, salted         Carbohydrate            0.06\n",
       "3  Butter, salted               Energy          717.00\n",
       "4  Butter, salted              Alcohol            0.00"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Load USDA nutrition facts database\n",
    "usda = pd.read_excel('../data/USDA_Database.xlsx')\n",
    "usda = usda[[\"SR description\",\"Nutrient description\", \"Nutrient value\"]]\n",
    "usda.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 2: Build ingredient lists from Recipes 1M+ Database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "leLi9zL4WMlc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 357 unique ingredients.\n",
      "Example ingredients: \n",
      "['apple juice, canned or bottled, unsweetened, without added ascorbic acid'\n",
      " 'apples, raw, with skin'\n",
      " 'applesauce, canned, unsweetened, without added ascorbic acid (includes usda commodity)'\n",
      " 'apricots, dried, sulfured, uncooked'\n",
      " 'artichokes, (globe or french), raw']\n"
     ]
    }
   ],
   "source": [
    "all_ingredients_full = []\n",
    "raw_ingredients = df[['ingredients']].values\n",
    "for ri in raw_ingredients:\n",
    "    for i in ri[0]:\n",
    "        all_ingredients_full += [i['text']]\n",
    "all_ingredients_full_unique = np.unique(all_ingredients_full)\n",
    "print(\"There are {} unique ingredients.\".format(len(all_ingredients_full_unique)))\n",
    "print(\"Example ingredients: \")\n",
    "print(all_ingredients_full_unique[10:15])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "EnSWml1KVYQs"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 213 unique ingredients (first words only).\n",
      "Example ingredients: \n",
      "['asparagus' 'avocados' 'bacon' 'baking chocolate' 'bananas']\n"
     ]
    }
   ],
   "source": [
    "all_ingredients_firstword = []\n",
    "for ing in all_ingredients_full:\n",
    "    all_ingredients_firstword.append(ing.split(',')[0])\n",
    "all_ingredients_unique = np.unique(all_ingredients_firstword,return_index=False)\n",
    "print(\"There are {} unique ingredients (first words only).\".format(len(all_ingredients_unique)))\n",
    "print('Example ingredients: ')\n",
    "print(all_ingredients_unique[10:15])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Save ingredient lists\n",
    "np.save('../data/Ingredient_List.npy', all_ingredients_full_unique)\n",
    "np.save('../data/Ingredient_List_First_Word.npy', all_ingredients_unique)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix No. 1\n",
    "Weights between ingredient nodes are equal to the number of recipes in which the two ingredients appear together in the Recipe1M+ dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "xJwmpjHoXRQo"
   },
   "outputs": [],
   "source": [
    "#Create adjacency matrix with weights equal to the number of recipes in which two ingredients appear together\n",
    "Adj_mat_1 = np.zeros([len(all_ingredients_full_unique),len(all_ingredients_full_unique)])\n",
    "ingredients_df = df['ingredients']\n",
    "for recipe in ingredients_df.values:\n",
    "    ingredients = []\n",
    "    for ingredient in recipe:\n",
    "        ingredients += [ingredient['text']]\n",
    "    indices = np.where(np.in1d(all_ingredients_full_unique,ingredients))\n",
    "    combos = list(itertools.combinations(indices[0],2))\n",
    "    for combo in combos:\n",
    "        Adj_mat_1[combo[0],combo[1]] += 1\n",
    "        Adj_mat_1[combo[1],combo[0]] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "id": "_OV9-Z10iSPw",
    "outputId": "f39dc196-72e1-4f2e-bc60-82f626961ed5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8385.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Visualize and save\n",
    "plt.spy(Adj_mat_1)\n",
    "print(np.max(Adj_mat_1))\n",
    "np.save('../data/Adjacency_Matrix_1.npy',Adj_mat_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The ingredients most commonly used together are salt, table and wheat flour, white, all-purpose, unenriched.\n"
     ]
    }
   ],
   "source": [
    "am1 = np.unravel_index(np.argmax(Adj_mat_1, axis=None), Adj_mat_1.shape)\n",
    "print(\"The ingredients most commonly used together are {0} and {1}.\".format(all_ingredients_full_unique[am1[0]], all_ingredients_full_unique[am1[1]]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "a0O-AIll3tco"
   },
   "source": [
    "### Adjacency Matrix No. 2\n",
    "Same as adjacency matrix #1 but using only the first word of each ingredient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5NouZJVn34Qc"
   },
   "outputs": [],
   "source": [
    "Adj_mat_2 = np.zeros([len(all_ingredients_unique),len(all_ingredients_unique)])\n",
    "ingredients_df = df['ingredients']\n",
    "for recipe in ingredients_df.values:\n",
    "    ingredients = []\n",
    "    for ingredient in recipe:\n",
    "        ingredients += [ingredient['text'].split(',')[0]]\n",
    "    indices = np.where(np.in1d(all_ingredients_unique,ingredients))\n",
    "    combos = list(itertools.combinations(indices[0],2))\n",
    "    for combo in combos:\n",
    "        Adj_mat_2[combo[0],combo[1]] += 1\n",
    "        Adj_mat_2[combo[1],combo[0]] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "id": "sFjwinlL6F3k",
    "outputId": "89e759a7-668a-4783-8f24-89bd30ce3f14"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10666.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.spy(Adj_mat_2)\n",
    "print(np.max(Adj_mat_2))\n",
    "np.save('../data/Adjacency_Matrix_2.npy',Adj_mat_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "QuCfnBVL7ewL"
   },
   "source": [
    "### Adjacency Matrix No. 3\n",
    "Build an adjacency matrix Cora-style (like we did in the first assignment): Have each recipe represent a \"feature\" for each ingredient. The feature equals 1 if the recipe includes that ingredient and 0 otherwise. Then use an RBF kernel to compute the distances between ingredients based on their feature vectors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "heIXn_Y97mE4"
   },
   "outputs": [],
   "source": [
    "num_recipes = len(df)\n",
    "feature = np.zeros([len(all_ingredients_full_unique),num_recipes])\n",
    "ingredients_df = df['ingredients']\n",
    "recipe_counter = 0\n",
    "for recipe in ingredients_df.values:\n",
    "    ingredients = []\n",
    "    for ingredient in recipe:\n",
    "        ingredients += [ingredient['text']]\n",
    "    indices = np.where(np.in1d(all_ingredients_full_unique,ingredients))\n",
    "    feature[indices,recipe_counter] = 1\n",
    "    recipe_counter = recipe_counter + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "colab_type": "code",
    "id": "7O7rQ7eAFyD_",
    "outputId": "b984d654-0cf8-4502-fcd5-bf2c2d12d33e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32.63865095634238\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot distances between nodes\n",
    "distances = pdist(feature, metric='euclidean')\n",
    "plt.hist(distances)\n",
    "mean_dist = np.mean(distances)\n",
    "print(mean_dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def epsilon_similarity_graph(X: np.ndarray, sigma=1, epsilon=0):\n",
    "    \"\"\" X (n x d): coordinates of the n data points in R^d.\n",
    "        sigma (float): width of the kernel\n",
    "        epsilon (float): threshold\n",
    "        Return:\n",
    "        adjacency (n x n ndarray): adjacency matrix of the graph.\n",
    "    \"\"\"\n",
    "    dist = squareform(pdist(X))\n",
    "    adjacency = np.exp(- dist ** 2 / (2 * sigma ** 2))\n",
    "    adjacency[adjacency < epsilon] = 0\n",
    "    np.fill_diagonal(adjacency, 0)\n",
    "    return adjacency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "tFK1zizTDkrC",
    "outputId": "d6720d34-4d17-4256-b4b3-fd4a682b521e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(feature, sigma, epsilon=0.2)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GG-p3PFiIAVL"
   },
   "outputs": [],
   "source": [
    "#Save the matrix\n",
    "np.save('../data/Adjacency_Matrix_3.npy',adjacency)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "hb14iGzLsBxJ"
   },
   "source": [
    "## Side note\n",
    "Testing the credibility of nutrition information for each ingredient from the Recipes 1M+ database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "iJgk7vEe1k74",
    "outputId": "c911f9f2-22a7-4d5b-8473-93f32d7727f7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 10: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 11: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 14: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 67: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 187: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 239: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 310: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 329: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 454: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 538: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 646: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 686: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 954: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 1159: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 1378: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 1515: 4 1/2 cup of broccoli nutrition facts: {'fat': 1.53, 'nrg': 139.5, 'pro': 11.565, 'sat': 0.15750000000000003, 'sod': 135.0, 'sug': 6.9750000000000005}\n",
      "Recipe 1569: 12 pound of broccoli nutrition facts: {'fat': 20.139484799999998, 'nrg': 1850.65536, 'pro': 153.49553279999998, 'sat': 2.1228105599999996, 'sod': 1796.2243199999998, 'sug': 92.53276799999999}\n",
      "Recipe 1731: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 1741: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 1986: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 1996: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 2076: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 2344: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 2569: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 2837: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 2946: 3 pound of broccoli nutrition facts: {'fat': 5.0348711999999995, 'nrg': 462.66384, 'pro': 38.373883199999995, 'sat': 0.5307026399999999, 'sod': 449.05607999999995, 'sug': 23.133191999999998}\n",
      "Recipe 3074: 6 ounce of broccoli nutrition facts: {'fat': 0.6293588999999999, 'nrg': 57.83298, 'pro': 4.796735399999999, 'sat': 0.06633782999999999, 'sod': 56.132009999999994, 'sug': 2.8916489999999997}\n",
      "Recipe 3185: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 3455: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 3459: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 3469: 12 cup of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 3649: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 3695: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 4370: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 4655: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 4656: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 4676: 34 pound of broccoli nutrition facts: {'fat': 57.0618736, 'nrg': 5243.52352, 'pro': 434.90400959999994, 'sat': 6.014629919999999, 'sod': 5089.30224, 'sug': 262.176176}\n",
      "Recipe 4718: 12 cup of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 5007: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 5083: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 5086: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 5106: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 5168: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 5257: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 5416: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 5588: 3 quart of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 5674: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 5676: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 5749: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 5973: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 6096: 16 pound of broccoli nutrition facts: {'fat': 26.8526464, 'nrg': 2467.54048, 'pro': 204.6607104, 'sat': 2.8304140799999997, 'sod': 2394.96576, 'sug': 123.377024}\n",
      "Recipe 6126: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 6305: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 6355: 4 1/2 cup of broccoli nutrition facts: {'fat': 1.53, 'nrg': 139.5, 'pro': 11.565, 'sat': 0.15750000000000003, 'sod': 135.0, 'sug': 6.9750000000000005}\n",
      "Recipe 6401: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 6425: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 6481: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 6863: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 6966: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 6982: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 7013: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 7253: 3/4 cup of broccoli nutrition facts: {'fat': 0.255, 'nrg': 23.25, 'pro': 1.9275, 'sat': 0.026250000000000006, 'sod': 22.5, 'sug': 1.1625}\n",
      "Recipe 7354: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 7458: 8 cup of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 7759: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 8061: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 8079: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 8160: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 8179: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 8180: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 8358: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 8362: 3/4 cup of broccoli nutrition facts: {'fat': 0.255, 'nrg': 23.25, 'pro': 1.9274999999999998, 'sat': 0.026250000000000002, 'sod': 22.5, 'sug': 1.1625}\n",
      "Recipe 8443: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 8488: 1/4 cup of broccoli nutrition facts: {'fat': 0.085, 'nrg': 7.75, 'pro': 0.6425, 'sat': 0.00875, 'sod': 7.5, 'sug': 0.38749999999999996}\n",
      "Recipe 8619: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 8628: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 8638: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 8695: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 8955: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 8981: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 9132: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 9197: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 9258: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 9343: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 9388: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 9437: 8 ounce of broccoli nutrition facts: {'fat': 0.8391452, 'nrg': 77.11064, 'pro': 6.3956472, 'sat': 0.08845043999999999, 'sod': 74.84268, 'sug': 3.855532}\n",
      "Recipe 9605: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 9700: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 9706: 3 pound of broccoli nutrition facts: {'fat': 5.0348711999999995, 'nrg': 462.66384, 'pro': 38.373883199999995, 'sat': 0.5307026399999999, 'sod': 449.05607999999995, 'sug': 23.133191999999998}\n",
      "Recipe 9900: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 9997: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 10011: 6 ounce of broccoli nutrition facts: {'fat': 0.6293588999999999, 'nrg': 57.83298, 'pro': 4.796735399999999, 'sat': 0.06633782999999999, 'sod': 56.132009999999994, 'sug': 2.8916489999999997}\n",
      "Recipe 10144: 1/2 pound of broccoli nutrition facts: {'fat': 0.8391452, 'nrg': 77.11064, 'pro': 6.3956472, 'sat': 0.08845043999999999, 'sod': 74.84268, 'sug': 3.855532}\n",
      "Recipe 10241: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 10278: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 10464: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 10581: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 10936: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 10947: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 11649: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 11679: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 11694: 2/3 cup of broccoli nutrition facts: {'fat': 0.22666666666666668, 'nrg': 20.666666666666668, 'pro': 1.7133333333333334, 'sat': 0.023333333333333334, 'sod': 20.0, 'sug': 1.0333333333333332}\n",
      "Recipe 11781: 8 cup of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 11881: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 11900: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 12296: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 12357: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 12475: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 12675: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 12774: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 12834: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 13380: 100 g of broccoli nutrition facts: {'fat': 0.37, 'nrg': 34.0, 'pro': 2.82, 'sat': 0.039, 'sod': 33.0, 'sug': 1.7000000000000002}\n",
      "Recipe 13486: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 13764: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 13981: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 14268: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 14464: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 14534: 14 cup of broccoli nutrition facts: {'fat': 4.760000000000001, 'nrg': 434.0, 'pro': 35.98, 'sat': 0.49000000000000005, 'sod': 420.0, 'sug': 21.7}\n",
      "Recipe 14572: 14 pound of broccoli nutrition facts: {'fat': 23.496065599999998, 'nrg': 2159.09792, 'pro': 179.07812159999997, 'sat': 2.4766123199999996, 'sod': 2095.59504, 'sug': 107.954896}\n",
      "Recipe 14671: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 14678: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 14680: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 14707: 60 g of broccoli nutrition facts: {'fat': 0.222, 'nrg': 20.400000000000002, 'pro': 1.692, 'sat': 0.0234, 'sod': 19.8, 'sug': 1.02}\n",
      "Recipe 14713: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 14885: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 15096: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 15102: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 15302: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 15309: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 15396: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 15396: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 15423: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 15750: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 15907: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 15922: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 15963: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 15987: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 16312: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 16319: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 16699: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 16725: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 16882: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 17201: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 17415: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 17467: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 17662: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 17747: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 17927: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 18162: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 18390: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 18493: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 18530: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 18851: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 18902: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 18954: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 19025: 2 1/2 pound of broccoli nutrition facts: {'fat': 4.1957260000000005, 'nrg': 385.55320000000006, 'pro': 31.978236, 'sat': 0.4422522, 'sod': 374.21340000000004, 'sug': 19.27766}\n",
      "Recipe 19277: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 19307: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 19363: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 19453: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 19544: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 19608: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 20167: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 20215: 7 cup of broccoli nutrition facts: {'fat': 2.3800000000000003, 'nrg': 217.0, 'pro': 17.99, 'sat': 0.24500000000000002, 'sod': 210.0, 'sug': 10.85}\n",
      "Recipe 20263: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 20303: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 20392: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 20483: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 20607: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 20655: 700 g of broccoli nutrition facts: {'fat': 2.5900000000000003, 'nrg': 238.00000000000003, 'pro': 19.74, 'sat': 0.273, 'sod': 231.0, 'sug': 11.9}\n",
      "Recipe 20670: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 20679: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 20720: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 21135: 1/2 pound of broccoli nutrition facts: {'fat': 0.8391452, 'nrg': 77.11064, 'pro': 6.3956472, 'sat': 0.08845043999999999, 'sod': 74.84268, 'sug': 3.855532}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 21215: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 21527: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 21643: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 21669: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 21723: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 21817: 3/4 pound of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 21822: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 21941: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 21976: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 22068: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 22094: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 22181: 500 g of broccoli nutrition facts: {'fat': 1.85, 'nrg': 170.0, 'pro': 14.1, 'sat': 0.195, 'sod': 165.0, 'sug': 8.5}\n",
      "Recipe 22372: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 22373: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 22745: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 22956: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 23085: 12 cup of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 23108: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 23154: 2 1/2 cup of broccoli nutrition facts: {'fat': 0.8500000000000001, 'nrg': 77.5, 'pro': 6.425, 'sat': 0.08750000000000001, 'sod': 75.0, 'sug': 3.875}\n",
      "Recipe 23226: 4 1/2 cup of broccoli nutrition facts: {'fat': 1.53, 'nrg': 139.5, 'pro': 11.565, 'sat': 0.15750000000000003, 'sod': 135.0, 'sug': 6.9750000000000005}\n",
      "Recipe 23271: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 23479: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 23667: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 23830: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 23913: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 23924: 3/4 cup of broccoli nutrition facts: {'fat': 0.255, 'nrg': 23.25, 'pro': 1.9275, 'sat': 0.026250000000000006, 'sod': 22.5, 'sug': 1.1625}\n",
      "Recipe 24004: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 24124: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 24148: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 24166: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 24373: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 24495: 2 1/2 pound of broccoli nutrition facts: {'fat': 4.1957260000000005, 'nrg': 385.55320000000006, 'pro': 31.978236, 'sat': 0.4422522, 'sod': 374.21340000000004, 'sug': 19.27766}\n",
      "Recipe 24705: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 24718: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 24965: 12 pound of broccoli nutrition facts: {'fat': 20.139484799999998, 'nrg': 1850.65536, 'pro': 153.49553279999998, 'sat': 2.1228105599999996, 'sod': 1796.2243199999998, 'sug': 92.53276799999999}\n",
      "Recipe 25137: 20 g of broccoli nutrition facts: {'fat': 0.07400000000000001, 'nrg': 6.800000000000001, 'pro': 0.564, 'sat': 0.0078, 'sod': 6.6000000000000005, 'sug': 0.34}\n",
      "Recipe 25157: 2 1/2 pound of broccoli nutrition facts: {'fat': 4.1957260000000005, 'nrg': 385.55320000000006, 'pro': 31.978236, 'sat': 0.4422522, 'sod': 374.21340000000004, 'sug': 19.27766}\n",
      "Recipe 25349: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 25469: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 25868: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 25926: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 25946: 300 g of broccoli nutrition facts: {'fat': 1.11, 'nrg': 102.00000000000001, 'pro': 8.459999999999999, 'sat': 0.11699999999999999, 'sod': 99.0, 'sug': 5.1000000000000005}\n",
      "Recipe 25953: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 25989: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 25995: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 26180: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 26390: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 26413: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 26522: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 26586: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 26740: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 26740: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 26970: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 26991: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 27017: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 27301: 4 quart of broccoli nutrition facts: {'fat': 5.44, 'nrg': 496.0, 'pro': 41.12, 'sat': 0.56, 'sod': 480.0, 'sug': 24.8}\n",
      "Recipe 27350: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 27351: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 27426: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 27467: 12 cup of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 27597: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 27853: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 27958: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 28238: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 28557: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 28561: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 28698: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 28846: 1/4 cup of broccoli nutrition facts: {'fat': 0.085, 'nrg': 7.75, 'pro': 0.6425, 'sat': 0.00875, 'sod': 7.5, 'sug': 0.38749999999999996}\n",
      "Recipe 29028: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 29144: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 29415: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 29442: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 29452: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 29458: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 29581: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 29627: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 29747: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 29758: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 29861: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 29889: 20 pound of broccoli nutrition facts: {'fat': 33.565808000000004, 'nrg': 3084.4256000000005, 'pro': 255.825888, 'sat': 3.5380176, 'sod': 2993.7072000000003, 'sug': 154.22128}\n",
      "Recipe 29910: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 29911: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 29944: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 29988: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 30001: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 30092: 4 1/2 cup of broccoli nutrition facts: {'fat': 1.53, 'nrg': 139.5, 'pro': 11.565, 'sat': 0.15750000000000003, 'sod': 135.0, 'sug': 6.9750000000000005}\n",
      "Recipe 30182: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 30246: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 30358: 14 cup of broccoli nutrition facts: {'fat': 4.760000000000001, 'nrg': 434.0, 'pro': 35.98, 'sat': 0.49000000000000005, 'sod': 420.0, 'sug': 21.7}\n",
      "Recipe 30393: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 30470: 1/3 cup of broccoli nutrition facts: {'fat': 0.11333333333333333, 'nrg': 10.333333333333332, 'pro': 0.8566666666666666, 'sat': 0.011666666666666669, 'sod': 10.0, 'sug': 0.5166666666666667}\n",
      "Recipe 30663: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 30698: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 30721: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 30755: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 30845: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 30916: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 30965: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 30971: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 31091: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 31348: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 31399: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 31438: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 31448: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 31537: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 31827: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 31896: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 31926: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 32048: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 32101: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 32207: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 32571: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 32813: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 32829: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 32832: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 32889: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 32954: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 33092: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 33262: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 33296: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 33509: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 33558: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 33635: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 34001: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 34172: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 34208: 2 1/2 pound of broccoli nutrition facts: {'fat': 4.1957260000000005, 'nrg': 385.55320000000006, 'pro': 31.978236, 'sat': 0.4422522, 'sod': 374.21340000000004, 'sug': 19.27766}\n",
      "Recipe 34234: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 34237: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 34239: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 34447: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 34621: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 34915: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 34949: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 35272: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 35382: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 35568: 8 cup of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 35635: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 35768: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 35840: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 35866: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 35875: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 35899: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 36260: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 36391: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 36414: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 36659: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 36772: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 36917: 8 cup of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 37010: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 37098: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 37189: 3 pound of broccoli nutrition facts: {'fat': 5.0348711999999995, 'nrg': 462.66384, 'pro': 38.373883199999995, 'sat': 0.5307026399999999, 'sod': 449.05607999999995, 'sug': 23.133191999999998}\n",
      "Recipe 37199: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 37235: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 37615: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 37727: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 37758: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 37795: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 37831: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 37845: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 37950: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 37963: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 37983: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 38547: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 38636: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 38645: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 38651: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 38736: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 38845: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 38979: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 39045: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 39160: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 39213: 12 cup of broccoli nutrition facts: {'fat': 4.08, 'nrg': 372.0, 'pro': 30.839999999999996, 'sat': 0.42000000000000004, 'sod': 360.0, 'sug': 18.6}\n",
      "Recipe 39317: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 39557: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 39728: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 39752: 1 1/4 cup of broccoli nutrition facts: {'fat': 0.42500000000000004, 'nrg': 38.75, 'pro': 3.2125, 'sat': 0.043750000000000004, 'sod': 37.5, 'sug': 1.9375}\n",
      "Recipe 39824: 1 1/4 pound of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 40008: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 40046: 2 1/2 pound of broccoli nutrition facts: {'fat': 4.1957260000000005, 'nrg': 385.55320000000006, 'pro': 31.978236, 'sat': 0.4422522, 'sod': 374.21340000000004, 'sug': 19.27766}\n",
      "Recipe 40397: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 40444: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 40540: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 40720: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 40798: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 41004: 3/4 pound of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 41006: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 41013: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 41039: 3/4 cup of broccoli nutrition facts: {'fat': 0.255, 'nrg': 23.25, 'pro': 1.9275, 'sat': 0.026250000000000006, 'sod': 22.5, 'sug': 1.1625}\n",
      "Recipe 41462: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 41630: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 41632: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 41657: 8 cup of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 41662: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 41723: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 41737: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 41745: 2 quart of broccoli nutrition facts: {'fat': 2.72, 'nrg': 248.0, 'pro': 20.56, 'sat': 0.28, 'sod': 240.0, 'sug': 12.4}\n",
      "Recipe 41802: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 41811: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 41856: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 41983: 12 pound of broccoli nutrition facts: {'fat': 20.139484799999998, 'nrg': 1850.65536, 'pro': 153.49553279999998, 'sat': 2.1228105599999996, 'sod': 1796.2243199999998, 'sug': 92.53276799999999}\n",
      "Recipe 42163: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 42247: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 42464: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 42542: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 42576: 9 cup of broccoli nutrition facts: {'fat': 3.06, 'nrg': 279.0, 'pro': 23.13, 'sat': 0.31500000000000006, 'sod': 270.0, 'sug': 13.950000000000001}\n",
      "Recipe 42900: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 42962: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 43142: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 43174: 4-5 cup of broccoli nutrition facts: {'fat': 3.06, 'nrg': 279.0, 'pro': 23.13, 'sat': 0.31500000000000006, 'sod': 270.0, 'sug': 13.950000000000001}\n",
      "Recipe 43181: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 43216: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 43258: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 43385: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 43449: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 43471: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 43532: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 43589: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 43663: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 43698: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 43699: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 43710: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 43808: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 43828: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 44008: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 44052: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 44098: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 44138: 5 cup of broccoli nutrition facts: {'fat': 1.7000000000000002, 'nrg': 155.0, 'pro': 12.85, 'sat': 0.17500000000000002, 'sod': 150.0, 'sug': 7.75}\n",
      "Recipe 44309: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 44488: 16 ounce of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 44714: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 44919: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 45031: 12 pound of broccoli nutrition facts: {'fat': 20.139484799999998, 'nrg': 1850.65536, 'pro': 153.49553279999998, 'sat': 2.1228105599999996, 'sod': 1796.2243199999998, 'sug': 92.53276799999999}\n",
      "Recipe 45136: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n",
      "Recipe 45205: 5 pound of broccoli nutrition facts: {'fat': 8.391452000000001, 'nrg': 771.1064000000001, 'pro': 63.956472, 'sat': 0.8845044, 'sod': 748.4268000000001, 'sug': 38.55532}\n",
      "Recipe 45273: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 45360: 20 ounce of broccoli nutrition facts: {'fat': 2.0978630000000003, 'nrg': 192.77660000000003, 'pro': 15.989118, 'sat': 0.2211261, 'sod': 187.10670000000002, 'sug': 9.63883}\n",
      "Recipe 45374: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 45829: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 45904: 8 ounce of broccoli nutrition facts: {'fat': 0.8391452, 'nrg': 77.11064, 'pro': 6.3956472, 'sat': 0.08845043999999999, 'sod': 74.84268, 'sug': 3.855532}\n",
      "Recipe 45931: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 46052: 1/2 pound of broccoli nutrition facts: {'fat': 0.8391452, 'nrg': 77.11064, 'pro': 6.3956472, 'sat': 0.08845043999999999, 'sod': 74.84268, 'sug': 3.855532}\n",
      "Recipe 46099: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 46167: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 46371: 750 g of broccoli nutrition facts: {'fat': 2.775, 'nrg': 255.00000000000003, 'pro': 21.15, 'sat': 0.2925, 'sod': 247.5, 'sug': 12.750000000000002}\n",
      "Recipe 46450: 1/4 cup of broccoli nutrition facts: {'fat': 0.085, 'nrg': 7.75, 'pro': 0.6425, 'sat': 0.00875, 'sod': 7.5, 'sug': 0.3875}\n",
      "Recipe 46535: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 46603: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 46636: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 46671: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 46788: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 47185: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 47368: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 47384: 1/2 cup of broccoli nutrition facts: {'fat': 0.17, 'nrg': 15.5, 'pro': 1.285, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 47523: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 47614: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 47637: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 47646: 12 ounce of broccoli nutrition facts: {'fat': 1.2587177999999999, 'nrg': 115.66596, 'pro': 9.593470799999999, 'sat': 0.13267565999999997, 'sod': 112.26401999999999, 'sug': 5.783297999999999}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recipe 47715: 14 pound of broccoli nutrition facts: {'fat': 23.496065599999998, 'nrg': 2159.09792, 'pro': 179.07812159999997, 'sat': 2.4766123199999996, 'sod': 2095.59504, 'sug': 107.954896}\n",
      "Recipe 47890: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 47940: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 47971: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 47975: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 48085: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 48166: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 48215: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n",
      "Recipe 48233: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 48633: 1 1/2 quart of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 48649: 12 pound of broccoli nutrition facts: {'fat': 20.139484799999998, 'nrg': 1850.65536, 'pro': 153.49553279999998, 'sat': 2.1228105599999996, 'sod': 1796.2243199999998, 'sug': 92.53276799999999}\n",
      "Recipe 48650: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 48716: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 48950: 1 1/2 cup of broccoli nutrition facts: {'fat': 0.51, 'nrg': 46.5, 'pro': 3.8549999999999995, 'sat': 0.052500000000000005, 'sod': 45.0, 'sug': 2.325}\n",
      "Recipe 49010: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 49080: 6 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 49126: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 49205: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 49234: 10 ounce of broccoli nutrition facts: {'fat': 1.0489315000000001, 'nrg': 96.38830000000002, 'pro': 7.994559, 'sat': 0.11056305, 'sod': 93.55335000000001, 'sug': 4.819415}\n",
      "Recipe 49475: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 49638: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 49675: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 49985: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 50031: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 50045: 6.0 cup of broccoli nutrition facts: {'fat': 2.04, 'nrg': 186.0, 'pro': 15.419999999999998, 'sat': 0.21000000000000002, 'sod': 180.0, 'sug': 9.3}\n",
      "Recipe 50064: 2 pound of broccoli nutrition facts: {'fat': 3.3565808, 'nrg': 308.44256, 'pro': 25.5825888, 'sat': 0.35380175999999997, 'sod': 299.37072, 'sug': 15.422128}\n",
      "Recipe 50326: 3 cup of broccoli nutrition facts: {'fat': 1.02, 'nrg': 93.0, 'pro': 7.709999999999999, 'sat': 0.10500000000000001, 'sod': 90.0, 'sug': 4.65}\n",
      "Recipe 50329: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 50398: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 50410: 4 cup of broccoli nutrition facts: {'fat': 1.36, 'nrg': 124.0, 'pro': 10.28, 'sat': 0.14, 'sod': 120.0, 'sug': 6.2}\n",
      "Recipe 50576: 2 cup of broccoli nutrition facts: {'fat': 0.68, 'nrg': 62.0, 'pro': 5.14, 'sat': 0.07, 'sod': 60.0, 'sug': 3.1}\n",
      "Recipe 50849: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 50877: 1/2 cup of broccoli nutrition facts: {'fat': 0.16999999999999998, 'nrg': 15.5, 'pro': 1.2849999999999997, 'sat': 0.0175, 'sod': 15.0, 'sug': 0.775}\n",
      "Recipe 50928: 1 cup of broccoli nutrition facts: {'fat': 0.34, 'nrg': 31.0, 'pro': 2.57, 'sat': 0.035, 'sod': 30.0, 'sug': 1.55}\n",
      "Recipe 51039: 1 1/2 pound of broccoli nutrition facts: {'fat': 2.5174355999999998, 'nrg': 231.33192, 'pro': 19.186941599999997, 'sat': 0.26535131999999995, 'sod': 224.52803999999998, 'sug': 11.566595999999999}\n",
      "Recipe 51152: 1 pound of broccoli nutrition facts: {'fat': 1.6782904, 'nrg': 154.22128, 'pro': 12.7912944, 'sat': 0.17690087999999998, 'sod': 149.68536, 'sug': 7.711064}\n"
     ]
    }
   ],
   "source": [
    "length = range(len(df))\n",
    "for i in length:\n",
    "    rec_ingr = df.iloc[i][\"ingredients\"]\n",
    "    for ingredient_ndx in range(len(rec_ingr)):\n",
    "        if rec_ingr[ingredient_ndx]['text'] == 'broccoli, raw':\n",
    "            quantity = df.iloc[i][\"quantity\"][ingredient_ndx]['text']\n",
    "            unit = df.iloc[i][\"unit\"][ingredient_ndx]['text']\n",
    "            nutr = df.iloc[i][\"nutr_per_ingredient\"][ingredient_ndx]\n",
    "            print(\"Recipe {0}: {1} {2} of broccoli nutrition facts: {3}\".format(i,quantity,unit,nutr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Sgl3m2dvZtxh"
   },
   "source": [
    "We can conclude that there is a bug in the \"nutrients per ingredient\" section, because this is scaled to the quantity of the ingredient but is not consistent across units. For example, 10 oz of broccoli is less than 2 cups but has 34% more fat. Furthermore, some recipes contain erroneous quantities, such as 16 pounds of broccoli. Therefore, we use the USDA Nutrition Information Database to get nutrition statistics for each ingredient."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix No. 4\n",
    "The features of each ingredient are ten commonly-found nutrients parsed from the USDA database. If the full ingredient in the Recipes 1M+ database does not exist in the USDA database, match only the first word of the ingredient. Then, use an RBF kernel to create an adjacency matrix from the nutrition information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "colab_type": "code",
    "id": "beCMPm2rhk67",
    "outputId": "14f83b40-6d87-46df-fecf-17865005cb6f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "177450\n",
      "177450\n",
      "1534\n",
      "527\n"
     ]
    }
   ],
   "source": [
    "#Match first two words of unique ingredients to ingredients in USDA database\n",
    "usda_ingr_names = usda[\"SR description\"]\n",
    "#save first two words\n",
    "usda_ingr_first2words = []\n",
    "usda_ingr_firstword = []\n",
    "for ingr in usda_ingr_names:\n",
    "    word_vec = ingr.split(',')\n",
    "    if len(word_vec)>2:\n",
    "        usda_ingr_first2words.append(word_vec[0].lower() + ',' + word_vec[1])\n",
    "    else:\n",
    "        usda_ingr_first2words.append(ingr.lower())\n",
    "    if len(word_vec)>1:\n",
    "        usda_ingr_firstword.append(word_vec[0].lower())\n",
    "    else:\n",
    "        usda_ingr_firstword.append(ingr.lower())\n",
    "usda_ingr_first2words_unique = np.unique(usda_ingr_first2words)\n",
    "usda_ingr_firstword_unique = np.unique(usda_ingr_firstword)\n",
    "print(len(usda_ingr_first2words))\n",
    "print(len(usda_ingr_firstword))\n",
    "print(len(usda_ingr_first2words_unique))\n",
    "print(len(usda_ingr_firstword_unique))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "d0Ksgw9sm_ia"
   },
   "outputs": [],
   "source": [
    "ingredients_matched = []; #List of all ingredients that are found in both the USDA and Recipes 1M+ datasets\n",
    "ingredients_matched_mask = [];\n",
    "count = 0;\n",
    "for ingr in all_ingredients_full_unique:\n",
    "    vec = ingr.split(',')\n",
    "    if len(vec) > 2:\n",
    "        ingr_name = vec[0] + ',' + vec[1]\n",
    "    else:\n",
    "        ingr_name = ingr\n",
    "    if ingr_name in usda_ingr_first2words_unique:\n",
    "        ingredients_matched.append(ingr)\n",
    "        ingredients_matched_mask.append(True)\n",
    "    else: #match only first word\n",
    "        if vec[0] in usda_ingr_firstword_unique:\n",
    "            ingredients_matched.append(ingr)\n",
    "            ingredients_matched_mask.append(True)\n",
    "        else:\n",
    "            ingredients_matched_mask.append(False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "tYjIO9kvTqU7"
   },
   "outputs": [],
   "source": [
    "def getIngredientNutritionInfo(ingredient,stringType):\n",
    "    if stringType == \"firstWord\":\n",
    "        vecToUse = usda_ingr_firstword\n",
    "    else:\n",
    "        vecToUse = usda_ingr_first2words\n",
    "    nutr_info = np.zeros([1,10])\n",
    "    index_of_ingr = vecToUse.index(ingredient)\n",
    "    fragment = usda.iloc[index_of_ingr:index_of_ingr+40]\n",
    "    descriptions = fragment[\"Nutrient description\"].values\n",
    "    values = fragment[\"Nutrient value\"].values\n",
    "    for d in range(len(descriptions)):\n",
    "        desc = descriptions[d]\n",
    "        if (desc == \"Protein\"):\n",
    "            nutr_info[:,0] = values[d]\n",
    "        elif (desc == \"Total Fat\"):\n",
    "            nutr_info[:,1] = values[d]\n",
    "        elif desc == \"Carbohydrate\":\n",
    "            nutr_info[:,2] = values[d]\n",
    "        elif desc == \"Energy\":\n",
    "            nutr_info[:,3] = values[d]\n",
    "        elif desc == \"Calcium\":\n",
    "            nutr_info[:,4] = values[d]\n",
    "        elif desc == \"Iron\":\n",
    "            nutr_info[:,5] = values[d]\n",
    "        elif desc == \"Sugars, total\":\n",
    "            nutr_info[:,6] = values[d]\n",
    "        elif desc == \"Sodium\":\n",
    "            nutr_info[:,7] = values[d]\n",
    "        elif desc == \"Vitamin B-12\":\n",
    "            nutr_info[:,8] = values[d]\n",
    "        elif desc == \"Fiber, total dietary\":\n",
    "            nutr_info[:,9] = values[d]\n",
    "    return nutr_info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2T2Ue6-GLN5a"
   },
   "outputs": [],
   "source": [
    "ingredient_nutrition_facts = np.zeros([len(ingredients_matched),10])\n",
    "for i in range(len(ingredients_matched)):\n",
    "    ingr = ingredients_matched[i]\n",
    "    vec = ingr.split(',')\n",
    "    if len(vec) > 2:\n",
    "        ingr_name = vec[0] + ',' + vec[1]\n",
    "    else:\n",
    "        ingr_name = ingr\n",
    "    if ingr_name in usda_ingr_first2words_unique:\n",
    "        nutr = getIngredientNutritionInfo(ingr_name,\"firstTwoWords\")\n",
    "        ingredient_nutrition_facts[i,:] = nutr\n",
    "    else: #match only first word\n",
    "        ingr_name = vec[0]\n",
    "        if ingr_name in usda_ingr_firstword_unique:\n",
    "            nutr = getIngredientNutritionInfo(ingr_name,\"firstWord\")\n",
    "            ingredient_nutrition_facts[i,:] = nutr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "colab_type": "code",
    "id": "dqwGYNqFXdbs",
    "outputId": "4b66305c-4699-49f4-c0cb-e5dc5bdc56ee"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "557.630089343887\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "distances = pdist(ingredient_nutrition_facts, metric='euclidean')\n",
    "plt.hist(distances)\n",
    "mean_dist = np.median(distances)\n",
    "print(mean_dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "ssIwtErPgPUu",
    "outputId": "78b2ddef-88ef-4cd3-fac8-652ef54d5d4e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(ingredient_nutrition_facts, sigma, epsilon=0.2)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 367
    },
    "colab_type": "code",
    "id": "9T9IaNDYgY1T",
    "outputId": "8f841eb7-2f39-4e1b-c4e3-f66c1f6514cc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([36607.,     0.,     0.,     0.,  2354.,  2694.,  2780.,  2208.,\n",
       "         2086.,  2996.,  2576.,  2152.,  1960.,  2520.,  3672.,  3996.,\n",
       "         3928.,  5166.,  5888., 14386.]),\n",
       " array([0.  , 0.05, 0.1 , 0.15, 0.2 , 0.25, 0.3 , 0.35, 0.4 , 0.45, 0.5 ,\n",
       "        0.55, 0.6 , 0.65, 0.7 , 0.75, 0.8 , 0.85, 0.9 , 0.95, 1.  ]),\n",
       " <a list of 20 Patch objects>)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(np.ravel(adjacency),bins=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_kQ9ztVFgoZS"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Ingredient_List_USDA_Nutrition_Info.npy', ingredients_matched)\n",
    "np.save('../data/Adjacency_Matrix_4.npy', adjacency)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix No. 5\n",
    "Normalize the nutrition facts to the mean nutrition fact so that some nutrients are not weighted more than others"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 284
    },
    "colab_type": "code",
    "id": "uzSS-9NZsfrc",
    "outputId": "e6ebacfd-e5f0-463e-dc1d-70830ea16fde"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.096383664290723\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ingredient_nutrition_facts_normalized = ingredient_nutrition_facts / np.mean(ingredient_nutrition_facts, axis=0)\n",
    "distances = pdist(ingredient_nutrition_facts_normalized, metric='euclidean')\n",
    "plt.hist(distances)\n",
    "mean_dist = np.median(distances)\n",
    "print(mean_dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "iezpOYEhsti6",
    "outputId": "6646cebe-7e59-4f73-8cff-f1c7af92e181"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(ingredient_nutrition_facts_normalized, sigma, epsilon=0.2)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SyiDLhi3s0_n"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_5.npy', adjacency)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix No. 6\n",
    "Same as #5 but sparsified for vizualization and community detection purposes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.4\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(ingredient_nutrition_facts_normalized, sigma, epsilon=0.1)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([57503.,     0.,  3950.,  3142.,  3262.,  2466.,  2500.,  2216.,\n",
       "         1414.,  1138.,  1068.,   984.,  1406.,  1560.,  1196.,  1754.,\n",
       "         2150.,  2618.,  3338.,  4304.]),\n",
       " array([0.  , 0.05, 0.1 , 0.15, 0.2 , 0.25, 0.3 , 0.35, 0.4 , 0.45, 0.5 ,\n",
       "        0.55, 0.6 , 0.65, 0.7 , 0.75, 0.8 , 0.85, 0.9 , 0.95, 1.  ]),\n",
       " <a list of 20 Patch objects>)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(np.ravel(adjacency),bins=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_6.npy', adjacency)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9Xk-pK-guCM8"
   },
   "outputs": [],
   "source": [
    "#Save nutrition facts matrix and nutrient labels for future reference\n",
    "np.save('../data/Nutrition_Facts_Per_Ingredient.npy', ingredient_nutrition_facts)\n",
    "nutrition_fact_column_labels = [\"Protein\", \"Total Fat\",\"Carbohydrate\",\"Energy\",\"Calcium\",\"Iron\",\"Sugars, total\",\"Sodium\",\"Vitamin B-12\",\"Fiber, total dietary\"]\n",
    "np.save('../data/Nutrition_Facts_Column_Labels.npy', nutrition_fact_column_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Find vegan ingredients\n",
    "Vitamin B-12 is only found in animal products. Therefore, we find all vegan ingredients by thresholding the ingredient list by their B-12 quantity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 952
    },
    "colab_type": "code",
    "id": "re8_tK99u191",
    "outputId": "76ce4016-53f3-4df6-c456-79a492a4d9c8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Some vegan ingredients:\n",
      "['beans, kidney, all types, mature seeds, raw' 'beans, snap, green, raw'\n",
      " 'beets, raw' 'beverages, carbonated, cola, regular'\n",
      " 'beverages, carbonated, ginger ale']\n"
     ]
    }
   ],
   "source": [
    "vegan_ingredient_mask = ingredient_nutrition_facts[:,8]  < 0.05\n",
    "vegan_ingredients = np.array(ingredients_matched)[vegan_ingredient_mask]\n",
    "problems = ['animal fat, bacon grease', 'fat, goose', 'lard', 'soup, chicken broth or bouillon, dry', 'soup, cream of chicken, canned, condensed','soup, swanson chicken broth 99% fat free','margarine-like, margarine-butter blend, soybean oil and butter']\n",
    "indices = []\n",
    "for problem in problems:\n",
    "    index = vegan_ingredients.tolist().index(problem)\n",
    "    indices.append(index)\n",
    "vegan_ingredients = np.delete(vegan_ingredients,indices)\n",
    "vegan_ingredient_mask[indices] = False\n",
    "print('Some vegan ingredients:')\n",
    "print(vegan_ingredients[20:25])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 226 vegan ingredients\n"
     ]
    }
   ],
   "source": [
    "print(\"There are {} vegan ingredients\".format(len(vegan_ingredients)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jfuXIVPAchZ0"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Vegan_Ingredient_Mask_USDA_Nutrition_Info.npy', vegan_ingredient_mask)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "9tcziOO36hG8"
   },
   "source": [
    "### Adjacency Matrix No. 7\n",
    "Concatenate recipe fetures and nutrition featues and then apply an RBF kernel. Normalize to the feature mean.\n",
    "Potential problem: There are more recipe features than nutrition features, so these may be weighted more heavily"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "5Fd6YGj-dPCo",
    "outputId": "213e39ab-0c4a-4838-a844-a4d90e1c1360"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(313, 10)\n",
      "(313, 51235)\n"
     ]
    }
   ],
   "source": [
    "print(ingredient_nutrition_facts.shape)\n",
    "feature_matched = feature[ingredients_matched_mask]\n",
    "print(feature_matched.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6sJPrdgp7boJ"
   },
   "outputs": [],
   "source": [
    "#Try concatenating both\n",
    "combined_feature_matrix = np.zeros([ingredient_nutrition_facts.shape[0], ingredient_nutrition_facts.shape[1] + feature_matched.shape[1]])\n",
    "combined_feature_matrix[:,0:(ingredient_nutrition_facts.shape[1])] = ingredient_nutrition_facts\n",
    "combined_feature_matrix[:,ingredient_nutrition_facts.shape[1]:(ingredient_nutrition_facts.shape[1]+ feature_matched.shape[1])] = feature_matched"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "XE-lKtqL8PZe",
    "outputId": "80182e34-f77a-4bb2-ee83-a91b10edb604"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(313, 51177)\n"
     ]
    }
   ],
   "source": [
    "#Normalize features based on mean value\n",
    "#Possible problem: there are many more recipe features than nutrition ones\n",
    "feature_means = np.mean(combined_feature_matrix,axis=0)\n",
    "combined_feature_matrix = combined_feature_matrix[:, feature_means != 0]\n",
    "feature_means = feature_means[feature_means != 0]\n",
    "combined_feature_matrix_normalized = combined_feature_matrix / feature_means;\n",
    "print(combined_feature_matrix_normalized.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "colab_type": "code",
    "id": "J6CN9ZCh-IS8",
    "outputId": "73312a62-f025-4c9d-e1ba-a3664f91a270"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1829.2480483951592\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAD4CAYAAADsKpHdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAUZUlEQVR4nO3df6zd9X3f8edrduPSNCb8cCLXdndN4kQCNJn4yiPLErG6LQ6pAqmgs6UWV6NyQonULJM2e/kj2SQk6JoyoS3OnMKANOFHISlWCGsYVI0mEeglccH8cLkEN9zYw7ciIt7SsJq898f53PTL9fG1fc+17w8/H9JX53ve3+/nez7vOO3rfn+ck1QVkiT9o9megCRpbjAQJEmAgSBJagwESRJgIEiSmsWzPYHpOvfcc2toaGi2pyFJ88oTTzzxt1W1rN+2eRsIQ0NDjIyMzPY0JGleSfI3R9vmJSNJEmAgSJIaA0GSBBgIkqTGQJAkAQaCJKkxECRJgIEgSWoMBEkSMI+/qTxfDW17YFY+d98NH5qVz5U0f3iGIEkCDARJUmMgSJIAA0GS1BgIkiTgOAIhya1JDibZ06ndnWR3W/Yl2d3qQ0n+rrPt850x65I8lWQ0yc1J0upL2vFGkzyWZGjm25QkHcvxnCHcBmzsFqrqX1bV2qpaC9wHfKWz+YWJbVX1sU59B7AVWNOWiWNeA/ygqt4J3ATcOK1OJEkDOWYgVNU3gVf6bWt/5f8GcOdUx0iyHFhaVY9WVQF3AFe0zZcDt7f1e4ENE2cPkqRTZ9B7CO8HXq6q5zu11Um+k+Qvkry/1VYAY519xlptYttLAFV1GHgVOKffhyXZmmQkycj4+PiAU5ckdQ0aCJt549nBAeAXq+oi4JPAl5MsBfr9xV/tdaptbyxW7ayq4aoaXras7/9GtCRpmqb90xVJFgO/DqybqFXVa8Brbf2JJC8A76J3RrCyM3wlsL+tjwGrgLF2zDM5yiUqSdLJM8gZwi8Dz1XVTy8FJVmWZFFbP4/ezePvVtUB4FCSi9v9gauB+9uwXcCWtn4l8Ei7zyBJOoWO57HTO4FHgXcnGUtyTdu0iSNvJn8AeDLJX9G7Qfyxqpr4a/9a4I+AUeAF4MFWvwU4J8kovctM2wboR5I0Tce8ZFRVm49S/+0+tfvoPYbab/8R4MI+9R8DVx1rHpKkk8tvKkuSAANBktQYCJIkwECQJDUGgiQJMBAkSY2BIEkCDARJUmMgSJIAA0GS1BgIkiTAQJAkNQaCJAkwECRJjYEgSQIMBElSYyBIkgADQZLUGAiSJOA4AiHJrUkOJtnTqX0myfeT7G7LZZ1t25OMJtmb5NJOfV2Sp9q2m5Ok1ZckubvVH0syNLMtSpKOx/GcIdwGbOxTv6mq1rbl6wBJzgc2ARe0MZ9LsqjtvwPYCqxpy8QxrwF+UFXvBG4CbpxmL5KkARwzEKrqm8Arx3m8y4G7quq1qnoRGAXWJ1kOLK2qR6uqgDuAKzpjbm/r9wIbJs4eJEmnziD3ED6e5Ml2SemsVlsBvNTZZ6zVVrT1yfU3jKmqw8CrwDn9PjDJ1iQjSUbGx8cHmLokabLpBsIO4B3AWuAA8NlW7/eXfU1Rn2rMkcWqnVU1XFXDy5YtO7EZS5KmNK1AqKqXq+r1qvoJ8AVgfds0Bqzq7LoS2N/qK/vU3zAmyWLgTI7/EpUkaYZMKxDaPYEJHwEmnkDaBWxqTw6tpnfz+PGqOgAcSnJxuz9wNXB/Z8yWtn4l8Ei7zyBJOoUWH2uHJHcClwDnJhkDPg1ckmQtvUs7+4CPAlTV00nuAZ4BDgPXVdXr7VDX0nti6QzgwbYA3AJ8MckovTODTTPRmCTpxBwzEKpqc5/yLVPsfz1wfZ/6CHBhn/qPgauONQ9J0snlN5UlSYCBIElqDARJEmAgSJIaA0GSBBgIkqTGQJAkAQaCJKkxECRJgIEgSWoMBEkSYCBIkhoDQZIEGAiSpMZAkCQBBoIkqTEQJEmAgSBJagwESRJwHIGQ5NYkB5Ps6dT+U5LnkjyZ5KtJ3trqQ0n+Lsnutny+M2ZdkqeSjCa5OUlafUmSu1v9sSRDM9+mJOlYjucM4TZg46TaQ8CFVfVPgL8Gtne2vVBVa9vysU59B7AVWNOWiWNeA/ygqt4J3ATceMJdSJIGdsxAqKpvAq9Mqn2jqg63t98CVk51jCTLgaVV9WhVFXAHcEXbfDlwe1u/F9gwcfYgSTp1ZuIewr8CHuy8X53kO0n+Isn7W20FMNbZZ6zVJra9BNBC5lXgnH4flGRrkpEkI+Pj4zMwdUnShIECIcmngMPAl1rpAPCLVXUR8Engy0mWAv3+4q+Jw0yx7Y3Fqp1VNVxVw8uWLRtk6pKkSRZPd2CSLcCvARvaZSCq6jXgtbb+RJIXgHfROyPoXlZaCexv62PAKmAsyWLgTCZdopIknXzTOkNIshH4d8CHq+pHnfqyJIva+nn0bh5/t6oOAIeSXNzuD1wN3N+G7QK2tPUrgUcmAkaSdOoc8wwhyZ3AJcC5ScaAT9N7qmgJ8FC7//ut9kTRB4D/mOQw8Drwsaqa+Gv/WnpPLJ1B757DxH2HW4AvJhmld2awaUY6kySdkGMGQlVt7lO+5Sj73gfcd5RtI8CFfeo/Bq461jxm0tC2B07lx0nSvOA3lSVJgIEgSWoMBEkSYCBIkhoDQZIEGAiSpMZAkCQBBoIkqTEQJEmAgSBJagwESRJgIEiSGgNBkgQYCJKkxkCQJAEGgiSpMRAkSYCBIElqDARJEmAgSJKaYwZCkluTHEyyp1M7O8lDSZ5vr2d1tm1PMppkb5JLO/V1SZ5q225OklZfkuTuVn8sydDMtihJOh7Hc4ZwG7BxUm0b8HBVrQEebu9Jcj6wCbigjflckkVtzA5gK7CmLRPHvAb4QVW9E7gJuHG6zUiSpu+YgVBV3wRemVS+HLi9rd8OXNGp31VVr1XVi8AosD7JcmBpVT1aVQXcMWnMxLHuBTZMnD1Ikk6d6d5DeHtVHQBor29r9RXAS539xlptRVufXH/DmKo6DLwKnNPvQ5NsTTKSZGR8fHyaU5ck9bN4ho/X7y/7mqI+1Zgji1U7gZ0Aw8PDffdRf0PbHpi1z953w4dm7bMlHb/pniG83C4D0V4PtvoYsKqz30pgf6uv7FN/w5gki4EzOfISlSTpJJtuIOwCtrT1LcD9nfqm9uTQano3jx9vl5UOJbm43R+4etKYiWNdCTzS7jNIkk6hY14ySnIncAlwbpIx4NPADcA9Sa4BvgdcBVBVTye5B3gGOAxcV1Wvt0NdS++JpTOAB9sCcAvwxSSj9M4MNs1IZ5KkE3LMQKiqzUfZtOEo+18PXN+nPgJc2Kf+Y1qgSJJmj99UliQBBoIkqTEQJEmAgSBJagwESRJgIEiSGgNBkgQYCJKkxkCQJAEGgiSpMRAkSYCBIElqDARJEmAgSJIaA0GSBBgIkqTGQJAkAQaCJKkxECRJwACBkOTdSXZ3lh8m+USSzyT5fqd+WWfM9iSjSfYmubRTX5fkqbbt5iQZtDFJ0omZdiBU1d6qWltVa4F1wI+Ar7bNN01sq6qvAyQ5H9gEXABsBD6XZFHbfwewFVjTlo3TnZckaXpm6pLRBuCFqvqbKfa5HLirql6rqheBUWB9kuXA0qp6tKoKuAO4YobmJUk6TjMVCJuAOzvvP57kySS3Jjmr1VYAL3X2GWu1FW19cv0ISbYmGUkyMj4+PkNTlyTBDARCkjcBHwb+pJV2AO8A1gIHgM9O7NpneE1RP7JYtbOqhqtqeNmyZQPNW5L0RjNxhvBB4NtV9TJAVb1cVa9X1U+ALwDr235jwKrOuJXA/lZf2acuSTqFZiIQNtO5XNTuCUz4CLCnre8CNiVZkmQ1vZvHj1fVAeBQkovb00VXA/fPwLwkSSdg8SCDk/wc8CvARzvl30+ylt5ln30T26rq6ST3AM8Ah4Hrqur1NuZa4DbgDODBtkiSTqGBAqGqfgScM6n2W1Psfz1wfZ/6CHDhIHORJA3GbypLkgADQZLUGAiSJMBAkCQ1BoIkCTAQJEmNgSBJAgwESVJjIEiSAANBktQYCJIkwECQJDUGgiQJMBAkSY2BIEkCDARJUmMgSJIAA0GS1BgIkiRgwEBIsi/JU0l2JxlptbOTPJTk+fZ6Vmf/7UlGk+xNcmmnvq4dZzTJzUkyyLwkSSduJs4Q/kVVra2q4fZ+G/BwVa0BHm7vSXI+sAm4ANgIfC7JojZmB7AVWNOWjTMwL0nSCTgZl4wuB25v67cDV3Tqd1XVa1X1IjAKrE+yHFhaVY9WVQF3dMZIkk6RQQOhgG8keSLJ1lZ7e1UdAGivb2v1FcBLnbFjrbairU+uHyHJ1iQjSUbGx8cHnLokqWvxgOPfV1X7k7wNeCjJc1Ps2+++QE1RP7JYtRPYCTA8PNx3H0nS9Ax0hlBV+9vrQeCrwHrg5XYZiPZ6sO0+BqzqDF8J7G/1lX3qkqRTaNqBkOTNSd4ysQ78KrAH2AVsabttAe5v67uATUmWJFlN7+bx4+2y0qEkF7eni67ujJEknSKDXDJ6O/DV9oToYuDLVfU/kvwlcE+Sa4DvAVcBVNXTSe4BngEOA9dV1evtWNcCtwFnAA+2RZJ0CqX3YM/8Mzw8XCMjI9MaO7TtgRmejeaifTd8aLanIM05SZ7ofE3gDfymsiQJMBAkSY2BIEkCDARJUmMgSJIAA0GS1BgIkiTAQJAkNQaCJAkwECRJjYEgSQIMBElSYyBIkgADQZLUGAiSJMBAkCQ1BoIkCTAQJEmNgSBJAgYIhCSrkvx5kmeTPJ3k91r9M0m+n2R3Wy7rjNmeZDTJ3iSXdurrkjzVtt2cJIO1JUk6UYsHGHsY+DdV9e0kbwGeSPJQ23ZTVf1Bd+ck5wObgAuAXwD+Z5J3VdXrwA5gK/At4OvARuDBAeYmSTpB0z5DqKoDVfXttn4IeBZYMcWQy4G7quq1qnoRGAXWJ1kOLK2qR6uqgDuAK6Y7L0nS9MzIPYQkQ8BFwGOt9PEkTya5NclZrbYCeKkzbKzVVrT1yXVJ0ik0cCAk+XngPuATVfVDepd/3gGsBQ4An53Ytc/wmqLe77O2JhlJMjI+Pj7o1CVJHQMFQpKfoRcGX6qqrwBU1ctV9XpV/QT4ArC+7T4GrOoMXwnsb/WVfepHqKqdVTVcVcPLli0bZOqSpEkGecoowC3As1X1h5368s5uHwH2tPVdwKYkS5KsBtYAj1fVAeBQkovbMa8G7p/uvCRJ0zPIU0bvA34LeCrJ7lb798DmJGvpXfbZB3wUoKqeTnIP8Ay9J5Sua08YAVwL3AacQe/pIp8wkqRTbNqBUFX/i/7X/78+xZjrgev71EeAC6c7F0nS4PymsiQJMBAkSY2BIEkCDARJUmMgSJKAwR47lea0oW0PzMrn7rvhQ7PyudKgPEOQJAEGgiSpMRAkSYCBIElqDARJEmAgSJIaA0GSBPg9BEkzYLa+8wF+72MmeYYgSQIMBElS4yUjaYZ5+UTzlWcIkiTAMwRpQZnNsxPNfwaCpHnNX7WdOXPmklGSjUn2JhlNsm225yNJp5s5EQhJFgH/FfggcD6wOcn5szsrSTq9zIlAANYDo1X13ar6f8BdwOWzPCdJOq3MlXsIK4CXOu/HgH86eackW4Gt7e3/SbL3OI9/LvC3A81wfjhd+gR7XajmTa+5caDhs9nnPz7ahrkSCOlTqyMKVTuBnSd88GSkqoanM7H55HTpE+x1oTpdep2rfc6VS0ZjwKrO+5XA/lmaiySdluZKIPwlsCbJ6iRvAjYBu2Z5TpJ0WpkTl4yq6nCSjwN/BiwCbq2qp2fwI074MtM8dbr0Cfa6UJ0uvc7JPlN1xKV6SdJpaK5cMpIkzTIDQZIELPBAmO8/h5FkVZI/T/JskqeT/F6rn53koSTPt9ezOmO2t373Jrm0U1+X5Km27eYk/R71nXVJFiX5TpKvtfcLstckb01yb5Ln2r/vexdir0n+dfvv7p4kdyb52YXSZ5JbkxxMsqdTm7HekixJcnerP5Zk6KQ3VVULcqF3c/oF4DzgTcBfAefP9rxOsIflwHva+luAv6b30x6/D2xr9W3AjW39/NbnEmB1639R2/Y48F563/l4EPjgbPd3lJ4/CXwZ+Fp7vyB7BW4Hfqetvwl460Lrld4XTl8Ezmjv7wF+e6H0CXwAeA+wp1Obsd6A3wU+39Y3AXef9J5m+z/Uk/iP9V7gzzrvtwPbZ3teA/Z0P/ArwF5geastB/b265HeU1vvbfs816lvBv7bbPfTp7+VwMPAL3UCYcH1Cixt/48yk+oLqlf+4RcIzqb3ROPXgF9dSH0CQ5MCYcZ6m9inrS+m983mnKxeqmpBXzLq93MYK2ZpLgNrp4sXAY8Bb6+qAwDt9W1tt6P1vKKtT67PNf8Z+LfATzq1hdjrecA48N/b5bE/SvJmFlivVfV94A+A7wEHgFer6hsssD4nmcnefjqmqg4DrwLnnLSZs7DvIRzXz2HMB0l+HrgP+ERV/XCqXfvUaor6nJHk14CDVfXE8Q7pU5sXvdL7a+89wI6qugj4v/QuLxzNvOy1XT+/nN4lkl8A3pzkN6ca0qc25/s8TtPp7ZT3vZADYUH8HEaSn6EXBl+qqq+08stJlrfty4GDrX60nsfa+uT6XPI+4MNJ9tH7tdtfSvLHLMxex4Cxqnqsvb+XXkAstF5/GXixqsar6u+BrwD/jIXXZ9dM9vbTMUkWA2cCr5y0mbOwA2He/xxGe9rgFuDZqvrDzqZdwJa2voXevYWJ+qb2dMJqYA3weDt1PZTk4nbMqztj5oSq2l5VK6tqiN6/1SNV9ZsszF7/N/BSkne30gbgGRZer98DLk7yc21+G4BnWXh9ds1kb91jXUnv/yZO7pnRbN+UOck3fC6j92TOC8CnZns+05j/P6d3ivgksLstl9G7jvgw8Hx7Pbsz5lOt3710nsQAhoE9bdt/4STfnBqw70v4h5vKC7JXYC0w0v5t/xQ4ayH2CvwH4Lk2xy/Se8pmQfQJ3Env3sjf0/tr/pqZ7A34WeBPgFF6TyKdd7J78qcrJEnAwr5kJEk6AQaCJAkwECRJjYEgSQIMBElSYyBIkgADQZLU/H/EdrnzTTVq7gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "distances = pdist(combined_feature_matrix_normalized, metric='euclidean')\n",
    "plt.hist(distances)\n",
    "mean_dist = np.median(distances)\n",
    "print(mean_dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "1R-6sa9g-W1n",
    "outputId": "870bad2e-1924-4929-98eb-70689c69d039"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(combined_feature_matrix_normalized, sigma, epsilon=0.2)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "WXVwY33C_WmL"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_7.npy', adjacency)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qCpbapQo_OUR"
   },
   "source": [
    "### Adjacency Matrix 8\n",
    "Same as 7 but without feature normalization (to try to weigh nutrition information more than recipe features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "colab_type": "code",
    "id": "qAn6Pgv9_P8Z",
    "outputId": "86193c81-7f57-4a76-f052-cffd27eaafaf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "558.9922892776581\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "distances = pdist(combined_feature_matrix, metric='euclidean')\n",
    "plt.hist(distances)\n",
    "mean_dist = np.median(distances)\n",
    "print(mean_dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "Ut4-kEWi_wvM",
    "outputId": "a8914993-9564-4f63-f34b-ee065b007ed7"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQIAAAD8CAYAAACcoKqNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO1dXehlV3X/Lcc0ihFMNAn/JmMzyhQ6So35D1NLitjWNjEviQ+R8UECTRkfkqJQKYlCG1sEW6pvVTpi6FBs0gGVBCm16aBIQRJnbL4mcZrRBDPOkKlfGPuQmrj6cM9p9uzZH2t/nbPPvesHl3vvOXuvtfba+6yz1trr3EvMDIVCsdl4xdwCKBSK+aGGQKFQqCFQKBRqCBQKBdQQKBQKqCFQKBTowBAQ0fVEdIKIThLRHTPwf4aIHiOih4no6HDsEiJ6gIieGt4vbsT7biI6S0SPG8e8vInozkFPJ4jousZy3EVEPxj08jAR3dBajoH2TiL6GhE9SUTHiehDw/E59OKTZVLdENGriOghInpkkOPjw/F6OmHm2V4AdgD4LoA3AfgVAI8A2DOxDM8AeIN17G8A3DF8vgPAXzfi/U4A1wB4PMYbwJ5BPxcC2DXobUdDOe4C8BFH22ZyDPS3AFwzfH4tgP8aeM6hF58sk+oGAAG4aPh8AYAHAbyjpk7m9gj2ATjJzN9j5v8FcC+AG2eWCVjJcGj4fAjATS2YMPM3APxYyPtGAPcy8wvM/DSAk1jpr5UcPjSTY5DlDDN/e/j8PIAnAVyBefTik8WHJrLwCj8fvl4wvBgVdTK3IbgCwLPG91MIK7oFGMC/EdExIjowHLucmc8Aq8UA4LIJ5fHxnkNXtxPRo0PoMLqdk8lBRFcBeDtWd8BZ9WLJAkysGyLaQUQPAzgL4AFmrqqTuQ0BOY5NXfN8LTNfA+A9AG4jondOzF+KqXX1WQBvBnA1gDMAPjWlHER0EYAvAvgwM/8s1LS1PA5ZJtcNM7/EzFcDuBLAPiJ6a0jkVDnmNgSnAOw0vl8J4PSUAjDz6eH9LIAvY+VCPUdEWwAwvJ+dUCQf70l1xczPDYvvlwA+h5ddy+ZyENEFWF14X2DmLw2HZ9GLS5Y5dcPMPwXwdQDXo6JO5jYE3wKwm4h2EdGvANgP4P6pmBPRa4joteNnAH8I4PFBhluGZrcAuG8qmQK87wewn4guJKJdAHYDeKiVEOMCG/BerPTSXA4iIgCfB/AkM3/aODW5XnyyTK0bIrqUiF43fH41gHcD+A5q6qRWtrcgI3oDVtnY7wL42MS834RVdvURAMdH/gBeD+AIgKeG90sa8b8HK9fyF1hZ8VtDvAF8bNDTCQDvaSzHPwJ4DMCjw8Laai3HQPt3sHJjHwXw8PC6YSa9+GSZVDcAfhPAfw78Hgfw57F1mioHDZ0UCsUGY+7QQKFQdAA1BAqFQg2BQqFQQ6BQKKCGQKFQoKEhoMSnCo3y3lnRixyAyuJDL7L0IgdQLksTQ0BEOwD8HVZlu3sAvJ+I9kS69aLUXuQAVBYfepGlFzmAQllaeQS9PlWoUCgceGUjuq6nn37LbDC4MqMV237jG98IIopWN21vb+PYsWNFwoVojHLYbba3twEAx44d+//PLuTK5pPJJUsqTVN2V5sR9nmbr2R+pDKltpPKYo/ZhK2LHL2a9M214oLJw8UrpPscjDoJjOuHzHyps2+LykIiuhnAdcz8x8P3DwDYx8x/4mnPwKrceVXe7YekTQwhGuM5u82op/GcD7mySWTKpWnK7mozwj5fQ9c+mVLbpfZzzZGti5zxuXTqWw8mDxevkO5LEBjXMWbe6zrRKjTIegordpHVQk2l175QFIo50MoQZD9VOBqD2gbBpFeT9lTPaizpmRDHQzOT81ekoYkhYOYXAdwO4KtY/bzTYWY+7mvvi7FiC8o+7nlyK8mdG9ub765zIUgXot3ONx6XzL5xhnjYbUe9jC+fru0+MZg0XXRD331jdMnieh/5uxCiE0JIz6lGJ2c9u8bge4rQN1cxObt4+tCV+PHFtaXxcoxGrzmCHNpz5Qhs/ZTE+jn9fNAcwfQ5giqYKmegUGw6ujYEQD1j0Cqpp8nCNlC9TovuDQGgnsEmQud7WrQqKCqGfUfo1RhMJdOS7pBLklWxQheGYHt7G0ePHhW3z7n4XFn3EH27vZn99p3Llc0ln/RcCs0YjdxdkhQ5pGMJ6VjSz04Mx3YfUuRJWUdSXrVvKC56IQPd7a6BD7kZ9BT6Pe4a5GT0JbsGuTLl0pli10BSWWgfl9B10dFdgxnRU5gwlxtci29oX7sWegwVRpmkso01EeuKbg1BtAAi0RiUFIFI6bZEL4ZPAleBy9T8FWno1hBI0JNnoFAsGd0aghSXTZpEaoGp3MUluaV2iXEujdD3lL6KOLo1BClI9QyWtlCWJq9ieejWEKS6/CnGYGnhxFyxdi40R7A8dGsIcqB3zvWFXtxtsVaGYESrwhgpL0U5VK/TYi0Nge4mKBRpWEtDALhLVFvxUdSH6nVadPGsgQs1FoKvtLN0S8vXX43NCrXLknP4T+0R2uXES0O3hqBWnbur1r7kl3dCE11LZhddYDkGwdbP1HLPuVOxRCMAdBIa2L9Z6HtiDDj/Z6/s86Ftq5Qn4FxehOkRuIpmTEMjgTR0CZ0LbdG5xmku2FHP5hhCegvxkjzZaM+rTdc8Zrez5fO91yhmSkHuhe/Ts2++fP1dOvTRCqHbpw99d9eSu/nAS/Skn+vpN7O/D62ePsyhbXswPn2OyHnC0cfXRM5ThC6aqU8J2nA9NVhDpzFDvISnD7sNDWpb89KwIER3CmO6lLAAaJMjSC0x7uEGtyR0ExrYbs34HqpSi7mYLnq+kMOm69p1sEMAV/IwxN/Fx+f6mufNc77xumCPwaUfV2gQCydyEJJbEhpIaIz9xvcSgxRbH5J2vr4+PdrHQ/p2zWUoNIjJuVGhgX2xhehraKChQUgODQ0mQqvQwPwcWgiu466k4BRuaMuwoAXtuWmmJm0VnYQGcyG2Feg7VuIiK6aBzk8aijwCInoGwPMAXgLwIjPvJaJLAPwzgKsAPAPgfcz8kzIx2yHXRZwSvcvXAq3HrIbiXNTwCH6Xma82Yo87ABxh5t0Ajgzfu4ZmmRWbjhahwY0ADg2fDwG4qQGP6kgtMpryDr1p3sCSsdS5Kto1IKKnAfwEAAP4e2Y+SEQ/ZebXGW1+wswXh+js3buXU/7XQKFIQevk4VI8SiJqtmtwLTOfJqLLADxARN9JEOoAgAPG93PO194+lNLwben4totc31vJtMnbh7m8zWO6fehHkSFg5tPD+1ki+jKAfQCeI6ItZj5DRFsAznr6HgRwEPB7BC4F51pf3yKQFo6EFpGEXuyu5DM6Jv2QAXIdcxUUhfiObcx+te50UkNqyhTTvwux8YRkk9Cz20voSnQp0XOKZyOZ63PO5U40Eb0GwCuY+fnh8wMA/hLA7wP4ETN/kojuAHAJM/9ZhFbzgiIpDd9dVD2CNL4m5vYITDk23CNoEhpcDuDLA8NXAvgnZv5XIvoWgMNEdCuA7wO4OYd4D0U0PW0ttpJhiuKfkgsuh4bEa1xCTD8lsg0BM38PwNscx3+ElVewFlhKIigXud5GCnowpDbWfV5T0UWJccq/IdfMEfjgiyt9rp4rvi+RT3JOoofcHEHoew5SdO/qk9rPfDfny9c/ha7dfsocQQpScwSLKzFOVVgwQWKdC13QoYvfbO/jRxR/Gs48n3MXrXnnTaGV0jaUP5DSC513XaQSr8c3r1JZfP3ttuarJ2zM04dSxPjGkkUtkoWS8zHaObItNVnoM+iaLOz86UNfaOBzb1KMV+ji9dFynTMtvq3oGD2zf8551zlJuNBLaGDSkbj80vG6UMNwxeitY2jQhSFIQQ2FSWJx1wXhs/ASmVKz2CFZpMdK5IpdeCkGOXTh5I5Nwksih+TmINFDys1AQndqLC5HUAO+Czjkqrm+p7jcsbjQPh9zG6U8pXKF+LuQ6pX5ZPKNLTVWt/tK9G2+SxKvIe9AanRDScs50Z1HkOMq5l5AsRyEpH8oTJAgZcGHjFFIXkk86+ITyouM5026Nb2IUH9feBHTuSSed32P3QRycwQxmpL5TfVCfOjOEMTcwnHQkjgtFP/H+vkmwZ5YX5iQ6sqm5hckLqwkRxCSyfwsce8lXoREJt9FEwrbQv1cayZG2xU6uGQOzUFujiA2zlDf0DnNESA8eVI+oUVS6u5JJ1PSvnabUuTwSInTY+0kc5lDd52w9jmCkJsv6eejJYlDS/lJzuXSnBI5cuTmCMz2c4y/F52nots6Ah9S4u8cGr6YzhdShHIatWTKhS+2t9uMcGW1W8nUCqGY2ReHl9CP5RZCOQKp256KwLi8dQTdegS1DVQNehK3s0aY4OLRg8GWYpS3RO7c0CC1bW0saZ5MdJsjqH3XaHUX8nkEte98S3I5a8haGhoo0tCtR7B0tPAMFDKo0UjHRhqCqRbKkoxBbxfPUvS2LthIQzDlIluKMehdxt7lWzo20hBI7n6x7UMpnbFdbCHPtX1Yexu0FnJzBPYWr0KGjTQEc2ApnsHSoTrOgxoCD6SVhSkLL2QMUisLJZBWFva4PVmyfTi2721MPUMNwcTo1TNQN3qzoYZgBvRoDHqTRzEtujUELSsLe1j0qcagB5mlqFFZWMpfkQatLKxAN3fhpVQgLsl118rC5aFbj2BTULq1qDgfqq90dOERLOF/DXw/TGLSS/1xER/vkOwpxyQ/AhJ6Ik4ir08W3/mYjKZMsXY+XjYNyXzHZJf+MImLruS3D6Z4yC5Yq9JDPOV6DNnnMk/1GGvseOix5Nq8Y+ckdHNkS+Hp04fNd+r5kzwinEM/xUDGjMnYthaaPIZMRHcT0Vkietw4dgkRPUBETw3vFxvn7iSik0R0goiuyxlIa5Tuy0ssfA7t2rsJU9YR9OSOT5mobFn1OSUkOYJ/AHC9dewOAEeYeTeAI8N3ENEeAPsBvGXo8xki2pEjWEtFSWjP5Sm5jEHru2htmGXLpV6M73uMvyINUUPAzN8A8GPr8I0ADg2fDwG4yTh+LzO/wMxPAzgJYF8lWWeHucBaG6pa4YaEV491+UvbNUh9/qQ35O4aXM7MZwBgeL9sOH4FgGeNdqeGY8loeUdOSTr5+rX2GCSJvhpo4UbPXUcATH9BtggXp0TtXQOX9p2aIaIDAA5U5r9WyE1oKZZ7Qc6FXEPwHBFtMfMZItoCcHY4fgrATqPdlQBOuwgw80EABwFg7969XPrfh7HjoYyxnVUPbR+6+vvo2XKEzvky66YcKZln1xik24chHi5Ikpw1tw8l27Q1tw9D7VO2D0PtU7xUaRLY19+F3NDgfgC3DJ9vAXCfcXw/EV1IRLsA7AbwkISgHau6hDYvuNR4PbSF5oqRzfYjX9vddX13jceW1TVOk5Z9zDUW18UTu8DMcy74xmH3cY3NPma+ckIcScLUls02mqFdHtu4+l4mrVQvI6Q7X1tXG1uWFA9RnP+xF7hjwd8D4AyAX2B1x78VwOux2i14ani/xGj/MQDfBXACwHti9Ic+bL94dUJ8POUVojGes9uMMD/7vreQKYe+LV8q7dzx2KhFM4e3b65yZfHRCfHz8QrpvuQVoHeUPddgtwVFPnCiRUylMZ6z24x6it15S7bLpJ6NlIc5Fp9sIdlr6NonUyvYYzYRCg9z6cc8L996GmmZbWshMK7l/a+B4lxIFl5PmFPGJeinN3TxrIELkrt2LiR9ffHo+NmUocbFmRr3SXSQM04pQl6R3aalB2AjdNcNQeI9+frloDdj1a0hCKFlaOBr47ojh9zQmjKF3NzQOKTjNGlKETKUUt4tkDsXoSR1rF8Oz948u40JDVopfc7J7G0xjTANZa07Zo/jXCdsjCFYV/RqDBTLQrehQW23MrXuoDeEZM7NgLfC3HKocUxHtx7B3BMZKvpoYVQkF3qsv91GmiwUF50IYe9R5yCWg5D0n8Igzb1Oa6FbQzD3XSWUBW8x+ZILPYacKr7SC9YnRwsDk4LaY/Jh7nVaC90aAkUe1C1W5EANwQKQemGrMVCkootk4ZJ+vNTsbxcUxcpNR9qp533xv+RYLFSIFeDUMCgh3YdyG5KiJRcvm4ZkvmP0fO0ldCUhZW3DLZnrc871cOdwPWvgy4LnZsfNfkt71iC3Qk+SPGzxrEGIr5SmS//Sfj7oswb+Zw269gh8Cq5RpJJiyUO8a95BY31LjXYO/VpVkzk0a3snrvnPobmOHoHmCBQKxeaEBlIa6xYaxORrERqE+tYI7WLtfNDQQB9DbpZF78GQSjHVboLpCtcI4xTt0UWOwIWWJcZLRM0KxhpeVYzHnNDt03R06xFMkTxZEmrK3/pCGT2BEo+glL8iDd0aAkVb6F1TYaKL0MC1fVhSBOJCSqLITvDYbUPJxFDiKOW8La/dV1JQJH32wJcUDfWx+8WO++TwjSM0FknhlotXCkJrJXX7MKetCynGW7cPFUlQz0ABqCEQoST55eo7dzLNRqoxGO/Y6465xjgH342sIwghZb865FLPJZevLxCXzTWmuUqMayEWXtSozfAhFF6GdFQKrSMIQBrX+dqYGXB7n7yGMW1Rfiod5/jqOUxIiY2n3LHIvYB786g2xhCkKl7afooJnXLR9GoMUnTQ20W2BHSxa+DC3AUvLvfKtXMwFVrpw5fDKBnbJl6IdhiwNEQ9AiK6m4jOEtHjxrG7iOgHRPTw8LrBOHcnEZ0kohNEdJ1EiO3t7fNcOt8LON/9q/ka6dt8zO92GxOlfFuNJZV+6Xh6mD8bzC//4en4uYZOfeshtJ5qrRvXyzeuIAQDfyeAawA8bhy7C8BHHG33AHgEwIUAdmH1Z6g7BDxY+uJVh2avkb7NZ4T52fe9hG/NMdvy+dr4zpeOqWQsLv3njNk3V7l69dGJ8Qjpd0Ide/8ENeoRMPM3APw41m7AjQDuZeYXmPlpACcB7BP2tfnmdBPRq0m7tpxz87GR4+rai2xqzKWruXmXoCRZeDsRPTqEDhcPx64A8KzR5tRw7DwQ0QEiOkpEzt8oa/nQUU3aU8XDc8bduXUGc9UbbGKOohS5huCzAN4M4GoAZwB8ajjumgHnCmLmg8y8lz37muoRzMPHh9Ty1k31CJaKLEPAzM8x80vM/EsAn8PL7v8pADuNplcCOF0moqIXLDUjrogjyxAQ0Zbx9b0Axh2F+wHsJ6ILiWgXgN0AHsrkkdNNRE9Dg3z6EmPQIjTQOoK2iNYRENE9AN4F4A1EdArAXwB4FxFdjZXb/wyADwIAMx8nosMAngDwIoDbmPmlNqLPi7nvjszz/ddhanluDVlrj3ecOyld6VzPvS5y0e2zBj60vgBG+jafUU+xiS55HsDX13VOogdzLDmyxXikXERju5w+Of1ccxQ6LqHrohNaD6FnDVw0ayGgK33WQFEfS737Kc6HGoIE+MpxW9Gegm8ppDmDHLolNGLtx/MpuRFX29aGMJd+ar+NMQQ1JsxFo9ZCCG21teRbAzFjkCOr3SenqElyPnVL1EYvSdzSfhtjCKZ8aKcHWlNDw4RlY2MMQSvo4n8ZagyWiy4Mwfb29jnfTbctVKFmH3c9TGEft3m4EHIbJQs9xd20v7vGE6Jrj1PCw0XbPJbSx4bPGITmwzfPoXnzvYdQK96WrqNUWWLzb/ZzrXWXTqVydmEIFGlIjW/tfq2hnsHyoIZggUjNeNv9poAag2VBDYGiGdQYLAfdVhaGKsmkVWY5kFa05Va+1ZIrtS9Qt7LQV2nnQ2plYS3EKgJTqw1DfVw0pHrKqRz1IdDXW1nY7W8W+pCrILOfhMbYxuwDnDuZrkROqxLj2jRdtEse7HGVY5vnUnWTa2hzdFVimKRG0GWMQ/2m3krW0KAQS977nxIaJvSNLjwC138fAm6Lmbug7H6x7UOfW2eeCz2cVCqf61wq7ZQx+87XuHhD228+pLjVrn629xfbhkxByjqS8qptJFPCHWCBOYLWoUHsAp/r6cMc2rZBi9XK14pTXaGBfU4amqXKEpsbV44gBS6dxgy4Pn2oUDigYUJ/UEOgmAVqDPpCFzkCF2on4Ux6EtqhR39dcWhN+VLO5dIspV2DZsg9D+1KpPBubWzWxZh16xG0TJ5IaIcSZ+Z7LTljiaTSpFaMdk19mzRTttZsGqHvNXjXwLrsGnVrCBSbAw0T5ocaAkUXUGMwL9QQKLqBGoP5oMnCBP6aLKzHN9TXlUDUZGFbdOERxP4WHXg5CWRWibkSQr5+5nkboeSSydcH00DYPFz8Y8mslAq/VFqpybTQvLj0nDp/rr6hOXbNa2lyMCRzCVx0JOtWMjZf4tqn0xi0stDTzm4z6kkrC2V8TZRUCErmLMTblsOeR60sXKELj0ChcEFzBtMhagiIaCcRfY2IniSi40T0oeH4JUT0ABE9NbxfbPS5k4hOEtEJIrqu5QAU6w01BtNA4hG8COBPmfk3ALwDwG1EtAfAHQCOMPNuAEeG7xjO7QfwFgDXA/gMEe0IMXDlCAB33JSSI3D1G2SMLi7TJR3727mA8VgoR2B/98nuCkNcstgIxYH2OH16MceQG7vG4lHf/Pl0YrZz6TI0z/aYYgiN0dfe9TlE38VLKkOqHkLHk5QQegG4D8AfADgBYGs4tgXgxPD5TgB3Gu2/CuC3IzRZ+uJVh6JXiMZ4zm4zwvzs+95KplyaIdlCstfQdU2akr6uOXHNVa4sPjohfj5eNdZNop6OsucaTMoRENFVAN4O4EEAlzPzGaw4nwFw2dDsCgDPGt1ODceSwAJLOye91nTn4lMD9iLLpWF/l4YJS9JVLxDXERDRRQC+CODDzPyzgNvlOnHezBDRAQAHAvykoolQs47APtZ64bWsZ++tjsBHw1fDEeurRkEGkUdARBdgZQS+wMxfGg4/R0Rbw/ktAGeH46cA7DS6XwngtE2TmQ8y8172bGe09Ahy7yrSY7VRcmeV0q5Jv5VHMCJmfFuMad0h2TUgAJ8H8CQzf9o4dT+AW4bPt2CVOxiP7yeiC4loF4DdAB5KFaylR7BELEn+nIRdDg+90OtBEhpcC+ADAB4jooeHYx8F8EkAh4noVgDfB3AzADDzcSI6DOAJrHYcbmPml1IFi7l/c9ObGkuS375Ac+SWhm9L0kvP6Lay0IfWE28mpSSVha7vJXxTz0lomrK72oyoVVkokSm1nVQ/oaSiVhYusLJw7hxBDt2W6MFgS1EjR5ACV72EIg369GEFulMt9qVgDllTC8YU56ILQ+D7XwMXcic5ZaG43Euz0s129cy2MTcxZ1yuc9LqyNidMuSu+vqkIqR739gk7WI8gXPn0tdfQte3JRlbR673XBlSIJlrE10YAsD9hJpvu84uhw2V4IZi+FjMLGkf4ufbC7d5SfbI7XjTdV4KiVEa5XPxLbnjlyx4nw5svUnzO2Z/l5ylno0rRyAx9r71bcsnlUHSvhtDIL0TuZSZYml9iyTWP9UI2LRDExJbuBKPw3eBhMbkg2sMobug1LBI2krlch0fdRDy2FLp1jB6qR6Bb31LL2of3RC6TRZOgdQFOXXcuZQ4t0c51zFP0DL52o1HIEWNu4mEhs8qSyy3hF7KeZ8XETuWM84WSNVZ7FiurmKejpTeOmJxHkHrjLTpTkr51pCpBf0edxqmnL8ex98rFmcIprLQNe/2Ofxq0O/xbtZaJvNOP9X4ezM4OfJ0W1noS9TUyOaGIOXry0S3kq1k3LnZ/lyeoZ2XqecvlnQurSwMIVRZGNJRKQLj8lYWdpEj8NURxLLErZDCNycuz0Up7dT+tRJuU+pIQl+ynVdC326XumtQipwt5sWFBgpFCdZxN6EGFhkaDH2q8c9xZUOhQgv0FhrY4ZC0jqB1aGAjJmfq3TMnNIjxcvErne+NCA1i50ohdWWlW1+10GtokLoFN9fNx7xAYu56jbVXmnDO1VNOaKAegZBvrF0vcvn6AuoR2GW8UplteqUegc8zqYXFegQutN63j0GyS1DTiMYuutzsfStdpdRZlGTnU2lIL9TUCzG2+xCC6zmC3nIVmixMgF0zbx+bEj0toqWj5kXZW02BFGoIEmAulhZeQQqWuuB6Raut0qVADUECbI9gzjLWpS64nlFiDOb2EEuhhiABtkcwvhTrg9Iio6Wuh400BLWruOZALb5TGLSs2veEZKQNaUbfR9eX1V/q3V6CjTQELYuRloreFnlJ7UGNsfSW1W+NbrcP54bvTuE6V2PBtNgulfTrzQCMKPEIzPYlc5P7YNIS0a1H0KM1duUIWtDOOV/Ct8VY1iWHEvMMcsfXm14W5xHMZaVjHkLpxC4t15BSWTglcguvQnRK15xr27k3dOsR+NCrIhXrjVrGrteko+RPUHcS0deI6EkiOk5EHxqO30VEPyCih4fXDUafO4noJBGdIKLrYjy2t7ed7qR9zLd9V/MlpW2360WumjoL9TEhCWtqjKNk7LacpiwxmV16lMDF3yVXbZ3E5Hch+tARrf7yfIuZv01ErwVwDMBNAN4H4OfM/LdW+z0A7gGwD8CvAvh3AL/OgT9CTX3oqNSihmiM5+w2o55ij5Y2eHT0HN45NEP9Q7Ln6tpeUzUeOpL2C61n10NHOTr10fTx8/GqsW58MlZ/6IiZzwA4M3x+noieBHBFoMuNAO5l5hcAPE1EJ7EyCt/0dZjjF4pCNEJ3i9jdsEQ2yZ21hGYO/Ra6zqVZKkuKRyBB6OnCkbb0icMaejbhohcyNkk5AiK6CsDbATw4HLqdiB4loruJ6OLh2BUAnjW6nULYcCgUi0NPCdIaEO8aENFFAL4I4MPM/DMi+iyAvwLAw/unAPwRAJfZOU9jRHQAwAHj+7kdKocGLksdaucKDSSWvjSzHHLfYy62T2ZpaCDh4ZM9dnfz6T7kMufKkooYXUk44FoPHYYGXogMARFdgJUR+AIzf2lg9pxx/nMAvjJ8PQVgp9H9SgCnHcIeBHAQAPbu3ctThAYSN9mcyJBLK2lTKmPsXGroIpEtdzwpYYeUR03dhu7gqXRDeg+tkZS5LYFvrEWhAa16fx7Ak8z8aeP4ltHsvQAeHz7fD2A/EcHsqfIAAAYzSURBVF1IRLsA7AbwkGQALVCaDPLRar0FNPcW09z8a0N6oaWO227v6z9uG5qvVsgxKhKP4FoAHwDwGBE9PBz7KID3E9HVWLn9zwD44CDEcSI6DOAJAC8CuI0DOwatUdPSpt5da/GaA3PznwupBsPltbhCSFfblsnCVEh2Df4D7rj/XwJ9PgHgEwVyVUNKUifW1hdv28dqTOrcyahQtts8vy6IjdeGLw9gnhs/j+siliOYU5+LKzFORYpyS+LcVF6lsrRGLK6dW77aSN1OtI2+JFHYs0ewuBJjhaJ3zH13z0G3P2fuQ+72YSp935aYr4jE/F7CN/WchKZ0+1C6hVuC3C3JVPqxcMa3pZdKX3LRd7R92PfPmfsqC32YKlGXuj3n65PKN/WclGZK6FPKN5VPSTtpP8n2r5S+Kx/kqiw0208ZGqRuH6pH4KG/KR5Brky5dHI9gtR+KQnO1EKlkEcQOxeiWQvqEVTCpnoENfnWkCNXt6kJztQxuu76rhqT0G6C3Uc9AizHI3BN8BQeQS6WnCMopT91jsC1HnzHTFqmbLWQ4xF0u2vQ0kD1YPxSsDR5bdTwRFJoSNqm5ghqzEHPuwndGgKFYgq09E58/Ho0BhtjCFoov4XbLD03ZzZfQsd89YLcXZGSHQYXejQGXScLayV3XP0kF13o4qstW4pcKbxSk3StDM6UycJU1EgWunJLoVzSiJpl6baMPl4ubIxHoFD0hp48AzUECsWM6MUY6Pahh75r+xBYv4KiXrcPQ9u3kn5Tbx/GeLh4heoKSqDbh0J6NWlPZUh7MNhSzJ0snFNXubzn9gy6NQQKxaZhTmPQrSGo7ZK6SkBr022J1nxq68R8TY05eNbCXMagW0OwFCzJZQ9h0yo5e5RpxBzGoFtDsJSFOdXdp+eFa8POEeToyO6TQkOiq969hqmNQbeGoDZ6KUZZR0xR5biJep7SGHRRWejCUnIEU6FnmUvu3i0wdwa+JnK3OVOxMR6BQrFUTGHY1BB4ILXAU9z9WvKYM7sfQqmX0dt4StHaGHRrCNbFtauFljmO2oU/cxcUrStaGoONyRGkYoq4TIqeZJGgB1nX1QC1yhlsjEfQqsR4KixJ5hoeQcmuwZJ0lYMWnoHkT1BfRUQPEdEjRHSciD4+HL+EiB4goqeG94uNPncS0UkiOkFE11WVOBM93KUUcvS2E9EbahsDiUfwAoDfY+a3AbgawPVE9A4AdwA4wsy7ARwZvoOI9gDYD+AtAK4H8Bki2lFN4kxoHcGyoHUEcdQ0BlFDwCv8fPh6wfBiADcCODQcPwTgpuHzjQDuZeYXmPlpACcB7KsibQFa3VH0TtUGqlcZahkDUY6AiHbQ6i/RzwJ4gJkfBHA5M58BgOH9sqH5FQCeNbqfGo7NCvUIlgXVqxw1jIHIEDDzS8x8NYArAewjoreG5HKROK8R0QEiOkpE8n82USgUTpQag6RdA2b+KYCvYxX7P0dEW4MQW1h5C8DKA9hpdLsSwGkHrYPMvNf3iykKhSINJcZAsmtwKRG9bvj8agDvBvAdAPcDuGVodguA+4bP9wPYT0QXEtEuALsBPBTisb297dxyso+FjsdeZr8QjfGc3WZETNE5sklkyqFtyxui7To/Lqwcvj660nGU9HPB3H+3f0KsVKehn5iz36W6L1lHkrm2ISko2gJwaMj8vwLAYWb+ChF9E8BhIroVwPcB3DwIcJyIDgN4AsCLAG5j5pdCDI4dOyb+vbyQ4kMw+8Umz1W0MSoyZnVzk1ySBVXj9/V8tF3npbr26cpFt4SmtF+Inj2PNXTq42kbHJd+zba1kLNeuv3xUjUE557LoT2VIQjRtOnOZQhGOTbFEHh4Lu/HS2tvH5n0atKeaptrSdtpo/s9vubgr0jLGXRrCBQKRTmkxqBbQ9AyZOkhHErFkmROSVL5UHJXX5KupoDEGPSSI/hvAP8D4IdzywLgDehDDkBl8aEXWXqRA5DJ8mvMfKnrRBeGAACI6KgvkbGJcgAqiw+9yNKLHEC5LN2GBgqFYjqoIVAoFF0ZgoNzCzCgFzkAlcWHXmTpRQ6gUJZucgQKhWI+9OQRKBSKmaCGQKFQqCFQKBRqCBQKBdQQKBQKAP8H3DC4n6FG5Y8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(combined_feature_matrix, sigma, epsilon=0.2)\n",
    "plt.spy(adjacency)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "i6Z-6jGD_3zJ",
    "outputId": "02c73501-5772-45a4-d9b6-9f35809ba454"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "laplacian = compute_laplacian(adjacency,True)\n",
    "lamb_norm, U_norm = spectral_decomposition(laplacian)\n",
    "compute_number_connected_components(lamb_norm,threshold=1e-12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Wgxxm_Qt_-Ie"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_8.npy', adjacency)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UvioreppC1eu"
   },
   "source": [
    "### Adjacency Matrix 9:\n",
    "Try to combine the ingredient coocurrence adjacency matrix (#1) with the nutrition adjacency matrix (#5) by adding them. Make sure that the coocurrence weights are not weighted more heavily than the nutrition weights by dividing the cooccurrence matrix by its maximum value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 918
    },
    "colab_type": "code",
    "id": "5rIZOkfiD7fh",
    "outputId": "c3e7bd9f-8d6a-457f-f2dd-ec37501d0b3e"
   },
   "outputs": [],
   "source": [
    "distances = pdist(ingredient_nutrition_facts_normalized, metric='euclidean')\n",
    "mean_dist = np.median(distances)\n",
    "sigma_coeff = 0.8\n",
    "sigma = sigma_coeff*mean_dist\n",
    "adjacency = epsilon_similarity_graph(ingredient_nutrition_facts_normalized, sigma, epsilon=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 232
    },
    "colab_type": "code",
    "id": "rzWCWng_EbCv",
    "outputId": "5e0bd207-1929-4537-bb5b-6e7471968a69"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "normalized_recipe_adj_mat = Adj_mat_1/np.max(Adj_mat_1)\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[ingredients_matched_mask,:]\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[:,ingredients_matched_mask]\n",
    "combined_mat = normalized_recipe_adj_mat + adjacency\n",
    "plt.spy(combined_mat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "xDYrzEGbFgug"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_9.npy', combined_mat)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix 10: \n",
    "Same as 9 but divide the recipes matrix by its mean value instead of the max to increase its weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UNF2tUuc-HLj"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQIAAAD8CAYAAACcoKqNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO19X+hnx3Xf52SjKCEWRIojsUirehP2ZR1aWb9lCbiEtIRI1svKUJfNg1moyuZBLvZDHqQYWoW8JAE7L6kNcm2yFNeqwDZaSmkjhINfipVdR5K1lmWtKmGttWgb3GCnD2oknzz87rVnZ+fMnJk7f879/u4Hvnzv996Zc86cmTlz5sy5vx8xMzZs2HCw8TOjBdiwYcN4bIZgw4YNmyHYsGHDZgg2bNiAzRBs2LABmyHYsGEDDBgCIrqfiF4mostE9MgA/q8T0beI6DkiujDdu42IniaiV6bvWxvx/gIRXSOiF517Im8ienTS08tEdF9jOR4jou9PenmOiB5oLcdE+wgRfY2IXiKiS0T08en+CL1IsnTVDRH9PBE9S0TPT3L84XS/nk6YedgHwCEArwL4VQA/B+B5AMc7y/A6gPd69/4UwCPT9SMA/qQR798EcC+AF1O8ARyf9HMzgKOT3g41lOMxAL8fKNtMjon+YQD3Tte3APjuxHOEXiRZuuoGAAF4z3R9E4BvAPiNmjoZ7RGcBHCZmf83M/9/AE8AODVYJmBfhnPT9TkAD7ZgwsxfB/ADJe9TAJ5g5reZ+TUAl7Gvv1ZySGgmxyTLVWb+5nT9IwAvAbgTY/QiySKhiSy8j7+fft40fRgVdTLaENwJ4A3n9xXEFd0CDOAviegiEZ2d7t3BzFeB/cEA4PaO8ki8R+jqY0T0wrR1mN3ObnIQ0fsAfAD7K+BQvXiyAJ11Q0SHiOg5ANcAPM3MVXUy2hBQ4F7vnOcPMvO9AD4E4GEi+s3O/LXoravPAvg1APcAuArgUz3lIKL3APgygE8w8w9jRVvLE5Clu26Y+V1mvgfAXQBOEtGvx0TOlWO0IbgC4Ijz+y4Ab/YUgJnfnL6vAfgq9l2ot4joMABM39c6iiTx7qorZn5rGnw/BvA5/NS1bC4HEd2E/Yn3RWb+ynR7iF5CsozUDTP/HYC/AnA/KupktCH4awDHiOgoEf0cgNMAzvdiTkS/SES3zNcAfgfAi5MMZ6ZiZwA81UumCO/zAE4T0c1EdBTAMQDPthJiHmATPox9vTSXg4gIwOcBvMTMn3YeddeLJEtv3RDRrxDRL03XvwDgtwF8BzV1UivauyAi+gD2o7GvAvhkZ96/iv3o6vMALs38AfwygGcAvDJ939aI/5ew71r+A/at+EMx3gA+OenpZQAfaizHfwbwLQAvTAPrcGs5Jtr/HPtu7AsAnps+DwzSiyRLV90A+KcA/mbi9yKAf58ap7ly0FRpw4YNBxijtwYbNmwwgM0QbNiwYTMEGzZs2AzBhg0bsBmCDRs2oKEhoMy3Cp303qGwIgewySLBiixW5ACWy9LEEBDRIQD/Eftpu8cB/C4RHU9Us6JUK3IAmywSrMhiRQ5goSytPAKrbxVu2LAhhEYZc/8KwH9yfn8UwJ97Zc4CuDB9+O6772bsZ3HtzGdvb28VNHf501Jfd999N+/t7d3Aw73nf0vyderX/9M1s5CIPgLgPmb+t9PvjwI4ycz/TijPAPYFotCLU+VoQXMk75HtWSNa6mueOy59dz4R0U/4x+QI0WmEi8x8IvSg1dag6C2sWWHA9QrVYFSqdKm8pXyk3zl1e8Dz/nYGsf4uGbNW9NPKEBS/VegagxyMWiVnvpZX6RGyWdZHbbgrutTumD4s6KqJIWDmdwB8DMD/xP6fd3qSmS9J5ff29ty6111LK0tGvOK679B1jK7Qvii9EG+Jh6ZNQNhbktrpl1nKX6OTENyJkeqfnHsxHfjtCdXT9rGkQ61OpbEceubz1OpA0klKNz5MvH04xwhcuBNgrWCutz+dadWkuaEcqfFptL+6xwiqoHSbcBCw6UWHVnoKTe4SXjHPrSd+dih3BZZaVWMWeQh6e1f+oB6p/95tzpnQoye/C9MewYzNM7gRlo2bJdl6jBvX0EoBQ+vj16wh8JW5ttOE2ljSjlg0uxVik2LNCI1Bv43acWpJN2YNQQglxsC6Jc6FpcGzBtTWl+Z4MHRtvd9WZQgA2RgsCdQsQW9D0zu4tHZDKh2Rhp7l0Ms9Qkwdb8YMRe6xbQnMGoJYw3M8g16JHK07atSEXKIjC0YktMVcskrH6oa2QqnyGjl6bLHMGoIUfGOQyuNuKUeMvwX0WFEkfqONQSv+kqchJR9pE56ksq1h/vgwBoMJG9ehpVw5tEcECgEbR7et+KcWIs0Rakk6cius1iOYYfloMSZXqcwlg8SqfnqgtUegjVfFYgrScw2NWjBnCLQK8a2xVC92puvncrdyoWN55rHyuTy0v6V89hIZpACZ1FZtu0oTc0K8NbLlbmVCHkBsW+CPUU2A0KWbs3UoGcPmtgaaFU9ys0LKDd3rva+X+GvLu5AGUIxHzjOtjLFy7qDPbXtuOb+sf+3rq/YxX6iey9flr9G7Rp4l40eCOY9gCUqUtja4+2/LmCPd1uXMhSbol+v1WNDRThmCGaMjsC2xhva03GaNhm+IQ23MSTG2slDtpCHYxZVoTdjF1OIZ0rgqXXysjNOdNATA7hqDNeQtAHYGeKsUY/fbf69Cu0W11JcmDMH8F4o0EezQtXQKEKIRiijH7qdkcJ+Hfmv2jyG5pYh2qpz2FCQ3kp5L0x3cGp3HeKSe+TLkfkJtkMqlxmaoTKpe7F4IGj3GeIdg+i8U1bKUNWmN4jvTGtWWUoyWtyX/WJ/4z0brYcL6/kJRiwm0C1hDQpHkGYxAD/6+9xNqfyoFfvT4NGsIaitmV4xBaRt6tn1XdB2D5P6PNnylMGsIWqDFAB014HP5jtoajUar/q4R/HMNx2gDYtYQ5CpG6vBQZpsVdywHSwbKiEFm5Qgxlgpco/9TwUutbKPHollDUGvFC9GpGbzpNdiXDpTRA80KUsd8mvoSvfl3jmGIeRg9YdYQxKA5npkhZXkt3SZIRzdWoXnBRXs/p/4InZSs0LnlQv0ea6/mqHIkzBqC2MDNsZ6SRzB/5wyANW4ptIjpKYWQ+z0SNVb6g4ZFbx8S0esAfgTgXQDvMPMJIroNwH8F8D4ArwP418z8f5eJ2Q7abYL05pjGtauxDVlSPyVj6+y7kegpQ2phsaAPCTU8gn/BzPc4iQqPAHiGmY8BeGb6bRoH4bjroKJXv1pw75egxdbgFIBz0/U5AA824KFGjnvbqiNHrwSlEfw1D2wJrbZ4Gv1a1udSQ8AA/pKILhLR2eneHcx8FQCm79uLCFdSWu5xTs8VpKR8idtdulqV6sMNpFob/LVODfyXjlI5ASFjLJVfGrwtwdK/UPRBZn6TiG4H8DQRfUdbcTIcZ5MF9fSy7sfo5OzpS1f7ErmAsnhD7xyE0R6Qi5ayhOJG0neKhhSDkvjVxiKPgJnfnL6vAfgqgJMA3iKiwwAwfV8T6j7OzCeklyBqoedK2BJLzpvXkv24NtQ4NrVyElVsCIjoF4nolvkawO8AeBHAeQBnpmJnADy1VMgRsGgMgHLDtqE+Yit3rkc5uo+WbA3uAPDVqQE/C+C/MPP/IKK/BvAkET0E4HsAPlJCvJZilrrFVl4kCe0jLcgVgwX5auvJT0aLpQm74yckR6j8KJj9ewSWUDqYWkzW2MDacCNa6smn7RoHbT917sf1/T2CWtDmesdQIx25FnqeGsx1N8hw9eOeICyhIz1r2Rc7bwhqwWrMQIPSPIK57prRWn7pqFDLXyNfjz7YeUNQoyPcslaCdWs1Sr3RU0+hlTu10ud4a2aPD1uiV0JRLp8SY1Cj/KjA0pIthcWEotrwtwbudwra8j10adYQ9BjopYrNNQY12uImFJWg94SctyNr31qkMI8FrRewhE9LmDUEtZTZSoE5xkCzwmtpSO3x6Y08muqdYpzDY+m7BjE9x9qbehb7Ld2rCXOGQNPgpVHwkOJLBq7WGPirRo8VI3Vk5f9ODT7NXlfDW4uaOgpF9jX8QvdibfO9IPdo0Ocb01csN0G6l/M8BLP/DTml8KX0l9JxaeSeFacMSCoSnQvfm9AaDSlxRiPbUt3W9GA0ufvaNkh9LeUT5Mrk8wiNlZoB8BlmPYJaVi9nJStF7aPFGquA5ly6FXpvDUbCXf1DsYLYeJaeLfUISmDOEORGXGuVWwrLeQY524YNedB6dqUxpRbeVgjmtgYzegzW2jykbUJuZ2rKh7YbsfI9E4p6G5qe/EJ69ftCmvCxutqYQSuY8whmWF1dU2jlGfhu5Laq6zBiQclNUrMAs4agFkYYlJbbhKX5BBvqoWZ68GiDYGJrsLe3hwsXLtxwP+doLoaRk6ZlEPEgGgPLsRgfI+XMPW0w+xpy6khOi1p0SnkD9c7V/T3l6FUkBitbmJZ6itHW9lPnflzfa8i5ytEGaKS6Gq/C36e73xKNlitYLbotjvnmBJreC00qe29pYlcLD8/CYmxia9ASGoubu9fzkz009Jda/qWR5Bj/liuSlROEUDJVjSSt1D33mXRqID3vCbMewa6hR9JRrOxoF33DjbDgCczYGUMwaqDnHhWVdn4qTTjFdxRGD/YWuSIS3Xk7pM0bKenLVjBrCHIH0JIYwRLkyrnUMxg9sXJhYZCPhHZbOlpPZg1BDka63PN3TsCthjGoHeAL0bNkdJZ6UrWChTk8Q/xjZUdiJ4KFPd5Uq813SQAxt25pwHSpXmseIS6l0SpY6LexdCFZKs9SmDUEo12lHsiZ0GtMY7XQh61kSJ1OaGI6mvcTemEntgYxjFZwCrnbBAuTa8M+ao6t0f2684ZgDcgxBiX72xYJQ2tBr3a7cRttjCBVpyfMGoKDNnBTxmDJ8WFvWMqYG/mKsnvPcn8BCkNARF8gomtE9KJz7zYiepqIXpm+b3WePUpEl4noZSK6r5Xgu4hWKbnS2XYrWB/0NRF6sSf3jUO/Tuz9hVbQeAR/AeB+794jAJ5h5mMAnpl+g4iOAzgN4P1Tnc8Q0aFq0kbQM48g5NLVcr+1noF/vRS5KbNrQehItIXHIgX+UlsDXyYtj9pIGgJm/jqAH3i3TwE4N12fA/Cgc/8JZn6bmV8DcBnAyUqyFqHlCuta8pqrrsYzyD16TNGreXzY46y+FH6f5SLUrtqLwQiUxgjuYOarADB93z7dvxPAG065K9O95hjxQk1LSMagdBD33hrE0m17onWKccqwlD7rjdrBwlCLgiaSiM4S0QUiuvEvkmwAEDYGVldaHy2yH0vlaE1Ps5VLbQ1Go9QQvEVEhwFg+r423b8C4IhT7i4Ab4YIMPPjzHyChT+UUMtKWlDyEvhJRyWryIgJacUjqI1QBuHSIK8FHZUagvMAzkzXZwA85dw/TUQ3E9FRAMcAPJsitre395PrUABOyoPXfCQ6Mbq+HD5CZTTXpW2aB1pue11ZJfqlOs3Fkv7R7Ms1fZnTnpQOU+Wk3yG+sXGTozftGI82OELwSwCuAvgH7K/4DwH4ZeyfFrwyfd/mlP8kgFcBvAzgQ8qBxf6H9x8s/tSiM5q3i9K6vds9Uve9+9dvs0YHbr900tUFaQ6a+JuFJ06c4NAfLw0h5IZJL3y4ZbXum+uKu+V9920uE0Kofkp+l54vc0yWUPtC9WL3Um3XQKIltVOin9PmWN1YuSWIyRHTQWj8SPXc8qk2aPtwoif+zUIThoB2+I+XpoxFCa2ebVrCa6Tue0CzUKR0kKOjCvq0/cdLpRhB6Ld7L2TEYs/856GPVN79HcP83P3274VoSc9SXoWkmxgPTZ2fuIxOfEJqT85iEtJ3il5K3lBZ7ZgJyZIzPkoDoin5XVlC8rplpTGWkv26duy6RzASsVUh11vIpeGXA9Kpq6Hnpf2Q4tkLLceR28aYwdboXVNuLrsAtj2CljBi6FT3NM9cy5+zl0/xa5WQZUH/LeFO4pgel+ihx1HszhuC0StSK7j5BEm3b4AOLGXNtYbfRq3ht6Qbs4YgV0kjVp7Ufq4FpAFlceW1MNBj0fratEPPUjqwYjDNGoLcDmvhkuXw7JVFJ7XHojGwJg/QfsLlBlKt6MisIRhtIa1Bs7JZMwYW+rC2PkKR+tAzjRwxWr1h1hBsuB7aKLwlY2BFjhaoZeQsGEvAsCHIHUQHJUagMQi5xmCbsOX0/KO9nNMXTcJRL5g1BLnKGRkZn69bypAbVPKTjmKTvZXcFgZ4KyMnufqaRCH/nlbGlgbbrCHY5VWqBCXZfKO2CTkZbWuD354Wxi4WEG4Fc4Yg5PpKA0qz4rn5+TE6qYkWG9DaAR+jqVlNfDc0xTeWkZaTiurf1+QtaBJsYsE2TcAtRmO+dvWUs5UL6dx14yVX32+ztnyoTo5sOc9DMGcISqyepo7fIUuTcLSRYfeZJvEkRSuHhlsmtTUIDVjNxNOgxEhqkHNGH2tXivbc7porcs6JQy6tGdqxDhj+l2cuNAGYHGOgvY7JsIR3CX13ddMGmSQakjySDEsnQA6vGL9cOkB4Fc9pT2nbS/u7ZFFL6UBD05xHsCGOJZPT0tFiL4wKWK5Nz2YNwRqOD12+vfgv5dfLGFgJFrbmn4o7aGNLo/Vk1hDkYkSKscu398qTw88foK2NQWoL0hOt+YeOkNeyiLnYGUMgYfRArI3SLUFo39hqAO6azltBCmaOgFlDkKsYC1a1JXIiwDk0W+pt9OAOofd2RauD0ePXrCGohRrn/6X0c8qU8NXmEKSODrV01ohWbQvlneTmRmhzVnr0wc4bghhGrlitJ1/uEdnaJ/xILI0JaI77WvePWUOQ2/Ccs+ec56V8l/CIDYTSfaUmgarWYEtlKo6ElDylhd8PLt1cvpa2v2YNQS1YG4gzSg3QkmNDTZnUatWC75oQMgCaiR3LRE1tBWYeLXW584ZgVwdijkHovU1wB+5oQzziWDeUJSllRUonOr2xM4Zg9IDrhR7trDWBLfRJbRmkIKF7HQrU5ngEI5A0BET0BSK6RkQvOvceI6LvE9Fz0+cB59mjRHSZiF4movtaCa5FTlCu1Qs1ufVanTLkYDYG2nqpF19aInUqUhPSyq559yPk8pfUawGNR/AXAO4P3P8zZr5n+vx3ACCi4wBOA3j/VOczRHSolrAzQoopCRbmJNmEUklTR0cx5MgVkiV31U7tX0sDe1LwLMWzJjR7c2kFX8oz9C1tDUr10UOPSUPAzF8H8AMlvVMAnmDmt5n5NQCXAZxcIF8QNaPlWtqhVNLQpwe0K8mM1CSP7VVTBkc64bByYhCbqLX6K7RIhJ7F6uU8a4ElMYKPEdEL09bh1unenQDecMpcme7dACI6S0QXiEj3b5CNwsJgb40lMYPR+llLTMU1SjlGvhZKDcFnAfwagHsAXAXwqel+SPqghpj5cWY+wcL/YlsLRp1K5J4E1OCnHewjvKSYLL35hbYGUtmc+y1RZAiY+S1mfpeZfwzgc/ip+38FwBGn6F0A3lwm4oYQRqy0WmNgKaGo9anBfO2v6LHyoWej9VRkCIjosPPzwwDmE4XzAE4T0c1EdBTAMQDPFvIoqbazWKKPmoNMYwx22SMIxRc0PGLB1NE6AhR/qoyIvgTgtwC8l4iuAPgPAH6LiO7Bvtv/OoDfAwBmvkRETwL4NoB3ADzMzO+2EV0HC0pugd5bA5/eiH1sDbhyt2pD7jbK/R4FGu2SAAARjReiAVoMNPfocPTgsSBDCi1ljNEObRdSx5wddHlRisltmYUrRa8JmDo6XJvea+URLE20soadMQQlWHMabS++KYNj3Ri0MpiaJKYUrARUgQNuCKy7tSH03FPm7HMtDOZc9DQS1oOFO2MIYnu1EXxboeeEyw1IrsEYLD3JqJWpOtOyYASAHTIEEqwo+iBgDcZAm/4rQXNsuEZPahWGQPsi0PwdS95I3c/pmFiOeQlSg3TJS0Ga+zHdaQNsOYNbq/tWSUExWbWJUyFoFh//Za2S92ekPkk9D8GcIVjaMVKG19zpPp0SjyFnVckZxNILPO5zv32lxqGlp7R0gvnlcibW0jI162nq5tIO9Z30O8fImDAEe3t71/3O6VR/MsSsoa+Q2ETqEc0NWXKNTJo2awZgrMzS9qeMgdabWuIl+fqqYUS1fCVZYnKG6obqpXTnj4PVeAQXL17MrpPzWmnu6uLTXQJ/QsQ6JZWW67dZsvwa2aUyrrxLo9oxY+DvtTUv4IS8vZiMqVTnnLFQWteXxYXWW4jFJVKyacexuf+GnKP0kIJCytbUT/EKldEEjmLlNHU0E1rz3NWDxtCUyhOqJ/GODeql7dbWyxkbMfht1MRS3HqaSe/f1/anBiY8ghBqueW7cmqwVB8j9WApOl4TGqOmNWijx6lZQ1ALuzYARw+YUuyiMRgZX6qNnTEES45y1oScI6GRCAWqdtEYuEgF9qTyFgzHzhiCpckda8FaDFts39qzT2rzChni0P4+Rw4LfbozhuCgoDRvYJRBTJ1MrA1S0DN0ROmXk2hZwGYINlRH7Fx8xlqNQSpnI3drkCrTC+aOD2fUspaWrO4S5Kw0qbqtkXM+Hzve7SlLLr2Y7KGJbX0cmvUIallJC9a2BnJWGivQyNjaM7Cip5RxGG0ozBqCDXGUJLxYRUtjMHqCaZJ+RssIbIZgtciZOL2NgCZG4KOVMeh1aiAdA8b0INUbYbTNxgg2hFGSiWY1RhCqVztm0DpGIPHwJ3PO6cEID8GsR2DBXbIENzfd/dbAQsKKFrU9gx4egftM0rUmsWgkzBqCXMWMUmQvvrXeCOyFJRlzNY1BK4+gJi8LfWnWEOSi1ltktfi2QunEGrE9WMK3ljHoHYQcNQ6XYmcMwQZ7qPHG5JrfunRhPbcgaQiI6AgRfY2IXiKiS0T08en+bUT0NBG9Mn3f6tR5lIguE9HLRHRfywakMHrv1QqW26XJvddiqTEYoScprVqCib70jzACRxqHAdw7Xd8C4LsAjgP4UwCPTPcfAfAn0/VxAM8DuBnAUQCvAjiU4MGtPrzPYMgnxnukXGv8WNSXJNOMHLk7te+CNAeTHgEzX2Xmb07XPwLwEoA7AZwCcG4qdg7Ag9P1KQBPMPPbzPwagMv46b9NV8OElWyInglBo6LSNXmWegajPYJUP1sZ51kxAiJ6H4APAPgGgDuY+SqwbywA3D4VuxPAG061K9O9Awcr5+GjstdaROxzJ46lfXgMo+VUGwIieg+ALwP4BDP/MFY0cO+G3iOis0R0gYguCPy0oplFy0w5KytJb+Qag1Z6SuUF+Nep3ILR/akyBER0E/aNwBeZ+SvT7beI6PD0/DCAa9P9KwCOONXvAvCmT5OZH2fmEyz8m2ar58ijMHqglKKF3BZeYdbmE8TyPiyNTc2pAQH4PICXmPnTzqPzAM5M12cAPOXcP01ENxPRUQDHADybK5glJVnAkjTUkZOmVT9aMAYScuSyMs417xp8EMBHAXyLiJ6b7v0BgD8G8CQRPQTgewA+AgDMfImIngTwbQDvAHiYmd/NFYwr5ZzXojMaS9sxQg+tefb4ewYSfLfe9wLcLYH5o0MAZEEQIrpBiNwOHjXQfbgy1JRpptWznWsxoiONgcR3RH8pcFHaiu9MZuGoqLj/2SUs9UB6LTKxbULvha6U3+gF2awh2LVJVQtr0UtvwygZg9ZbE+le7stho/vVrCHIhZUVoZUMB+U15CUIGYOWx4euXoWM2eTxoV8+JnPLPjzQf5ik1/6tRWJRDs3Rq01P+Pvy1qcWIT6h+znHhyPeYNwZj6BESWueIAdhhS9Fz6PF1OqdkiNUbkTfmjUEucrY9YnhD5g15BGMzl8YtSVausCMWKDMbA1iltW18K5L5j/zabj7aunapenX82lqVxqpvi9TTEYNv5DMobo+D39LFJIxpGO/bolRCsmd4iX1d4yGzy/2zKcjIdRXpQjlHvj3U+32ZZbappHVjCFINSKkNPc71jmhurHrmFzSIJNQQlcaIBrascEl7Vdzot8anaUg8c/9jtGYryUDn5IlJb/WQLv3JNncsZ4a55I8OfdDMLs1yMWo/b6GrzQwlmD0HngtmCdX7XbEPAtNee0zl3bLvjBrCNYcyPOREzFuhdJBVCqnlYw67TYrl96MXIOgoSnxOJCnBiWu11LarQKUufLWGEwufJfTrxcLqmmi4v65eeh6FGJbhRnSeX6MXm60P1cXIboH0iPYEEfu3ra1a6yVYzRinkGO7LntbF1+KcwEC3Ow1NWuNYi15XPlrT0INJHwms9jAU4L8IPLSwKgIRq5hiYV1DzQW4MYLG0NNHV6bA2WoMbWQCpv1UPwo/a57re0NSgZWyme29ZAgCWPQHvclMOr9CQiRi/Fr/RoKnTkVTKxWiHlDaWOViV6fvnSnAoNDrRHUKvRVlekXKypHfPE0hrKnvCNVOlpguT51Ig/aHnWhFlDkIvakfaeNDaMRY4xSCU1SUlBKVqlnmUt7IwhKFFSDcVaTmTaUJ6FWAtLjph7YmcMQa0A4Fqwq+2qjVigNuTeW/YSt62BAms9197QF6nAYA1jEKJbYxIfyK1BLeu3KyvnknaMOo60kFlYgpgx0GQPLjmCHIVVJhRZgpWc+hh6y2ddHxr4SUfu/VgdIO9NUSvjx6xHYEE5GvSScwmfta7MoxHyDEpX8VSsQuLTC2YNwTZwr8cSfYwwqrtifKT0X21CWCoekXrWC2YNwVo8gl5Ymz4sJhOVQhNAlLIpS1KLRxhRs4agFnZlMC6BlX3oCNTMUE2lK9dKJR9hRDX/BPUIEX2NiF4ioktE9PHp/mNE9H0iem76PODUeZSILhPRy0R0X45Aschrzv5MKutb7lSEW2PRU9Hikgi6L7+U4y7xcn9rB5VmJavxslSsb3Lo5ELqX+1Yi8UMXCx5V6RmhqzWM/lJgdgHwGEA907XtwD4LoDjAB4D8PuB8scBPA/gZgBHAbwK4FCCB/sf3n+w6k/NNuyCPtbeBz5NF6H7/j2Jllu+8eeCNAeTHgEzX2Xmb07XPwLwEoA7I1VOAXiCmd9m5tcAXAZwMqj/w/kAAArSSURBVMXH4ZdcGaTVMkQjVM8vF1uxa61OIfra1TVHBr+epq4km1aHGllC9yVaKbljuvF/+6t46Dq3Df5bi76XJt0L0Y29nFU6JkvGbFaMgIjeB+ADAL4x3foYEb1ARF8golune3cCeMOpdgVxw4G9vb1oo0OTNfTt05DqxQZaaHBIVjT28euGaGnaqJUjpDOt/DmyhcrF5I5dS3VTcsfopmTWyBFr3/zxX7euAa0eYvKm6klQJxQR0XsAfBnAJ5j5h0T0WQB/hH2X448AfArAvwEQ2iDdIAURnQVwNsFTJdvcyNBRz9xpfnkp+CNZcpdOqH5JAoort38vdN/lmeIXou3TivGM1U9Bm+4dkr8GYn2b0l8KtSZ+aFyFjh1D406Sa0nfqTwCIroJ+0bgi8z8lYnZW8z8LjP/GMDn8FP3/wqAI071uwC8GRD8cWY+wc7/a49FS2MdkIqy1ug8v5NCq3ANXn5bUu2Wnsd0ojVWoUy5GrqUaJTQ9uv47nnIfdfkAkjwx4HmaLEWNEZAKpeSUXNqQAA+D+AlZv60c/+wU+zDAF6crs8DOE1ENxPRUQDHADyr4BO8BtJ7SvfanwChCREaLPPvkk5NdYKmnu/ChWTyn9VaUUMrSGj1LOUXmqy1V1aJl3tfcp81kDy1+Xdpe1L61ExqjQ5SfDRbgw8C+CiAbxHRc9O9PwDwu0R0D/bd/tcB/N4kxCUiehLAtwG8A+BhZn5XwQdT/eCKGHKTpfoSvRRtaRsRKpPrWmrcNamNObxC2yStSx7T1RKDo+27nDKlMoSMQGz7qJWvZDxo+zNnu+k/1/KhXm5NVAgiBm4cxLEJLCk+tjf06cXcap+eZuKE6pY+T20J/Pp+nRxDEFuxY3RKENujh+IgOXRTkzGkn9g4KvV+ahuxFH1Jj+719H3R3Yq7MJFZOJ8azNDsIf3rkOsWW91TRkCaVKFtylJXU/ss5SbGJrdPJ+QmaxCqo+0XaTWW5HDvS3XdNsX4SjKn2uLLGarjQrtN0IzxFK9QuZSeJJh+DVnTqFiZ1KSVeKZiFCl6see5v2Ny+DKnBlfJINUavNSz3C1U6J7WGLiGTovStoWQu03waUuLUEgGrfFPwdTWwEWuIlu7ZCm+Nd3LGJ8N64HBPhO3BmY8gtTKOsPf26XKu/VKyqasc6xsqg1+uVC8QLsaaWMe2jaFZPDLhn5LsqbapvFWYrpyaWg9H592TH8+T1UALuIZaO+Hxon7WyqTGqs+zBiCHMsZigHkWl9tWW3Edul997d0DcTbWbNNWnmWtFtLq4RGyUocqxPjmaIZ6rOScSHVq6EDE8FC4MagUWi1j8UAXGsoBXliPKVymj17qlxIjtBvqa3u75Rno217rE6sbkjOVGxC4hfrT2kcxOhqysf6IHQ/l38ImvhNDmJ97j7PGQer8QhCFtkPDhHJyUMa2pIbr0EqQJRj6UPXWre/dOXy60iDdyn9UL1Qf8Z4aFdu6Z40VkLlc/nH5MrxDFK0NL9zPAMzhmApWii0dX0rWLrVyKU7EiNlSi0WI2Fma+DDorJ2GSUu6oZ8lAQye8CsIchVVs09WA2+rfnV0s+MmoY3FXfpidr8Y/tubYxjNgZWdAQYNASSglPBEClYGKLv080JFkpHMhqeMVpSkE8aVO5g8mmG6vonLDHeKdlS7Zv339LxmDSZNP0hySA9C026nH4Ptc1to8/Lf6aJfWmDiVrDIfVhDCYMgZ9i7EI7IEODPlQ3FXxLTQT3d6hjYquERD/WXqlMzDBq3c+Y7CFDkzIGLl+t8ZCeh2RMGaxQPY0e/PKxPosZ0RR9H66+JMPu1pWMeWr8aeQzlVnoK0Wypu6gix2npayx1DkhXqHfIcMSKivJJNUP6SGmj1CbfDqhclI7YvwknUtlNW1OtVUzFjTyhfi5v6U62vbmQNK/pl5MXonedG9dmYWp1SI2+UP3pckl1ZEGo0a+UEf5K2XOIA99SxMpVDfEJ7XKh+6lDJxmUkmyxYyI/1vydiQekgeY8lI095fAHWMpI+iWz4XWyJgxBNJAjjVCsw+TJmYu7RZHaSGZpLIp7yJ2L5dXDDl1c+Rzny2hG6KhbWOuvDmeisQvZAwk/hrdSdcpWUzECDbkI3d1GLEFzNmjt5ShN72URyvVKV31a8CsIcjZDy+lk0vTLxtypXORmjSxlSJFN1VeM3BL2javdFrUmATSNmq+V2qcYp6ljxKvyf0eYQzMGgINYvvOGjS1ZWtsH1KTRjOhJbo5ZbQBwhaowSPWFyX9FBtXrSbsCGNg1hDkKqLHQLWEngNliW6tbQ1KDak2QF0LvY2BWUOQi1EDrsa2oITfGmBB1tYLhLvVCJ0MpTyK2PjpaQxWaQhGBb7ca2kA9ELNQSLtnWvQH+2phfqoRn+5nkLpPl+7besxrlZpCEagNFi3BmiCXjXp9kRoktY6Es7hn3qeOhpsbQzMGgLtmSlQdyWORc1LjoVq8Y/JsoTWDAuTtgYkrybmESzp16UnRVo6rY2BmYQiHzmR6pqDuMaRUAv+fhltwkpt3lqUnnIsRar/Usk2S3UWMkTao1sNr1YnOGY9glz0TA8N0bcQGJPQO7FHGyxbMyTPQzoGrnm83UKnZg1BrdWs9YrUa7+5JEaRm9izFKH9+K5B2x/avIbc/qxtDDT/BPXniehZInqeiC4R0R9O928joqeJ6JXp+1anzqNEdJmIXiai+0oEqxkR3wUsieiP8AiWZPFtSKO2MdB4BG8D+JfM/M8A3APgfiL6DQCPAHiGmY8BeGb6DSI6DuA0gPcDuB/AZ4joUDWJDzhKPJARHkHPyHxKFst8l2aj1jIGSUPA+/j76edN04cBnAJwbrp/DsCD0/UpAE8w89vM/BqAywBOVpF2Q3FMordHMIJvCLX5h+gtyTFZKl8tY6CKERDRIdr/l+jXADzNzN8AcAczXwWA6fv2qfidAN5wql+Z7mWhliUfvSJZwRYjqIPQKcHSU4elW6gaxkBlCJj5XWa+B8BdAE4S0a/H5AqRuKEQ0VkiukBEFwSeGtGSGL0itcKuTbBW6B3E9e9pE4ZqJG8tGetZpwbM/HcA/gr7e/+3iOjwJMRh7HsLwL4HcMSpdheANwO0HmfmEyz86aRtoIex68dyu4DcQG4tLDEGmlODXyGiX5qufwHAbwP4DoDzAM5Mxc4AeGq6Pg/gNBHdTERHARwD8GyuYJb3wCP5liTqjDg1cL93EVIMIJXZKD0bnWegySw8DODcFPn/GQBPMvN/I6L/BeBJInoIwPcAfAQAmPkSET0J4NsA3gHwMDO/myvY2vIIWvNeEoUfld23y16dlEegvS/dqyVbbgaiqb9i7KJVKmVP1GzDTMv/7i1HDr9RacYhWVrQk2i77Y71V2v9BHiKf8XYbGbhhnoY5RGM4N0asfbEXPyUR9BiQc7ZJuy8IbDg8VjAqDyC0RiVSOVD83ZhC2iNgVlDsOURXI+l++5ReQS7Bk1CUS6t1sFcjTGw8hry3wL4f9P3aLwXleSoMBFEWQZMsmp6qYBhsnjbnvcC+NvYVqhjsDCqk4nXPxGfW3HjiOiCFMg4iHIAmywSrMhiRQ5guSxmtwYbNmzoh80QbNiwwZQheHy0ABOsyAFsskiwIosVOYCFspiJEWzYsGEcLHkEGzZsGITNEGzYsGEzBBs2bNgMwYYNG7AZgg0bNgD4R1N7K3rEoZfQAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "normalized_recipe_adj_mat = Adj_mat_1/np.mean(Adj_mat_1)\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[ingredients_matched_mask,:]\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[:,ingredients_matched_mask]\n",
    "combined_mat = normalized_recipe_adj_mat + adjacency\n",
    "plt.spy(combined_mat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ff5tWss2-JPd"
   },
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_10.npy', combined_mat)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix 11\n",
    "Combine the two matrices by multiplyting them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "normalized_recipe_adj_mat = Adj_mat_1/np.max(Adj_mat_1)\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[ingredients_matched_mask,:]\n",
    "normalized_recipe_adj_mat = normalized_recipe_adj_mat[:,ingredients_matched_mask]\n",
    "combined_mat = np.multiply(normalized_recipe_adj_mat, adjacency)\n",
    "plt.spy(combined_mat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_11.npy', combined_mat)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adjacency Matrix No. 12\n",
    "Add the nutrition and recipe adjacency matrices made using RBF kernels (#3 and #5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mat_3 = np.load('../data/Adjacency_Matrix_3.npy')\n",
    "mat_3 = mat_3[ingredients_matched_mask,:]\n",
    "mat_3 = mat_3[:,ingredients_matched_mask]\n",
    "mat_5 = np.load('../data/Adjacency_Matrix_5.npy')\n",
    "combined_mat = mat_3 + mat_5\n",
    "plt.spy(combined_mat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('../data/Adjacency_Matrix_12.npy', combined_mat)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "CreateAdjacencyMatrix.ipynb",
   "provenance": []
  },
  "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}