src/python/Criar Classificador de Igrejas Usando Snorkel-Copy1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Criar Classificador de Igrejas usando SnorkelML\n",
    "\n",
    "Olá!\n",
    "\n",
    "Neste notebook usaremos a fantástica ferramenta [SnorkelML](snorkel.org/get-started/) para criar um classificador para as igrejas evangélicas em língua portuguesa.\n",
    "\n",
    "Este trabalho faz parte do projeto dadascope. \n",
    "\n",
    "## Sumário\n",
    "\n",
    "1. Abrir banco de dados e selecionar razões sociais únicas.\n",
    "2. Criar funções de classificação (Labelling Functions) do teste.\n",
    "3. Criar modelos e aplicar nos bancos de dados de desenvolvimento e teste.\n",
    "4. Avaliar treinamento\n",
    "5. Criar banco de dados com classificação"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/henriquegomide/github_repos/other/igrejasevangelicas/src/python'"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import re\n",
    "\n",
    "from snorkel.labeling import labeling_function, PandasLFApplier, LFAnalysis\n",
    "from snorkel.labeling.model import MajorityLabelVoter, LabelModel\n",
    "from snorkel.analysis import get_label_buckets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Abrir banco de dados\n",
    "data = pd.read_csv('../../data/igreja_cnpj_ativo.zip',\n",
    "                   encoding='ISO-8859-1',\n",
    "                   compression='zip')\n",
    "\n",
    "# Dividir banco entre treino, dev, validação e teste\n",
    "df_copy  = data['razao_social'].unique()\n",
    "df_copy  = pd.Series(df_copy)\n",
    "df_copy  = pd.DataFrame({'index':df_copy.index, 'razao_social':df_copy.values})\n",
    "\n",
    "df_train = df_copy.sample(frac=0.95, random_state=0)\n",
    "\n",
    "# Após a realização da codificação manual\n",
    "df_dev   = df_train.sample(n=300)\n",
    "\n",
    "\n",
    "# Afer labeling is done, open data\n",
    "df_copy = pd.read_csv('../../data/human_labelling/df_valid.csv', \n",
    "                      sep=';')\n",
    "df_valid = df_copy.sample(frac=0.5, random_state=0)\n",
    "df_test = df_copy.drop(df_valid.index)\n",
    "df_dev = pd.read_csv('../../data/human_labelling/df_dev.csv',\n",
    "                    sep=';')\n",
    "\n",
    "# Create outcomes\n",
    "Y_dev   = df_dev.label.values\n",
    "Y_valid = df_valid.label.values\n",
    "Y_test  = df_test.label.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Criar labelling functions\n",
    "\n",
    "De acordo com a documentação do Snorkel, iremos tentar criar funções 'grosseiras' que tentarão classificar as igrejas usando palavras-chaves e expressões regulares."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Definir rótulos para conveniência\n",
    "ABSTAIN = -1\n",
    "NOT_TARGET = 0\n",
    "TARGET = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def keyword_lookup(x, keywords, label):\n",
    "    if any(word in x.razao_social.lower() for word in keywords):\n",
    "        return label\n",
    "    return ABSTAIN\n",
    "\n",
    "def make_keyword_lf(keywords, label=TARGET):\n",
    "    return LabelingFunction(\n",
    "        name=f\"keyword_{keywords[0]}\",\n",
    "        f=keyword_lookup,\n",
    "        resources=dict(keywords=keywords, label=label),\n",
    "    )\n",
    "\n",
    "\n",
    "# Definir objetivos\n",
    "keyword_assembleia = make_keyword_lf(keywords=[\"assembleia\", \"ass\", \"assem\",\n",
    "                                               \"assemb\", \"asembleia\"])\n",
    "\n",
    "keyword_evangelica = make_keyword_lf(keywords=[\"evangelica\", \"evang\"])\n",
    "\n",
    "keyword_pentecostal = make_keyword_lf(keywords=[\"pentecostal\"])\n",
    "\n",
    "keyword_iurd = make_keyword_lf(keywords=[\"universal do reino de deus\"])\n",
    "\n",
    "keyword_internacional = make_keyword_lf(keywords=[\"internacional\"])\n",
    "\n",
    "keyword_ministerio = make_keyword_lf(keywords=[\"ministerio\", \"minist\", \"min\"])\n",
    "\n",
    "keyword_igreja = make_keyword_lf(keywords=[\"igreja\"])\n",
    "\n",
    "keyword_sara = make_keyword_lf(keywords=[\"sara\"])\n",
    "\n",
    "keyword_protestantes = make_keyword_lf(keywords=[\"luterana\", \"batista\", \"metodista\", \n",
    "                                                \"presbiteriana\"])\n",
    "\n",
    "\n",
    "# Definir nomes que não são critérios\n",
    "keyword_funerario = make_keyword_lf(keywords=[\"funerarios\", \"servico\", \"servicos\",\n",
    "                                             \"funeraria\", \"funeral\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_associacao = make_keyword_lf(keywords=[\"associacao\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "\n",
    "keyword_espiritas = make_keyword_lf(keywords=[\"espirita\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_catolica = make_keyword_lf(keywords=[\"catolica\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_irmas = make_keyword_lf(keywords=[\"irmas\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_moradores = make_keyword_lf(keywords=[\"moradores\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_senhora = make_keyword_lf(keywords=[\"senhora\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_paroquia = make_keyword_lf(keywords=[\"paroquia\"],\n",
    "                                      label=NOT_TARGET)\n",
    "\n",
    "keyword_sao = make_keyword_lf(keywords=[\"sao\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_santo = make_keyword_lf(keywords=[\"santo\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_maconica = make_keyword_lf(keywords=[\"maconica\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_educacional = make_keyword_lf(keywords=[\"educacional\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_comunitario = make_keyword_lf(keywords=[\"comunitario\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_treinamento = make_keyword_lf(keywords=[\"treinamento\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_instituto = make_keyword_lf(keywords=[\"instituto\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "\n",
    "keyword_kardec = make_keyword_lf(keywords=[\"kardec\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_umbanda = make_keyword_lf(keywords=[\"umbanda\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_caboclo = make_keyword_lf(keywords=[\"caboclo\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_tenda = make_keyword_lf(keywords=[\"tenda\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_ogum = make_keyword_lf(keywords=[\"ogum\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_ubirajara = make_keyword_lf(keywords=[\"ubirajara\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_oxala = make_keyword_lf(keywords=[\"oxala\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_cacique = make_keyword_lf(keywords=[\"cacique\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_iemanja = make_keyword_lf(keywords=[\"yemanja\", \"iemanja\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_oxossi = make_keyword_lf(keywords=[\"oxossi\"],\n",
    "                              label=NOT_TARGET)\n",
    "\n",
    "keyword_coral = make_keyword_lf(keywords=[\"coral\"],\n",
    "                              label=NOT_TARGET)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criar lista de funções de rótulo\n",
    "lfs = [\n",
    "    keyword_assembleia,\n",
    "    keyword_evangelica,\n",
    "    keyword_funerario\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/henriquegomide/environments/datascience/lib/python3.8/site-packages/tqdm/std.py:668: FutureWarning: The Panel class is removed from pandas. Accessing it from the top-level namespace will also be removed in the next version\n",
      "  from pandas import Panel\n",
      "100%|██████████| 73688/73688 [00:42<00:00, 1729.07it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 1529.52it/s]\n",
      "100%|██████████| 219/219 [00:00<00:00, 1859.85it/s]\n"
     ]
    }
   ],
   "source": [
    "# Aplicar funções nos bancos de dados \n",
    "applier = PandasLFApplier(lfs=lfs)\n",
    "L_train = applier.apply(df=df_train)\n",
    "L_dev   = applier.apply(df=df_dev)\n",
    "L_valid = applier.apply(df=df_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>j</th>\n",
       "      <th>Polarity</th>\n",
       "      <th>Coverage</th>\n",
       "      <th>Overlaps</th>\n",
       "      <th>Conflicts</th>\n",
       "      <th>Correct</th>\n",
       "      <th>Incorrect</th>\n",
       "      <th>Emp. Acc.</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>keyword_assembleia</th>\n",
       "      <td>0</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.05</td>\n",
       "      <td>17</td>\n",
       "      <td>3</td>\n",
       "      <td>0.850000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_evangelica</th>\n",
       "      <td>1</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.05</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_pentecostal</th>\n",
       "      <td>2</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.00</td>\n",
       "      <td>19</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_universal do reino de deus</th>\n",
       "      <td>3</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_internacional</th>\n",
       "      <td>4</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_ministerio</th>\n",
       "      <td>5</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.01</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_igreja</th>\n",
       "      <td>6</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.61</td>\n",
       "      <td>0.54</td>\n",
       "      <td>0.07</td>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_sara</th>\n",
       "      <td>7</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_luterana</th>\n",
       "      <td>8</td>\n",
       "      <td>[1]</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.06</td>\n",
       "      <td>21</td>\n",
       "      <td>1</td>\n",
       "      <td>0.954545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_espirita</th>\n",
       "      <td>9</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_funerarios</th>\n",
       "      <td>10</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_associacao</th>\n",
       "      <td>11</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.04</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.750000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_catolica</th>\n",
       "      <td>12</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_senhora</th>\n",
       "      <td>13</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_paroquia</th>\n",
       "      <td>14</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_sao</th>\n",
       "      <td>15</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.07</td>\n",
       "      <td>0.07</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_santo</th>\n",
       "      <td>16</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.02</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_maconica</th>\n",
       "      <td>17</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_educacional</th>\n",
       "      <td>18</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_comunitario</th>\n",
       "      <td>19</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_treinamento</th>\n",
       "      <td>20</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_instituto</th>\n",
       "      <td>21</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_kardec</th>\n",
       "      <td>22</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_umbanda</th>\n",
       "      <td>23</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_caboclo</th>\n",
       "      <td>24</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_tenda</th>\n",
       "      <td>25</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_ogum</th>\n",
       "      <td>26</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_ubirajara</th>\n",
       "      <td>27</td>\n",
       "      <td>[0]</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_oxala</th>\n",
       "      <td>28</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_cacique</th>\n",
       "      <td>29</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_yemanja</th>\n",
       "      <td>30</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_oxossi</th>\n",
       "      <td>31</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>keyword_coral</th>\n",
       "      <td>32</td>\n",
       "      <td>[]</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     j Polarity  Coverage  Overlaps  \\\n",
       "keyword_assembleia                   0      [1]      0.20      0.18   \n",
       "keyword_evangelica                   1      [1]      0.31      0.28   \n",
       "keyword_pentecostal                  2      [1]      0.19      0.18   \n",
       "keyword_universal do reino de deus   3       []      0.00      0.00   \n",
       "keyword_internacional                4       []      0.00      0.00   \n",
       "keyword_ministerio                   5      [1]      0.16      0.13   \n",
       "keyword_igreja                       6      [1]      0.61      0.54   \n",
       "keyword_sara                         7      [1]      0.03      0.03   \n",
       "keyword_luterana                     8      [1]      0.22      0.22   \n",
       "keyword_espirita                     9      [0]      0.06      0.01   \n",
       "keyword_funerarios                  10      [0]      0.01      0.01   \n",
       "keyword_associacao                  11      [0]      0.04      0.04   \n",
       "keyword_catolica                    12      [0]      0.01      0.01   \n",
       "keyword_senhora                     13       []      0.00      0.00   \n",
       "keyword_paroquia                    14      [0]      0.01      0.01   \n",
       "keyword_sao                         15      [0]      0.08      0.07   \n",
       "keyword_santo                       16      [0]      0.03      0.03   \n",
       "keyword_maconica                    17      [0]      0.01      0.00   \n",
       "keyword_educacional                 18       []      0.00      0.00   \n",
       "keyword_comunitario                 19       []      0.00      0.00   \n",
       "keyword_treinamento                 20       []      0.00      0.00   \n",
       "keyword_instituto                   21       []      0.00      0.00   \n",
       "keyword_kardec                      22       []      0.00      0.00   \n",
       "keyword_umbanda                     23      [0]      0.01      0.01   \n",
       "keyword_caboclo                     24      [0]      0.01      0.01   \n",
       "keyword_tenda                       25      [0]      0.01      0.01   \n",
       "keyword_ogum                        26      [0]      0.01      0.00   \n",
       "keyword_ubirajara                   27      [0]      0.01      0.01   \n",
       "keyword_oxala                       28       []      0.00      0.00   \n",
       "keyword_cacique                     29       []      0.00      0.00   \n",
       "keyword_yemanja                     30       []      0.00      0.00   \n",
       "keyword_oxossi                      31       []      0.00      0.00   \n",
       "keyword_coral                       32       []      0.00      0.00   \n",
       "\n",
       "                                    Conflicts  Correct  Incorrect  Emp. Acc.  \n",
       "keyword_assembleia                       0.05       17          3   0.850000  \n",
       "keyword_evangelica                       0.05       31          0   1.000000  \n",
       "keyword_pentecostal                      0.00       19          0   1.000000  \n",
       "keyword_universal do reino de deus       0.00        0          0   0.000000  \n",
       "keyword_internacional                    0.00        0          0   0.000000  \n",
       "keyword_ministerio                       0.01       16          0   1.000000  \n",
       "keyword_igreja                           0.07       61          0   1.000000  \n",
       "keyword_sara                             0.00        3          0   1.000000  \n",
       "keyword_luterana                         0.06       21          1   0.954545  \n",
       "keyword_espirita                         0.01        6          0   1.000000  \n",
       "keyword_funerarios                       0.01        0          1   0.000000  \n",
       "keyword_associacao                       0.04        3          1   0.750000  \n",
       "keyword_catolica                         0.01        1          0   1.000000  \n",
       "keyword_senhora                          0.00        0          0   0.000000  \n",
       "keyword_paroquia                         0.00        1          0   1.000000  \n",
       "keyword_sao                              0.07        2          6   0.250000  \n",
       "keyword_santo                            0.02        1          2   0.333333  \n",
       "keyword_maconica                         0.00        1          0   1.000000  \n",
       "keyword_educacional                      0.00        0          0   0.000000  \n",
       "keyword_comunitario                      0.00        0          0   0.000000  \n",
       "keyword_treinamento                      0.00        0          0   0.000000  \n",
       "keyword_instituto                        0.00        0          0   0.000000  \n",
       "keyword_kardec                           0.00        0          0   0.000000  \n",
       "keyword_umbanda                          0.00        1          0   1.000000  \n",
       "keyword_caboclo                          0.00        1          0   1.000000  \n",
       "keyword_tenda                            0.00        1          0   1.000000  \n",
       "keyword_ogum                             0.00        1          0   1.000000  \n",
       "keyword_ubirajara                        0.00        1          0   1.000000  \n",
       "keyword_oxala                            0.00        0          0   0.000000  \n",
       "keyword_cacique                          0.00        0          0   0.000000  \n",
       "keyword_yemanja                          0.00        0          0   0.000000  \n",
       "keyword_oxossi                           0.00        0          0   0.000000  \n",
       "keyword_coral                            0.00        0          0   0.000000  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Summary labelling functions output\n",
    "LFAnalysis(L=L_dev, lfs=lfs).lf_summary(Y=Y_dev)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Criar modelos para classificar igrejas\n",
    "\n",
    "Hora de criar dois modelos e comparar a acurárcia de nossas classificações."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Majority Model\n",
    "majority_model = MajorityLabelVoter()\n",
    "preds_train = majority_model.predict(L=L_train)\n",
    "\n",
    "# Using LabelModel from Snorkel\n",
    "label_model = LabelModel(cardinality=2, verbose=True)\n",
    "label_model.fit(L_train=L_train, n_epochs=1000, lr=0.001, log_freq=100, seed=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:root:Metrics calculated over data points with non-abstain labels only\n",
      "WARNING:root:Metrics calculated over data points with non-abstain labels only\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Majority Vote Accuracy:   95.8%\n",
      "Label Model Accuracy:     94.9%\n"
     ]
    }
   ],
   "source": [
    "# Compare Model Metrics\n",
    "majority_acc = majority_model.score(L=L_valid, Y=Y_valid)[\"accuracy\"]\n",
    "print(f\"{'Majority Vote Accuracy:':<25} {majority_acc * 100:.1f}%\")\n",
    "\n",
    "label_model_acc = label_model.score(L=L_valid, Y=Y_valid)[\"accuracy\"]\n",
    "print(f\"{'Label Model Accuracy:':<25} {label_model_acc * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Aplicar modelo e salvar banco de dados\n",
    "\n",
    "Hora de aplicar o modelo com melhor acurácia e salvar o banco para posterior análise."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/henriquegomide/environments/datascience/lib/python3.8/site-packages/tqdm/std.py:668: FutureWarning: The Panel class is removed from pandas. Accessing it from the top-level namespace will also be removed in the next version\n",
      "  from pandas import Panel\n",
      "100%|██████████| 152269/152269 [01:15<00:00, 2005.06it/s]\n"
     ]
    }
   ],
   "source": [
    "L_train = applier.apply(df=data)\n",
    "data['is_evangelic'] = label_model.predict(L=L_train, tie_break_policy=\"abstain\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       " 1    106941\n",
       "-1     23966\n",
       " 0     21362\n",
       "Name: is_evangelic, dtype: int64"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['is_evangelic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "#data.to_csv('data/final/cnae_labelled_data.csv.gz', \n",
    "#           compression='gzip')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.2 64-bit ('datascience': venv)",
   "language": "python",
   "name": "python38264bitdatasciencevenv571850cfa66243cf9c4a7feda8aedcdb"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}