ad

08-Test2.ipynb

@@ -793,7 +793,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 185,
+   "execution_count": 190,
    "metadata": {},
    "outputs": [
     {
@@ -824,6 +824,7 @@
     "    s = s.replace(\"<START:location>\", \"\")\n",
     "    s = s.replace(\"<END>\", \"\")\n",
     "    s = re.sub(r\"[^\\w\\s]\", '', s)\n",
+    "    s = re.sub(r\"[^\\w\\s]\", '', s)\n",
     "    #s = re.sub(r\"[<START:location>|<END>]\", '', s)\n",
     "    return s\n",
     "f = open(\"./data/named_entity_recognition_sp_MX_locations.JSON\", \"r\", encoding=\"utf8\")\n",
@@ -833,6 +834,7 @@
     "jsonDoc = json.loads(jsonDoc)\n",
     "for i in jsonDoc:\n",
     "    texto = removerpunt(i[\"doc_locations\"])\n",
+    "    ##texto = re.findall(r'([A-Za-z][a-z]{2,15})', texto)\n",
     "    texto = texto.split(\" \")\n",
     "    listaPalabras.extend(list(set(texto)))\n",
     "\n",