DataLocale Restaurants — OpenRefine + ElasticSearch Export

refine-restaurants-cleanup.json

[
  {
    "op": "core/column-removal",
    "description": "Remove column Adresse suite",
    "columnName": "Adresse suite"
  },
  {
    "op": "core/text-transform",
    "description": "Text transform on cells in column Site web using expression grel:if(isNonBlank(value), 'http://'+replace(value, /^https?:\\/\\//, ''), null)",
    "engineConfig": {
      "facets": [
        {
          "invert": false,
          "expression": "value",
          "selectError": false,
          "omitError": false,
          "selectBlank": false,
          "name": "Statut",
          "omitBlank": false,
          "columnName": "Statut",
          "type": "list",
          "selection": [
            {
              "v": {
                "v": "Producteur",
                "l": "Producteur"
              }
            }
          ]
        }
      ],
      "mode": "row-based"
    },
    "columnName": "Site web",
    "expression": "grel:if(isNonBlank(value), 'http://'+replace(value, /^https?:\\/\\//, ''), null)",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "op": "core/text-transform",
    "description": "Text transform on cells in column Classement Logis using expression grel:if(isNonBlank(value), value.match(/.*(\\d+)\\D+$/)[0], 0)",
    "engineConfig": {
      "facets": [
        {
          "invert": true,
          "expression": "value",
          "selectError": false,
          "omitError": false,
          "selectBlank": true,
          "name": "Classement Logis",
          "omitBlank": false,
          "columnName": "Classement Logis",
          "type": "list",
          "selection": []
        }
      ],
      "mode": "row-based"
    },
    "columnName": "Classement Logis",
    "expression": "grel:if(isNonBlank(value), value.match(/.*(\\d+)\\D+$/)[0], 0)",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "op": "core/column-addition",
    "description": "Create column Style de vie at index 10 based on column Chaînes using expression grel:import re\nfrom sets import Set\n\ntags = Set([])\ncell_labels = [\"Chaînes\", \"Spécialités culinaires\", \"Marques et labels\"]\n\nfor l in cell_labels:\n  val = cells[l][\"value\"]\n\n  if re.search(\"vélo\", val, flags=re.IGNORECASE):\n    tags.add('Bouger')\n\n  if re.search(\"ferme\", val, flags=re.IGNORECASE):\n    tags.add('Activité')\n\n  if re.search(\"bien-être\", val, flags=re.IGNORECASE):\n    tags.add('Se détendre')\n\n  if re.search(\"bio|traditionnel\", val, flags=re.IGNORECASE):\n    tags.add('Se détendre')\n\nreturn \";\".join(tags)",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "Style de vie",
    "columnInsertIndex": 10,
    "baseColumnName": "Chaînes",
    "expression": "grel:import re\nfrom sets import Set\n\ntags = Set([])\ncell_labels = [\"Chaînes\", \"Spécialités culinaires\", \"Marques et labels\"]\n\nfor l in cell_labels:\n  val = cells[l][\"value\"]\n\n  if re.search(\"vélo\", val, flags=re.IGNORECASE):\n    tags.add('Bouger')\n\n  if re.search(\"ferme\", val, flags=re.IGNORECASE):\n    tags.add('Activité')\n\n  if re.search(\"bien-être\", val, flags=re.IGNORECASE):\n    tags.add('Se détendre')\n\n  if re.search(\"bio|traditionnel\", val, flags=re.IGNORECASE):\n    tags.add('Se détendre')\n\nreturn \";\".join(tags)",
    "onError": "set-to-blank"
  }
]

refine-restaurants-export.json

{"index": {"_type": "restaurants"}}
{"name" : {{jsonize(cells["Raison sociale"].value)}}, "categories" : {{jsonize(cells["Catégorie"].value.split(";"))}}, "specialties" : {{jsonize(cells["Spécialités culinaires"].value.split(";"))}}, "address" : {{jsonize(cells["Adresse 1"].value)}}, "postal_code" : {{jsonize(cells["Code postal"].value)}}, "city" : {{jsonize(cells["Commune"].value)}}, "phone" : {{jsonize(cells["Téléphone"].value.split(";"))}}, "website" : {{jsonize(cells["Site web"].value)}}, "lifestyle": {{jsonize(cells["Style de vie"].value.split(";"))}}, "ranking" : {{jsonize(cells["Classement Logis"].value)}}, "labels" : {{jsonize(cells["Marques et labels"].value.split(";"))}}, "handicap" : {{jsonize(cells["Tourisme et Handicap"].value.split(";"))}}, "location" : [{{jsonize(cells["Longitude"].value)}}, {{jsonize(cells["Latitude"].value)}}]}