DataLocale Hosting — OpenRefine + ElasticSearch Export

refine-hosting-cleanup.json

[
  {
    "op": "core/column-removal",
    "description": "Remove column File",
    "columnName": "File"
  },
  {
    "op": "core/text-transform",
    "description": "Text transform on cells in column Adresse using expression grel:if (isNonBlank(cells[\"Adresse suite\"].value), join([value, cells[\"Adresse suite\"].value], \"\\n\"), value)",
    "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": "Adresse",
    "expression": "grel:if (isNonBlank(cells[\"Adresse suite\"].value), join([value, cells[\"Adresse suite\"].value], \"\\n\"), value)",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "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 label using expression grel:if(isNonBlank(value), value.match(/.*(\\d+)\\D+$/)[0], 0)",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "columnName": "Classement label",
    "expression": "grel:if(isNonBlank(value), value.match(/.*(\\d+)\\D+$/)[0], 0)",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "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": [],
      "mode": "row-based"
    },
    "columnName": "Site web",
    "expression": "grel:if(isNonBlank(value), 'http://'+replace(value, /^https?:\\/\\//, ''), null)",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  }
]

refine-hosting-export.json

{"index": {"_type": "hosting"}}
{"name" : {{jsonize(cells["Raison sociale"].value)}}, "address" : {{jsonize(cells["Adresse"].value)}}, "postal_code" : {{jsonize(cells["Code postal"].value)}}, "city" : {{jsonize(cells["Commune"].value)}}, "phone" : {{jsonize(cells["Tél. information/réservation"].value)}}, "website" : {{jsonize(cells["Site web"].value)}}, "labels" : {{jsonize(cells["Label"].value.split(";"))}}, "label_ranking" : {{jsonize(cells["Classement label"].value)}}, "theme_labels" : {{jsonize(cells["Thèmes label"].value.split(";"))}}, "lifestyle" : {{jsonize(cells["Style de vie"].value.split(";"))}}, "handicap" : {{jsonize(cells["Tourisme et Handicap"].value)}}, "location" : [{{jsonize(cells["Longitude"].value)}}, {{jsonize(cells["Latitude"].value)}}]}