//
This script begins at line 11 and is adapted from a script developed by SiB Colombia. https://github.com/SIB-Colombia/data-quality-open-refine/blob/master/ValTaxonomicAPIWoRMS_ValTaxonomicaAPIWoRMS.txt

After creating your project in openrefine with your species data, click on the Undo/Redo tab and click the Apply button.  A new window will open. Copy lines 11 - 228 into the window in OpenRefine. Click the button to perform the operations.

After the script completes, complete a custom facet to evaluate whether the data in the scientificName field matches the data in the acceptedScientificName that was inserted by the script.

Use the returned information to edit any of the species names in your source data. Consider using additional fields from the returned data to become part of your source data.
//

[
  {
    "op": "core/text-transform",
    "description": "Text transform on cells in column scientificName",
    "engineConfig": {
      "mode": "row-based",
      "facets": []
    },
    "columnName": "scientificName",
    "expression": "grel:value.replace(/\\s+/,' ').trim()",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "op": "core/column-addition",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "baseColumnName": "scientificName",
    "expression": "grel:value.replace(' ','%20')",
    "onError": "set-to-blank",
    "newColumnName": "NomAPIw",
    "columnInsertIndex": 2,
    "description": "Create column NomAPIw at index 4 based on column scientificName using expression grel:value.replace(' ','%20')"
  },
 {
    "op": "core/column-addition-by-fetching-urls",
    "description": "Create column c at index 3 by fetching URLs based on column NomAPIw",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "callAPIworms",
    "columnInsertIndex": 3,
    "baseColumnName": "NomAPIw",
    "urlExpression": "grel:'http://www.marinespecies.org/rest/AphiaRecordsByMatchNames?scientificnames%5B%5D='+value+'&marine_only=false'",
    "onError": "set-to-blank",
    "delay": 0,
    "cacheResponses": true,
    "httpHeadersJson": [
      {
        "name": "authorization",
        "value": ""
      },
      {
        "name": "user-agent",
        "value": "OpenRefine 3.1 [b90e413]"
      },
      {
        "name": "accept",
        "value": "*/*"
      }
    ]
  },
  {
    "op": "core/text-transform",
    "description": "Text transform on cells in column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "columnName": "callAPIworms",
    "expression": "grel:value.replace('[','').replace(']','')",
    "onError": "keep-original",
    "repeat": false,
    "repeatCount": 10
  },
  {
    "op": "core/column-addition",
    "description": "Create column specificEpithet at index 2 based on column callAPIworms",
    "engineConfig": {
      "mode": "row-based",
      "facets": []
    },
    "newColumnName": "specificEpithet",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:split(value.parseJson().get(\"scientificname\"),' ')[1]",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-move",
    "description": "Move column specificEpithet to position 2",
    "columnName": "specificEpithet",
    "index": 2
  },
  {
    "op": "core/column-addition",
    "description": "Create column genus at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "genus",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('genus')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-addition",
    "description": "Create column family at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "family",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('family')",
    "onError": "set-to-blank"
  },
     {
    "op": "core/column-addition",
    "description": "Create column order at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "order",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('order')",
    "onError": "set-to-blank"
  },
   {
    "op": "core/column-addition",
    "description": "Create column class at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "class",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('class')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-addition",
    "description": "Create column phylum at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "phylum",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('phylum')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-addition",
    "description": "Create column kingdom at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "kingdom",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('kingdom')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-addition",
    "description": "Create column taxonRank at index 2 based on column callAPIworms",
    "engineConfig": {
      "facets": [],
      "mode": "row-based"
    },
    "newColumnName": "taxonRank",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('rank')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-addition",
    "description": "Create column acceptedScientificName at index 2 based on column callAPIworms",
    "engineConfig": {
      "mode": "row-based",
      "facets": []
    },
    "newColumnName": "acceptedScientificName",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('valid_name')",
    "onError": "set-to-blank"
  },
  {
    "op": "core/column-move",
    "description": "Move column scientificName to position 2",
    "columnName": "scientificName",
    "index": 2
  },
  {
    "op": "core/column-move",
    "description": "Move column previousIdentifications to position 4",
    "columnName": "previousIdentifications",
    "index": 4
  },
  {
    "op": "core/column-addition",
    "description": "Create column scientificNameID at index 2 based on column callAPIworms",
    "engineConfig": {
      "mode": "row-based",
      "facets": []
    },
    "newColumnName": "scientificNameID",
    "columnInsertIndex": 2,
    "baseColumnName": "callAPIworms",
    "expression": "grel:value.parseJson().get('lsid')",
    "onError": "set-to-blank"
  }
]