openrefine-kimws2019/doaj-openrefine.json

487 lines
12 KiB
JSON
Raw Normal View History

[
{
"op": "core/multivalued-cell-split",
"columnName": "Authors",
"keyColumnName": "Title",
"mode": "separator",
"separator": "|",
"regex": false,
"description": "Split multi-valued cells in column Authors"
},
{
"op": "core/multivalued-cell-join",
"columnName": "Authors",
"keyColumnName": "Title",
"separator": "|",
"description": "Join multi-valued cells in column Authors"
},
{
"op": "core/multivalued-cell-split",
"columnName": "Subjects",
"keyColumnName": "Title",
"mode": "separator",
"separator": "|",
"regex": false,
"description": "Split multi-valued cells in column Subjects"
},
{
"op": "core/multivalued-cell-join",
"columnName": "Subjects",
"keyColumnName": "Title",
"separator": "|",
"description": "Join multi-valued cells in column Subjects"
},
{
"op": "core/mass-edit",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Language",
"expression": "value",
"edits": [
{
"from": [
"English"
],
"fromBlank": false,
"fromError": false,
"to": "EN"
}
],
"description": "Mass edit cells in column Language"
},
{
"op": "core/multivalued-cell-split",
"columnName": "Authors",
"keyColumnName": "Title",
"mode": "separator",
"separator": "|",
"regex": false,
"description": "Split multi-valued cells in column Authors"
},
{
"op": "core/mass-edit",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Authors",
"expression": "value",
"edits": [
{
"from": [
"A. Khan Vakeel",
"Vakeel A. Khan"
],
"fromBlank": false,
"fromError": false,
"to": "A. Khan Vakeel"
},
{
"from": [
"Chandra Naveen",
"Naveen Chandra"
],
"fromBlank": false,
"fromError": false,
"to": "Chandra Naveen"
},
{
"from": [
"B. K. Revathi",
"B. K Revathi"
],
"fromBlank": false,
"fromError": false,
"to": "B. K. Revathi"
},
{
"from": [
"Santiago Garcia-Granda",
"Santiago García-Granda"
],
"fromBlank": false,
"fromError": false,
"to": "Santiago García-Granda"
},
{
"from": [
"Jian-Chao Yuan",
"Jianchao Yuan"
],
"fromBlank": false,
"fromError": false,
"to": "Jian-Chao Yuan"
},
{
"from": [
"Chang-Ge Zheng",
"ChangGe Zheng"
],
"fromBlank": false,
"fromError": false,
"to": "Chang-Ge Zheng"
},
{
"from": [
"Il'ya A. Gural'skiy",
"Il`ya A. Gural`skiy"
],
"fromBlank": false,
"fromError": false,
"to": "Il'ya A. Gural'skiy"
},
{
"from": [
"Rongbin Huang",
"Rong-Bin Huang"
],
"fromBlank": false,
"fromError": false,
"to": "Rong-Bin Huang"
},
{
"from": [
"Sheng-Lan Zhao",
"Shenglan Zhao"
],
"fromBlank": false,
"fromError": false,
"to": "Sheng-Lan Zhao"
}
],
"description": "Mass edit cells in column Authors"
},
{
"op": "core/mass-edit",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Authors",
"expression": "value",
"edits": [
{
"from": [
"R. A. Nagalakshmi",
"R.A. Nagalakshmi"
],
"fromBlank": false,
"fromError": false,
"to": "R. A. Nagalakshmi"
},
{
"from": [
"H. C. Devarajegowda",
"H.C. Devarajegowda"
],
"fromBlank": false,
"fromError": false,
"to": "H. C. Devarajegowda"
},
{
"from": [
"J. Kamal Raja",
"J. Kamalraja"
],
"fromBlank": false,
"fromError": false,
"to": "J. Kamal Raja"
},
{
"from": [
"R. V. Krishnakumar",
"R.V. Krishnakumar"
],
"fromBlank": false,
"fromError": false,
"to": "R. V. Krishnakumar"
},
{
"from": [
"Edward R. T. Tiekink",
"Edward R.T. Tiekink"
],
"fromBlank": false,
"fromError": false,
"to": "Edward R. T. Tiekink"
},
{
"from": [
"Guo -Qing Jiang",
"Guo-Qing Jiang"
],
"fromBlank": false,
"fromError": false,
"to": "Guo -Qing Jiang"
}
],
"description": "Mass edit cells in column Authors"
},
{
"op": "core/text-transform",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Publisher",
"expression": "value.trim()",
"onError": "keep-original",
"repeat": false,
"repeatCount": 10,
"description": "Text transform on cells in column Publisher using expression value.trim()"
},
{
"op": "core/text-transform",
"engineConfig": {
"facets": [
{
"type": "list",
"name": "Publisher",
"expression": "value",
"columnName": "Publisher",
"invert": false,
"omitBlank": false,
"omitError": false,
"selection": [
{
"v": {
"v": "Akshantala Enterprises",
"l": "Akshantala Enterprises"
}
},
{
"v": {
"v": "Society of Pharmaceutical Technocrats",
"l": "Society of Pharmaceutical Technocrats"
}
}
],
"selectBlank": false,
"selectError": false
}
],
"mode": "row-based"
},
"columnName": "Title",
"expression": "value.toTitlecase()",
"onError": "keep-original",
"repeat": false,
"repeatCount": 10,
"description": "Text transform on cells in column Title using expression value.toTitlecase()"
},
{
"op": "core/text-transform",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Date",
"expression": "value.toDate()",
"onError": "keep-original",
"repeat": false,
"repeatCount": 10,
"description": "Text transform on cells in column Date using expression value.toDate()"
},
{
"op": "core/column-addition",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"baseColumnName": "Date",
"expression": "grel:value.toString(\"dd MMMM yyyy\")",
"onError": "set-to-blank",
"newColumnName": "Formatted Date",
"columnInsertIndex": 5,
"description": "Create column Formatted Date at index 5 based on column Date using expression grel:value.toString(\"dd MMMM yyyy\")"
},
{
"op": "core/text-transform",
"engineConfig": {
"facets": [
{
"type": "list",
"name": "Authors",
"expression": "grel:value.contains(\",\").toString()",
"columnName": "Authors",
"invert": false,
"omitBlank": false,
"omitError": false,
"selection": [
{
"v": {
"v": "true",
"l": "true"
}
}
],
"selectBlank": false,
"selectError": false
}
],
"mode": "row-based"
},
"columnName": "Authors",
"expression": "grel:value.match(/(.*),(.*)/).reverse().join(\" \")",
"onError": "keep-original",
"repeat": false,
"repeatCount": 10,
"description": "Text transform on cells in column Authors using expression grel:value.match(/(.*),(.*)/).reverse().join(\" \")"
},
{
"op": "core/column-addition-by-fetching-urls",
"engineConfig": {
"facets": [
{
"type": "list",
"name": "ISSNs",
"expression": "grel:rowIndex == 0",
"columnName": "ISSNs",
"invert": false,
"omitBlank": false,
"omitError": false,
"selection": [
{
"v": {
"v": true,
"l": "true"
}
}
],
"selectBlank": false,
"selectError": false
}
],
"mode": "row-based"
},
"baseColumnName": "ISSNs",
"urlExpression": "grel:\"http://api.crossref.org/journals/\"+value\"",
"onError": "set-to-blank",
"newColumnName": "Journal details",
"columnInsertIndex": 9,
"delay": 5000,
"cacheResponses": true,
"httpHeadersJson": [
{
"name": "authorization",
"value": ""
},
{
"name": "user-agent",
"value": "OpenRefine 3.2-beta [8d89a2a]"
},
{
"name": "accept",
"value": "*/*"
}
],
"description": "Create column Journal details at index 9 by fetching URLs based on column ISSNs using expression grel:\"http://api.crossref.org/journals/\"+value\""
},
{
"op": "core/column-addition",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"baseColumnName": "Journal details",
"expression": "grel:value.parseJson().message.title",
"onError": "set-to-blank",
"newColumnName": "Journal Title",
"columnInsertIndex": 10,
"description": "Create column Journal Title at index 10 based on column Journal details using expression grel:value.parseJson().message.title"
},
{
"op": "core/recon",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Publisher",
"config": {
"mode": "standard-service",
"service": "http://refine.codefork.com/reconcile/viaf",
"identifierSpace": "http://rdf.freebase.com/ns/user/hangy/viaf",
"schemaSpace": "http://rdf.freebase.com/ns/type.object.id",
"type": {
"id": "/organization/organization",
"name": "Corporate Name"
},
"autoMatch": false,
"columnDetails": [],
"limit": 0
},
"description": "Reconcile cells in column Publisher to type /organization/organization"
},
{
"op": "core/recon-judge-similar-cells",
"engineConfig": {
"facets": [
{
"type": "list",
"name": "Publisher",
"expression": "value",
"columnName": "Publisher",
"invert": false,
"omitBlank": false,
"omitError": false,
"selection": [
{
"v": {
"v": "International Union of Crystallography",
"l": "International Union of Crystallography"
}
}
],
"selectBlank": false,
"selectError": false
}
],
"mode": "row-based"
},
"columnName": "Publisher",
"similarValue": "International Union of Crystallography",
"judgment": "matched",
"match": {
"id": "158070937",
"name": "International Union of Crystallography.",
"types": [
"/organization/organization"
],
"score": 0.9743589743589743
},
"shareNewTopics": false,
"description": "Match item International Union of Crystallography. (158070937) for cells containing \"International Union of Crystallography\" in column Publisher"
},
{
"op": "core/recon-match-best-candidates",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"columnName": "Publisher",
"description": "Match each cell to its best recon candidate in column Publisher"
},
{
"op": "core/column-addition",
"engineConfig": {
"facets": [],
"mode": "row-based"
},
"baseColumnName": "Publisher",
"expression": "grel:cell.recon.match.id",
"onError": "set-to-blank",
"newColumnName": "VIAF ID",
"columnInsertIndex": 12,
"description": "Create column VIAF ID at index 12 based on column Publisher using expression grel:cell.recon.match.id"
2019-04-01 20:15:29 +02:00
},
{
"op": "core/multivalued-cell-join",
"columnName": "Authors",
"keyColumnName": "Title",
"separator": "|",
"description": "Join multi-valued cells in column Authors"
}
]