doidoost_dismiss #418

Merged
claudio.atzori merged 30 commits from doidoost_dismiss into beta 2024-04-17 12:01:14 +02:00
20 changed files with 9357 additions and 2258 deletions
Showing only changes of commit 58dbe71d39 - Show all commits

View File

@ -5,8 +5,6 @@ import org.apache.commons.lang3.StringUtils;
public class DoiCleaningRule { public class DoiCleaningRule {
public static String clean(final String doi) { public static String clean(final String doi) {
return doi return doi
.toLowerCase() .toLowerCase()
@ -28,7 +26,6 @@ public class DoiCleaningRule {
if (!replaced.contains("10.")) if (!replaced.contains("10."))
return null; return null;
final String ret = replaced.substring(replaced.indexOf("10.")); final String ret = replaced.substring(replaced.indexOf("10."));
if (!ret.startsWith(CleaningFunctions.DOI_PREFIX)) if (!ret.startsWith(CleaningFunctions.DOI_PREFIX))

View File

@ -0,0 +1,26 @@
[
{
"paramName": "m",
"paramLongName": "master",
"paramDescription": "the master name",
"paramRequired": true
},
{
"paramName": "s",
"paramLongName": "sourcePath",
"paramDescription": "The base path of Crossref DUMP",
"paramRequired": true
},
{
"paramName": "t",
"paramLongName": "targetPath",
"paramDescription": "The target path",
"paramRequired": false
},
{
"paramName": "i",
"paramLongName": "isLookupUrl",
"paramDescription": "the Information System Service LookUp URL",
"paramRequired": true
}
]

View File

@ -0,0 +1,940 @@
[
{
"id": "100007630",
"uri": "http://dx.doi.org/10.13039/100007630",
"name": "College of Engineering and Informatics, National University of Ireland, Galway",
"synonym": []
},
{
"id": "100007731",
"uri": "http://dx.doi.org/10.13039/100007731",
"name": "Endo International",
"synonym": []
},
{
"id": "100008099",
"uri": "http://dx.doi.org/10.13039/100008099",
"name": "Food Safety Authority of Ireland",
"synonym": []
},
{
"id": "100008124",
"uri": "http://dx.doi.org/10.13039/100008124",
"name": "Department of Jobs, Enterprise and Innovation",
"synonym": []
},
{
"id": "100009098",
"uri": "http://dx.doi.org/10.13039/100009098",
"name": "Department of Foreign Affairs and Trade, Ireland",
"synonym": []
},
{
"id": "100009099",
"uri": "http://dx.doi.org/10.13039/100009099",
"name": "Irish Aid",
"synonym": []
},
{
"id": "100009770",
"uri": "http://dx.doi.org/10.13039/100009770",
"name": "National University of Ireland",
"synonym": []
},
{
"id": "100009985",
"uri": "http://dx.doi.org/10.13039/100009985",
"name": "Parkinson's Association of Ireland",
"synonym": []
},
{
"id": "100010399",
"uri": "http://dx.doi.org/10.13039/100010399",
"name": "European Society of Cataract and Refractive Surgeons",
"synonym": []
},
{
"id": "100010414",
"uri": "http://dx.doi.org/10.13039/100010414",
"name": "Health Research Board",
"synonym": [
"501100001590"
]
},
{
"id": "100010546",
"uri": "http://dx.doi.org/10.13039/100010546",
"name": "Deparment of Children and Youth Affairs, Ireland",
"synonym": []
},
{
"id": "100010993",
"uri": "http://dx.doi.org/10.13039/100010993",
"name": "Irish Nephrology Society",
"synonym": []
},
{
"id": "100011062",
"uri": "http://dx.doi.org/10.13039/100011062",
"name": "Asian Spinal Cord Network",
"synonym": []
},
{
"id": "100011096",
"uri": "http://dx.doi.org/10.13039/100011096",
"name": "Jazz Pharmaceuticals",
"synonym": []
},
{
"id": "100011396",
"uri": "http://dx.doi.org/10.13039/100011396",
"name": "Irish College of General Practitioners",
"synonym": []
},
{
"id": "100012734",
"uri": "http://dx.doi.org/10.13039/100012734",
"name": "Department for Culture, Heritage and the Gaeltacht, Ireland",
"synonym": []
},
{
"id": "100012754",
"uri": "http://dx.doi.org/10.13039/100012754",
"name": "Horizon Pharma",
"synonym": []
},
{
"id": "100012891",
"uri": "http://dx.doi.org/10.13039/100012891",
"name": "Medical Research Charities Group",
"synonym": []
},
{
"id": "100012919",
"uri": "http://dx.doi.org/10.13039/100012919",
"name": "Epilepsy Ireland",
"synonym": []
},
{
"id": "100012920",
"uri": "http://dx.doi.org/10.13039/100012920",
"name": "GLEN",
"synonym": []
},
{
"id": "100012921",
"uri": "http://dx.doi.org/10.13039/100012921",
"name": "Royal College of Surgeons in Ireland",
"synonym": []
},
{
"id": "100013029",
"uri": "http://dx.doi.org/10.13039/100013029",
"name": "Iris O'Brien Foundation",
"synonym": []
},
{
"id": "100013206",
"uri": "http://dx.doi.org/10.13039/100013206",
"name": "Food Institutional Research Measure",
"synonym": []
},
{
"id": "100013381",
"uri": "http://dx.doi.org/10.13039/100013381",
"name": "Irish Phytochemical Food Network",
"synonym": []
},
{
"id": "100013433",
"uri": "http://dx.doi.org/10.13039/100013433",
"name": "Transport Infrastructure Ireland",
"synonym": []
},
{
"id": "100013461",
"uri": "http://dx.doi.org/10.13039/100013461",
"name": "Arts and Disability Ireland",
"synonym": []
},
{
"id": "100013548",
"uri": "http://dx.doi.org/10.13039/100013548",
"name": "Filmbase",
"synonym": []
},
{
"id": "100013917",
"uri": "http://dx.doi.org/10.13039/100013917",
"name": "Society for Musicology in Ireland",
"synonym": []
},
{
"id": "100014251",
"uri": "http://dx.doi.org/10.13039/100014251",
"name": "Humanities in the European Research Area",
"synonym": []
},
{
"id": "100014364",
"uri": "http://dx.doi.org/10.13039/100014364",
"name": "National Children's Research Centre",
"synonym": []
},
{
"id": "100014384",
"uri": "http://dx.doi.org/10.13039/100014384",
"name": "Amarin Corporation",
"synonym": []
},
{
"id": "100014902",
"uri": "http://dx.doi.org/10.13039/100014902",
"name": "Irish Association for Cancer Research",
"synonym": []
},
{
"id": "100015023",
"uri": "http://dx.doi.org/10.13039/100015023",
"name": "Ireland Funds",
"synonym": []
},
{
"id": "100015037",
"uri": "http://dx.doi.org/10.13039/100015037",
"name": "Simon Cumbers Media Fund",
"synonym": []
},
{
"id": "100015319",
"uri": "http://dx.doi.org/10.13039/100015319",
"name": "Sport Ireland Institute",
"synonym": []
},
{
"id": "100015320",
"uri": "http://dx.doi.org/10.13039/100015320",
"name": "Paralympics Ireland",
"synonym": []
},
{
"id": "100015442",
"uri": "http://dx.doi.org/10.13039/100015442",
"name": "Global Brain Health Institute",
"synonym": []
},
{
"id": "100015776",
"uri": "http://dx.doi.org/10.13039/100015776",
"name": "Health and Social Care Board",
"synonym": []
},
{
"id": "100015992",
"uri": "http://dx.doi.org/10.13039/100015992",
"name": "St. Luke's Institute of Cancer Research",
"synonym": []
},
{
"id": "100017897",
"uri": "http://dx.doi.org/10.13039/100017897",
"name": "Friedreich\u2019s Ataxia Research Alliance Ireland",
"synonym": []
},
{
"id": "100018064",
"uri": "http://dx.doi.org/10.13039/100018064",
"name": "Department of Tourism, Culture, Arts, Gaeltacht, Sport and Media",
"synonym": []
},
{
"id": "100018172",
"uri": "http://dx.doi.org/10.13039/100018172",
"name": "Department of the Environment, Climate and Communications",
"synonym": []
},
{
"id": "100018175",
"uri": "http://dx.doi.org/10.13039/100018175",
"name": "Dairy Processing Technology Centre",
"synonym": []
},
{
"id": "100018270",
"uri": "http://dx.doi.org/10.13039/100018270",
"name": "Health Service Executive",
"synonym": []
},
{
"id": "100018529",
"uri": "http://dx.doi.org/10.13039/100018529",
"name": "Alkermes",
"synonym": []
},
{
"id": "100018542",
"uri": "http://dx.doi.org/10.13039/100018542",
"name": "Irish Endocrine Society",
"synonym": []
},
{
"id": "100018754",
"uri": "http://dx.doi.org/10.13039/100018754",
"name": "An Roinn Sl\u00e1inte",
"synonym": []
},
{
"id": "100018998",
"uri": "http://dx.doi.org/10.13039/100018998",
"name": "Irish Research eLibrary",
"synonym": []
},
{
"id": "100019428",
"uri": "http://dx.doi.org/10.13039/100019428",
"name": "Nabriva Therapeutics",
"synonym": []
},
{
"id": "100019637",
"uri": "http://dx.doi.org/10.13039/100019637",
"name": "Horizon Therapeutics",
"synonym": []
},
{
"id": "100020174",
"uri": "http://dx.doi.org/10.13039/100020174",
"name": "Health Research Charities Ireland",
"synonym": []
},
{
"id": "100020202",
"uri": "http://dx.doi.org/10.13039/100020202",
"name": "UCD Foundation",
"synonym": []
},
{
"id": "100020233",
"uri": "http://dx.doi.org/10.13039/100020233",
"name": "Ireland Canada University Foundation",
"synonym": []
},
{
"id": "100022943",
"uri": "http://dx.doi.org/10.13039/100022943",
"name": "National Cancer Registry Ireland",
"synonym": []
},
{
"id": "501100001581",
"uri": "http://dx.doi.org/10.13039/501100001581",
"name": "Arts Council of Ireland",
"synonym": []
},
{
"id": "501100001582",
"uri": "http://dx.doi.org/10.13039/501100001582",
"name": "Centre for Ageing Research and Development in Ireland",
"synonym": []
},
{
"id": "501100001583",
"uri": "http://dx.doi.org/10.13039/501100001583",
"name": "Cystinosis Foundation Ireland",
"synonym": []
},
{
"id": "501100001584",
"uri": "http://dx.doi.org/10.13039/501100001584",
"name": "Department of Agriculture, Food and the Marine, Ireland",
"synonym": []
},
{
"id": "501100001586",
"uri": "http://dx.doi.org/10.13039/501100001586",
"name": "Department of Education and Skills, Ireland",
"synonym": []
},
{
"id": "501100001587",
"uri": "http://dx.doi.org/10.13039/501100001587",
"name": "Economic and Social Research Institute",
"synonym": []
},
{
"id": "501100001588",
"uri": "http://dx.doi.org/10.13039/501100001588",
"name": "Enterprise Ireland",
"synonym": []
},
{
"id": "501100001589",
"uri": "http://dx.doi.org/10.13039/501100001589",
"name": "Environmental Protection Agency",
"synonym": []
},
{
"id": "501100001591",
"uri": "http://dx.doi.org/10.13039/501100001591",
"name": "Heritage Council",
"synonym": []
},
{
"id": "501100001592",
"uri": "http://dx.doi.org/10.13039/501100001592",
"name": "Higher Education Authority",
"synonym": []
},
{
"id": "501100001593",
"uri": "http://dx.doi.org/10.13039/501100001593",
"name": "Irish Cancer Society",
"synonym": []
},
{
"id": "501100001594",
"uri": "http://dx.doi.org/10.13039/501100001594",
"name": "Irish Heart Foundation",
"synonym": []
},
{
"id": "501100001595",
"uri": "http://dx.doi.org/10.13039/501100001595",
"name": "Irish Hospice Foundation",
"synonym": []
},
{
"id": "501100001596",
"uri": "http://dx.doi.org/10.13039/501100001596",
"name": "Irish Research Council for Science, Engineering and Technology",
"synonym": []
},
{
"id": "501100001597",
"uri": "http://dx.doi.org/10.13039/501100001597",
"name": "Irish Research Council for the Humanities and Social Sciences",
"synonym": []
},
{
"id": "501100001598",
"uri": "http://dx.doi.org/10.13039/501100001598",
"name": "Mental Health Commission",
"synonym": []
},
{
"id": "501100001600",
"uri": "http://dx.doi.org/10.13039/501100001600",
"name": "Research and Education Foundation, Sligo General Hospital",
"synonym": []
},
{
"id": "501100001601",
"uri": "http://dx.doi.org/10.13039/501100001601",
"name": "Royal Irish Academy",
"synonym": []
},
{
"id": "501100001603",
"uri": "http://dx.doi.org/10.13039/501100001603",
"name": "Sustainable Energy Authority of Ireland",
"synonym": []
},
{
"id": "501100001604",
"uri": "http://dx.doi.org/10.13039/501100001604",
"name": "Teagasc",
"synonym": []
},
{
"id": "501100001627",
"uri": "http://dx.doi.org/10.13039/501100001627",
"name": "Marine Institute",
"synonym": []
},
{
"id": "501100001628",
"uri": "http://dx.doi.org/10.13039/501100001628",
"name": "Central Remedial Clinic",
"synonym": []
},
{
"id": "501100001629",
"uri": "http://dx.doi.org/10.13039/501100001629",
"name": "Royal Dublin Society",
"synonym": []
},
{
"id": "501100001630",
"uri": "http://dx.doi.org/10.13039/501100001630",
"name": "Dublin Institute for Advanced Studies",
"synonym": []
},
{
"id": "501100001631",
"uri": "http://dx.doi.org/10.13039/501100001631",
"name": "University College Dublin",
"synonym": []
},
{
"id": "501100001633",
"uri": "http://dx.doi.org/10.13039/501100001633",
"name": "National University of Ireland, Maynooth",
"synonym": []
},
{
"id": "501100001634",
"uri": "http://dx.doi.org/10.13039/501100001634",
"name": "University of Galway",
"synonym": []
},
{
"id": "501100001635",
"uri": "http://dx.doi.org/10.13039/501100001635",
"name": "University of Limerick",
"synonym": []
},
{
"id": "501100001636",
"uri": "http://dx.doi.org/10.13039/501100001636",
"name": "University College Cork",
"synonym": []
},
{
"id": "501100001637",
"uri": "http://dx.doi.org/10.13039/501100001637",
"name": "Trinity College Dublin",
"synonym": []
},
{
"id": "501100001638",
"uri": "http://dx.doi.org/10.13039/501100001638",
"name": "Dublin City University",
"synonym": []
},
{
"id": "501100002081",
"uri": "http://dx.doi.org/10.13039/501100002081",
"name": "Irish Research Council",
"synonym": []
},
{
"id": "501100002736",
"uri": "http://dx.doi.org/10.13039/501100002736",
"name": "Covidien",
"synonym": []
},
{
"id": "501100002755",
"uri": "http://dx.doi.org/10.13039/501100002755",
"name": "Brennan and Company",
"synonym": []
},
{
"id": "501100002919",
"uri": "http://dx.doi.org/10.13039/501100002919",
"name": "Cork Institute of Technology",
"synonym": []
},
{
"id": "501100002959",
"uri": "http://dx.doi.org/10.13039/501100002959",
"name": "Dublin City Council",
"synonym": []
},
{
"id": "501100003036",
"uri": "http://dx.doi.org/10.13039/501100003036",
"name": "Perrigo Company Charitable Foundation",
"synonym": []
},
{
"id": "501100003037",
"uri": "http://dx.doi.org/10.13039/501100003037",
"name": "Elan",
"synonym": []
},
{
"id": "501100003496",
"uri": "http://dx.doi.org/10.13039/501100003496",
"name": "HeyStaks Technologies",
"synonym": []
},
{
"id": "501100003553",
"uri": "http://dx.doi.org/10.13039/501100003553",
"name": "Gaelic Athletic Association",
"synonym": []
},
{
"id": "501100003840",
"uri": "http://dx.doi.org/10.13039/501100003840",
"name": "Irish Institute of Clinical Neuroscience",
"synonym": []
},
{
"id": "501100003956",
"uri": "http://dx.doi.org/10.13039/501100003956",
"name": "Aspect Medical Systems",
"synonym": []
},
{
"id": "501100004162",
"uri": "http://dx.doi.org/10.13039/501100004162",
"name": "Meath Foundation",
"synonym": []
},
{
"id": "501100004210",
"uri": "http://dx.doi.org/10.13039/501100004210",
"name": "Our Lady's Children's Hospital, Crumlin",
"synonym": []
},
{
"id": "501100004321",
"uri": "http://dx.doi.org/10.13039/501100004321",
"name": "Shire",
"synonym": []
},
{
"id": "501100004981",
"uri": "http://dx.doi.org/10.13039/501100004981",
"name": "Athlone Institute of Technology",
"synonym": []
},
{
"id": "501100006518",
"uri": "http://dx.doi.org/10.13039/501100006518",
"name": "Department of Communications, Energy and Natural Resources, Ireland",
"synonym": []
},
{
"id": "501100006553",
"uri": "http://dx.doi.org/10.13039/501100006553",
"name": "Collaborative Centre for Applied Nanotechnology",
"synonym": []
},
{
"id": "501100006759",
"uri": "http://dx.doi.org/10.13039/501100006759",
"name": "CLARITY Centre for Sensor Web Technologies",
"synonym": []
},
{
"id": "501100009246",
"uri": "http://dx.doi.org/10.13039/501100009246",
"name": "Technological University Dublin",
"synonym": []
},
{
"id": "501100009269",
"uri": "http://dx.doi.org/10.13039/501100009269",
"name": "Programme of Competitive Forestry Research for Development",
"synonym": []
},
{
"id": "501100009315",
"uri": "http://dx.doi.org/10.13039/501100009315",
"name": "Cystinosis Ireland",
"synonym": []
},
{
"id": "501100010808",
"uri": "http://dx.doi.org/10.13039/501100010808",
"name": "Geological Survey of Ireland",
"synonym": []
},
{
"id": "501100011030",
"uri": "http://dx.doi.org/10.13039/501100011030",
"name": "Alimentary Glycoscience Research Cluster",
"synonym": []
},
{
"id": "501100011031",
"uri": "http://dx.doi.org/10.13039/501100011031",
"name": "Alimentary Health",
"synonym": []
},
{
"id": "501100011103",
"uri": "http://dx.doi.org/10.13039/501100011103",
"name": "Rann\u00eds",
"synonym": []
},
{
"id": "501100012354",
"uri": "http://dx.doi.org/10.13039/501100012354",
"name": "Inland Fisheries Ireland",
"synonym": []
},
{
"id": "501100014384",
"uri": "http://dx.doi.org/10.13039/501100014384",
"name": "X-Bolt Orthopaedics",
"synonym": []
},
{
"id": "501100014710",
"uri": "http://dx.doi.org/10.13039/501100014710",
"name": "PrecisionBiotics Group",
"synonym": []
},
{
"id": "501100014827",
"uri": "http://dx.doi.org/10.13039/501100014827",
"name": "Dormant Accounts Fund",
"synonym": []
},
{
"id": "501100016041",
"uri": "http://dx.doi.org/10.13039/501100016041",
"name": "St Vincents Anaesthesia Foundation",
"synonym": []
},
{
"id": "501100017501",
"uri": "http://dx.doi.org/10.13039/501100017501",
"name": "FotoNation",
"synonym": []
},
{
"id": "501100018641",
"uri": "http://dx.doi.org/10.13039/501100018641",
"name": "Dairy Research Ireland",
"synonym": []
},
{
"id": "501100018839",
"uri": "http://dx.doi.org/10.13039/501100018839",
"name": "Irish Centre for High-End Computing",
"synonym": []
},
{
"id": "501100019905",
"uri": "http://dx.doi.org/10.13039/501100019905",
"name": "Galway University Foundation",
"synonym": []
},
{
"id": "501100020036",
"uri": "http://dx.doi.org/10.13039/501100020036",
"name": "Dystonia Ireland",
"synonym": []
},
{
"id": "501100020221",
"uri": "http://dx.doi.org/10.13039/501100020221",
"name": "Irish Motor Neurone Disease Association",
"synonym": []
},
{
"id": "501100020270",
"uri": "http://dx.doi.org/10.13039/501100020270",
"name": "Advanced Materials and Bioengineering Research",
"synonym": []
},
{
"id": "501100020403",
"uri": "http://dx.doi.org/10.13039/501100020403",
"name": "Irish Composites Centre",
"synonym": []
},
{
"id": "501100020425",
"uri": "http://dx.doi.org/10.13039/501100020425",
"name": "Irish Thoracic Society",
"synonym": []
},
{
"id": "501100021102",
"uri": "http://dx.doi.org/10.13039/501100021102",
"name": "Waterford Institute of Technology",
"synonym": []
},
{
"id": "501100021110",
"uri": "http://dx.doi.org/10.13039/501100021110",
"name": "Irish MPS Society",
"synonym": []
},
{
"id": "501100021525",
"uri": "http://dx.doi.org/10.13039/501100021525",
"name": "Insight SFI Research Centre for Data Analytics",
"synonym": []
},
{
"id": "501100021694",
"uri": "http://dx.doi.org/10.13039/501100021694",
"name": "Elan Pharma International",
"synonym": []
},
{
"id": "501100021838",
"uri": "http://dx.doi.org/10.13039/501100021838",
"name": "Royal College of Physicians of Ireland",
"synonym": []
},
{
"id": "501100022542",
"uri": "http://dx.doi.org/10.13039/501100022542",
"name": "Breakthrough Cancer Research",
"synonym": []
},
{
"id": "501100022610",
"uri": "http://dx.doi.org/10.13039/501100022610",
"name": "Breast Cancer Ireland",
"synonym": []
},
{
"id": "501100022728",
"uri": "http://dx.doi.org/10.13039/501100022728",
"name": "Munster Technological University",
"synonym": []
},
{
"id": "501100022729",
"uri": "http://dx.doi.org/10.13039/501100022729",
"name": "Institute of Technology, Tralee",
"synonym": []
},
{
"id": "501100023273",
"uri": "http://dx.doi.org/10.13039/501100023273",
"name": "HRB Clinical Research Facility Galway",
"synonym": []
},
{
"id": "501100023378",
"uri": "http://dx.doi.org/10.13039/501100023378",
"name": "Lauritzson Foundation",
"synonym": []
},
{
"id": "501100023551",
"uri": "http://dx.doi.org/10.13039/501100023551",
"name": "Cystic Fibrosis Ireland",
"synonym": []
},
{
"id": "501100023970",
"uri": "http://dx.doi.org/10.13039/501100023970",
"name": "Tyndall National Institute",
"synonym": []
},
{
"id": "501100024094",
"uri": "http://dx.doi.org/10.13039/501100024094",
"name": "Raidi\u00f3 Teilif\u00eds \u00c9ireann",
"synonym": []
},
{
"id": "501100024242",
"uri": "http://dx.doi.org/10.13039/501100024242",
"name": "Synthesis and Solid State Pharmaceutical Centre",
"synonym": []
},
{
"id": "501100024313",
"uri": "http://dx.doi.org/10.13039/501100024313",
"name": "Irish Rugby Football Union",
"synonym": []
},
{
"id": "100007490",
"uri": "http://dx.doi.org/10.13039/100007490",
"name": "Bausch and Lomb Ireland",
"synonym": []
},
{
"id": "100007819",
"uri": "http://dx.doi.org/10.13039/100007819",
"name": "Allergan",
"synonym": []
},
{
"id": "100010547",
"uri": "http://dx.doi.org/10.13039/100010547",
"name": "Irish Youth Justice Service",
"synonym": []
},
{
"id": "100012733",
"uri": "http://dx.doi.org/10.13039/100012733",
"name": "National Parks and Wildlife Service",
"synonym": []
},
{
"id": "100015278",
"uri": "http://dx.doi.org/10.13039/100015278",
"name": "Pfizer Healthcare Ireland",
"synonym": []
},
{
"id": "100017144",
"uri": "http://dx.doi.org/10.13039/100017144",
"name": "Shell E and P Ireland",
"synonym": []
},
{
"id": "100022895",
"uri": "http://dx.doi.org/10.13039/100022895",
"name": "Health Research Institute, University of Limerick",
"synonym": []
},
{
"id": "501100001599",
"uri": "http://dx.doi.org/10.13039/501100001599",
"name": "National Council for Forest Research and Development",
"synonym": []
},
{
"id": "501100006554",
"uri": "http://dx.doi.org/10.13039/501100006554",
"name": "IDA Ireland",
"synonym": []
},
{
"id": "501100011626",
"uri": "http://dx.doi.org/10.13039/501100011626",
"name": "Energy Policy Research Centre, Economic and Social Research Institute",
"synonym": []
},
{
"id": "501100014531",
"uri": "http://dx.doi.org/10.13039/501100014531",
"name": "Physical Education and Sport Sciences Department, University of Limerick",
"synonym": []
},
{
"id": "501100014745",
"uri": "http://dx.doi.org/10.13039/501100014745",
"name": "APC Microbiome Institute",
"synonym": []
},
{
"id": "501100014826",
"uri": "http://dx.doi.org/10.13039/501100014826",
"name": "ADAPT - Centre for Digital Content Technology",
"synonym": []
},
{
"id": "501100020570",
"uri": "http://dx.doi.org/10.13039/501100020570",
"name": "College of Medicine, Nursing and Health Sciences, National University of Ireland, Galway",
"synonym": []
},
{
"id": "501100020871",
"uri": "http://dx.doi.org/10.13039/501100020871",
"name": "Bernal Institute, University of Limerick",
"synonym": []
},
{
"id": "501100023852",
"uri": "http://dx.doi.org/10.13039/501100023852",
"name": "Moore Institute for Research in the Humanities and Social Studies, University of Galway",
"synonym": []
}
]

View File

@ -0,0 +1,44 @@
<configuration>
<property>
<name>jobTracker</name>
<value>yarnRM</value>
</property>
<property>
<name>nameNode</name>
<value>hdfs://nameservice1</value>
</property>
<property>
<name>oozie.use.system.libpath</name>
<value>true</value>
</property>
<property>
<name>oozie.action.sharelib.for.spark</name>
<value>spark2</value>
</property>
<property>
<name>oozie.launcher.mapreduce.user.classpath.first</name>
<value>true</value>
</property>
<property>
<name>spark2ExtraListeners</name>
<value>com.cloudera.spark.lineage.NavigatorAppListener</value>
</property>
<property>
<name>spark2SqlQueryExecutionListeners</name>
<value>com.cloudera.spark.lineage.NavigatorQueryListener</value>
</property>
<property>
<name>spark2YarnHistoryServerAddress</name>
<value>http://iis-cdh5-test-gw.ocean.icm.edu.pl:18089 </value>
</property>
<property>
<name>spark2EventLogDir</name>
<value>/user/spark/spark2ApplicationHistory</value>
</property>
</configuration>

View File

@ -0,0 +1,54 @@
<workflow-app name="generate_crossref_Datasource" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>sourcePath</name>
<description>The base path of MAG DUMP CSV Tables</description>
</property>
<property>
<name>targetPath</name>
<description>The base path of MAG DUMP CSV Tables</description>
</property>
<property>
<name>isLookupUrl</name>
<description>The Information service Lookup URL</description>
</property>
</parameters>
<start to="generateOAF"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="generateOAF">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Crossref TO OAF</name>
<class>eu.dnetlib.dhp.collection.crossref.SparkMapDumpIntoOAF</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.executor.memoryOverhead=2g
--conf spark.sql.shuffle.partitions=3000
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>

View File

@ -1,10 +1,17 @@
package eu.dnetlib.dhp.collection.crossref package eu.dnetlib.dhp.collection.crossref
import com.fasterxml.jackson.databind.ObjectMapper
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup
import eu.dnetlib.dhp.schema.common.ModelConstants import eu.dnetlib.dhp.schema.common.ModelConstants
import eu.dnetlib.dhp.schema.oaf._ import eu.dnetlib.dhp.schema.oaf._
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils.{field, qualifier, structuredProperty, subject} import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils.{field, qualifier, structuredProperty, subject}
import eu.dnetlib.dhp.schema.oaf.utils.{DoiCleaningRule, GraphCleaningFunctions, IdentifierFactory, OafMapperUtils, PidType} import eu.dnetlib.dhp.schema.oaf.utils.{
DoiCleaningRule,
GraphCleaningFunctions,
IdentifierFactory,
OafMapperUtils,
PidType
}
import eu.dnetlib.dhp.utils.DHPUtils import eu.dnetlib.dhp.utils.DHPUtils
import org.apache.commons.lang.StringUtils import org.apache.commons.lang.StringUtils
import org.json4s import org.json4s
@ -37,18 +44,33 @@ case class funderInfo(id: String, uri: String, name: String, synonym: List[Strin
case class mappingFunder(name: String, DOI: Option[String], award: Option[List[String]]) {} case class mappingFunder(name: String, DOI: Option[String], award: Option[List[String]]) {}
case class CrossrefResult(oafType: String, body: String) {}
case object Crossref2Oaf { case object Crossref2Oaf {
val logger: Logger = LoggerFactory.getLogger(Crossref2Oaf.getClass) val logger: Logger = LoggerFactory.getLogger(Crossref2Oaf.getClass)
val mapper = new ObjectMapper
val irishFunder: List[funderInfo] = { val irishFunder: List[funderInfo] = {
val s = Source val s = Source
.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/doiboost/crossref/irish_funder.json")) .fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/collection/crossref/irish_funder.json"))
.mkString .mkString
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
lazy val json: org.json4s.JValue = parse(s) lazy val json: org.json4s.JValue = parse(s)
json.extract[List[funderInfo]] json.extract[List[funderInfo]]
} }
val invalidName = List(
",",
"none none",
"none, none",
"none &na;",
"(:null)",
"test test test",
"test test",
"test",
"&na; &na;"
)
def getIrishId(doi: String): Option[String] = { def getIrishId(doi: String): Option[String] = {
val id = doi.split("/").last val id = doi.split("/").last
irishFunder irishFunder
@ -64,8 +86,9 @@ case object Crossref2Oaf {
cf cf
} }
def generateDataInfo(): DataInfo = { def generateDataInfo(): DataInfo = {
generateDataInfo("0.9") generateDataInfo("0.91")
} }
def generateDataInfo(trust: String): DataInfo = { def generateDataInfo(trust: String): DataInfo = {
@ -84,6 +107,7 @@ case object Crossref2Oaf {
) )
di di
} }
def getOpenAccessQualifier(): AccessRight = { def getOpenAccessQualifier(): AccessRight = {
OafMapperUtils.accessRight( OafMapperUtils.accessRight(
@ -102,6 +126,7 @@ case object Crossref2Oaf {
ModelConstants.DNET_ACCESS_MODES ModelConstants.DNET_ACCESS_MODES
) )
} }
def getUnknownQualifier(): AccessRight = { def getUnknownQualifier(): AccessRight = {
OafMapperUtils.accessRight( OafMapperUtils.accessRight(
ModelConstants.UNKNOWN, ModelConstants.UNKNOWN,
@ -197,14 +222,81 @@ case object Crossref2Oaf {
} }
def isValidAuthorName(fullName: String): Boolean = {
if (fullName == null || fullName.isEmpty)
return false
if (invalidName.contains(fullName.toLowerCase.trim))
return false
true
}
def filterResult(publication: Result): Boolean = {
//Case empty publication
if (publication == null)
return false
if (publication.getId == null || publication.getId.isEmpty)
return false
//Case publication with no title
if (publication.getTitle == null || publication.getTitle.size == 0)
return false
val s = publication.getTitle.asScala.count(p =>
p.getValue != null
&& p.getValue.nonEmpty && !p.getValue.equalsIgnoreCase("[NO TITLE AVAILABLE]")
)
if (s == 0)
return false
// fixes #4360 (test publisher)
val publisher =
if (publication.getPublisher != null) publication.getPublisher.getValue else null
if (
publisher != null && (publisher.equalsIgnoreCase("Test accounts") || publisher
.equalsIgnoreCase("CrossRef Test Account"))
) {
return false;
}
//RELAXED this constraint
//Publication with no Author
if (publication.getAuthor == null || publication.getAuthor.size() == 0)
return true
//filter invalid author
val authors = publication.getAuthor.asScala.map(s => {
if (s.getFullname.nonEmpty) {
s.getFullname
} else
s"${s.getName} ${s.getSurname}"
})
val c = authors.count(isValidAuthorName)
if (c == 0)
return false
// fixes #4368
if (
authors.count(s => s.equalsIgnoreCase("Addie Jackson")) > 0 && "Elsevier BV".equalsIgnoreCase(
publication.getPublisher.getValue
)
)
return false
true
}
def mappingResult(result: Result, json: JValue, instanceType: Qualifier, originalType: String): Result = { def mappingResult(result: Result, json: JValue, instanceType: Qualifier, originalType: String): Result = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
//MAPPING Crossref DOI into PID //MAPPING Crossref DOI into PID
val doi: String = DoiCleaningRule.normalizeDoi((json \ "DOI").extract[String]) val doi: String = DoiCleaningRule.normalizeDoi((json \ "DOI").extract[String])
result.setPid(List(structuredProperty( result.setPid(
List(
structuredProperty(
doi, doi,
qualifier( qualifier(
PidType.doi.toString, PidType.doi.toString,
@ -213,7 +305,9 @@ case object Crossref2Oaf {
ModelConstants.DNET_PID_TYPES ModelConstants.DNET_PID_TYPES
), ),
null null
)).asJava) )
).asJava
)
//MAPPING Crossref DOI into OriginalId //MAPPING Crossref DOI into OriginalId
//and Other Original Identifier of dataset like clinical-trial-number //and Other Original Identifier of dataset like clinical-trial-number
@ -251,7 +345,9 @@ case object Crossref2Oaf {
val subtitles = val subtitles =
for { JString(title) <- json \ "subtitle" if title.nonEmpty } yield structuredProperty( for { JString(title) <- json \ "subtitle" if title.nonEmpty } yield structuredProperty(
title, title,
ModelConstants.SUBTITLE_QUALIFIER, null) ModelConstants.SUBTITLE_QUALIFIER,
null
)
result.setTitle((mainTitles ::: originalTitles ::: shortTitles ::: subtitles).asJava) result.setTitle((mainTitles ::: originalTitles ::: shortTitles ::: subtitles).asJava)
// DESCRIPTION // DESCRIPTION
@ -415,6 +511,7 @@ case object Crossref2Oaf {
else else
result result
} }
def generateIdentifier(oaf: Result, doi: String): String = { def generateIdentifier(oaf: Result, doi: String): String = {
val id = DHPUtils.md5(doi.toLowerCase) val id = DHPUtils.md5(doi.toLowerCase)
s"50|doiboost____|$id" s"50|doiboost____|$id"
@ -431,7 +528,12 @@ case object Crossref2Oaf {
List( List(
structuredProperty( structuredProperty(
orcid, orcid,
qualifier( ModelConstants.ORCID_PENDING, ModelConstants.ORCID_PENDING, ModelConstants.DNET_PID_TYPES, ModelConstants.DNET_PID_TYPES), qualifier(
ModelConstants.ORCID_PENDING,
ModelConstants.ORCID_PENDING,
ModelConstants.DNET_PID_TYPES,
ModelConstants.DNET_PID_TYPES
),
generateDataInfo() generateDataInfo()
) )
).asJava ).asJava
@ -473,14 +575,13 @@ case object Crossref2Oaf {
null null
} }
def convert(input: String, vocabularies: VocabularyGroup): List[Oaf] = { def convert(input: String, vocabularies: VocabularyGroup): List[CrossrefResult] = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
lazy val json: json4s.JValue = parse(input) lazy val json: json4s.JValue = parse(input)
var resultList: List[Oaf] = List() var resultList: List[CrossrefResult] = List()
val objectType = (json \ "type").extractOrElse[String](null) val objectType = (json \ "type").extractOrElse[String](null)
val objectSubType = (json \ "subtype").extractOrElse[String](null)
if (objectType == null) if (objectType == null)
return resultList return resultList
val typology = getTypeQualifier(objectType, vocabularies) val typology = getTypeQualifier(objectType, vocabularies)
@ -507,7 +608,7 @@ case object Crossref2Oaf {
createCrossrefCollectedFrom(), createCrossrefCollectedFrom(),
result.getDataInfo, result.getDataInfo,
result.getLastupdatetimestamp result.getLastupdatetimestamp
) ).map(s => CrossrefResult(s.getClass.getSimpleName, mapper.writeValueAsString(s)))
} }
result match { result match {
@ -522,10 +623,15 @@ case object Crossref2Oaf {
if (doisReference != null && doisReference.nonEmpty) { if (doisReference != null && doisReference.nonEmpty) {
val citation_relations: List[Relation] = generateCitationRelations(doisReference, result) val citation_relations: List[Relation] = generateCitationRelations(doisReference, result)
resultList = resultList ::: citation_relations resultList = resultList ::: citation_relations.map(s =>
CrossrefResult(s.getClass.getSimpleName, mapper.writeValueAsString(s))
)
} }
resultList = resultList ::: List(result) if (!filterResult(result))
resultList List()
else
resultList ::: List(result).map(s => CrossrefResult(s.getClass.getSimpleName, mapper.writeValueAsString(s)))
} }
private def createCiteRelation(source: Result, targetPid: String, targetPidType: String): List[Relation] = { private def createCiteRelation(source: Result, targetPid: String, targetPidType: String): List[Relation] = {
@ -836,8 +942,7 @@ case object Crossref2Oaf {
val item = new Dataset val item = new Dataset
item.setResourcetype(objectType) item.setResourcetype(objectType)
return item return item
} } else if (objectType.getClassid.equalsIgnoreCase("software")) {
else if (objectType.getClassid.equalsIgnoreCase("software")){
val item = new Software val item = new Software
item.setResourcetype(objectType) item.setResourcetype(objectType)
return item return item

View File

@ -1,106 +0,0 @@
package eu.dnetlib.dhp.collection.crossref
import eu.dnetlib.dhp.application.ArgumentApplicationParser
import org.apache.commons.io.IOUtils
import org.apache.hadoop.io.{IntWritable, Text}
import org.apache.spark.SparkConf
import org.apache.spark.sql.expressions.Aggregator
import org.apache.spark.sql.{Dataset, Encoder, SaveMode, SparkSession}
import org.json4s
import org.json4s.DefaultFormats
import org.json4s.jackson.JsonMethods.parse
import org.slf4j.{Logger, LoggerFactory}
object CrossrefDataset {
val logger: Logger = LoggerFactory.getLogger(CrossrefDataset.getClass)
def to_item(input: String): CrossrefDT = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
lazy val json: json4s.JValue = parse(input)
val ts: Long = (json \ "indexed" \ "timestamp").extract[Long]
val doi: String = DoiBoostMappingUtil.normalizeDoi((json \ "DOI").extract[String])
CrossrefDT(doi, input, ts)
}
def main(args: Array[String]): Unit = {
val conf: SparkConf = new SparkConf()
val parser = new ArgumentApplicationParser(
IOUtils.toString(
CrossrefDataset.getClass.getResourceAsStream(
"/eu/dnetlib/dhp/doiboost/crossref_to_dataset_params.json"
)
)
)
parser.parseArgument(args)
val spark: SparkSession =
SparkSession
.builder()
.config(conf)
.appName(CrossrefDataset.getClass.getSimpleName)
.master(parser.get("master"))
.getOrCreate()
import spark.implicits._
val crossrefAggregator = new Aggregator[CrossrefDT, CrossrefDT, CrossrefDT] with Serializable {
override def zero: CrossrefDT = null
override def reduce(b: CrossrefDT, a: CrossrefDT): CrossrefDT = {
if (b == null)
return a
if (a == null)
return b
if (a.timestamp > b.timestamp) {
return a
}
b
}
override def merge(a: CrossrefDT, b: CrossrefDT): CrossrefDT = {
if (b == null)
return a
if (a == null)
return b
if (a.timestamp > b.timestamp) {
return a
}
b
}
override def bufferEncoder: Encoder[CrossrefDT] = implicitly[Encoder[CrossrefDT]]
override def outputEncoder: Encoder[CrossrefDT] = implicitly[Encoder[CrossrefDT]]
override def finish(reduction: CrossrefDT): CrossrefDT = reduction
}
val workingPath: String = parser.get("workingPath")
val main_ds: Dataset[CrossrefDT] = spark.read.load(s"$workingPath/crossref_ds").as[CrossrefDT]
val update =
spark.createDataset(
spark.sparkContext
.sequenceFile(s"$workingPath/index_update", classOf[IntWritable], classOf[Text])
.map(i => CrossrefImporter.decompressBlob(i._2.toString))
.map(i => to_item(i))
)
main_ds
.union(update)
.groupByKey(_.doi)
.agg(crossrefAggregator.toColumn)
.map(s => s._2)
.write
.mode(SaveMode.Overwrite)
.save(s"$workingPath/crossref_ds_updated")
}
}

View File

@ -0,0 +1,61 @@
package eu.dnetlib.dhp.collection.crossref
import eu.dnetlib.dhp.application.AbstractScalaApplication
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup
import eu.dnetlib.dhp.schema.oaf.{Oaf, Publication, Dataset => OafDataset}
import eu.dnetlib.dhp.utils.ISLookupClientFactory
import org.apache.spark.sql._
import org.slf4j.{Logger, LoggerFactory}
class SparkMapDumpIntoOAF(propertyPath: String, args: Array[String], log: Logger)
extends AbstractScalaApplication(propertyPath, args, log: Logger) {
/** Here all the spark applications runs this method
* where the whole logic of the spark node is defined
*/
override def run(): Unit = {
val sourcePath = parser.get("sourcePath")
log.info("sourcePath: {}", sourcePath)
val targetPath = parser.get("targetPath")
log.info("targetPath: {}", targetPath)
val isLookupUrl: String = parser.get("isLookupUrl")
log.info("isLookupUrl: {}", isLookupUrl)
val isLookupService = ISLookupClientFactory.getLookUpService(isLookupUrl)
val vocabularies = VocabularyGroup.loadVocsFromIS(isLookupService)
require(vocabularies != null)
transformCrossref(spark, sourcePath, targetPath, vocabularies)
}
def transformCrossref(
spark: SparkSession,
sourcePath: String,
targetPath: String,
vocabularies: VocabularyGroup
): Unit = {
import spark.implicits._
val dump = spark.read.text(sourcePath).as[String]
dump
.flatMap(s => Crossref2Oaf.convert(s, vocabularies))
.write
.mode(SaveMode.Overwrite)
.partitionBy("oafType")
.option("compression", "gzip")
.text(targetPath)
}
}
object SparkMapDumpIntoOAF {
def main(args: Array[String]): Unit = {
val logger: Logger = LoggerFactory.getLogger(SparkMapDumpIntoOAF.getClass)
new SparkMapDumpIntoOAF(
log = logger,
args = args,
propertyPath = "/eu/dnetlib/dhp/collection/crossref/convert_crossref_dump_to_oaf_params.json"
).initialize().run()
}
}

View File

@ -14,7 +14,6 @@ import scala.collection.JavaConverters._
case class MAGPaper( case class MAGPaper(
paperId: Option[Long], paperId: Option[Long],
rank: Option[Int],
doi: Option[String], doi: Option[String],
docType: Option[String], docType: Option[String],
paperTitle: Option[String], paperTitle: Option[String],
@ -29,7 +28,6 @@ case class MAGPaper(
journalName: Option[String], journalName: Option[String],
journalIssn: Option[String], journalIssn: Option[String],
journalPublisher: Option[String], journalPublisher: Option[String],
journalWebpage: Option[String],
conferenceSeriesId: Option[Long], conferenceSeriesId: Option[Long],
conferenceInstanceId: Option[Long], conferenceInstanceId: Option[Long],
conferenceName: Option[String], conferenceName: Option[String],
@ -51,8 +49,6 @@ case class MAGPaper(
abstractText: Option[String], abstractText: Option[String],
// List of authors // List of authors
authors: Option[List[MAGAuthor]], authors: Option[List[MAGAuthor]],
// List of Fields of Study
fos: Option[List[MAGFieldOfStudy]],
urls: Option[List[String]] urls: Option[List[String]]
) )
@ -65,13 +61,6 @@ case class MAGAuthor(
GridId: Option[String] GridId: Option[String]
) )
case class MAGFieldOfStudy(
FieldOfStudyId: Option[Long],
DisplayName: Option[String],
MainType: Option[String],
Score: Option[Double]
)
object MagUtility extends Serializable { object MagUtility extends Serializable {
val datatypedict = Map( val datatypedict = Map(

View File

@ -62,9 +62,7 @@ class SparkCreateMagDenormalizedTable(propertyPath: String, args: Array[String],
//next step we create a table containing //next step we create a table containing
val authors = MagUtility.loadMagEntity(spark, "Authors", magBasePath) val authors = MagUtility.loadMagEntity(spark, "Authors", magBasePath)
val affiliations = MagUtility.loadMagEntity(spark, "Affiliations", magBasePath) val affiliations = MagUtility.loadMagEntity(spark, "Affiliations", magBasePath)
val paaf = MagUtility.loadMagEntity(spark, "PaperAuthorAffiliations", magBasePath) val paperAuthorAffiliations = MagUtility.loadMagEntity(spark, "PaperAuthorAffiliations", magBasePath)
val paperAuthorAffiliations = paaf.join(step0, paaf("PaperId") === step0("PaperId"), "leftsemi")
val j1 = paperAuthorAffiliations val j1 = paperAuthorAffiliations
.join(authors, paperAuthorAffiliations("AuthorId") === authors("AuthorId"), "inner") .join(authors, paperAuthorAffiliations("AuthorId") === authors("AuthorId"), "inner")
@ -116,26 +114,26 @@ class SparkCreateMagDenormalizedTable(propertyPath: String, args: Array[String],
step2.count() step2.count()
step1.unpersist() step1.unpersist()
val fos = MagUtility // val fos = MagUtility
.loadMagEntity(spark, "FieldsOfStudy", magBasePath) // .loadMagEntity(spark, "FieldsOfStudy", magBasePath)
.select($"FieldOfStudyId".alias("fos"), $"DisplayName", $"MainType") // .select($"FieldOfStudyId".alias("fos"), $"DisplayName", $"MainType")
//
val paperFieldsOfStudy = MagUtility // val paperFieldsOfStudy = MagUtility
.loadMagEntity(spark, "PaperFieldsOfStudy", magBasePath) // .loadMagEntity(spark, "PaperFieldsOfStudy", magBasePath)
.select($"FieldOfStudyId", $"Score", $"PaperId") // .select($"FieldOfStudyId", $"Score", $"PaperId")
//
val paperFoS = paperFieldsOfStudy // val paperFoS = paperFieldsOfStudy
.join(broadcast(fos), fos("fos") === paperFieldsOfStudy("FieldOfStudyId")) // .join(broadcast(fos), fos("fos") === paperFieldsOfStudy("FieldOfStudyId"))
.groupBy("PaperId") // .groupBy("PaperId")
.agg(collect_set(struct("FieldOfStudyId", "DisplayName", "MainType", "Score")).as("FoS")) // .agg(collect_set(struct("FieldOfStudyId", "DisplayName", "MainType", "Score")).as("FoS"))
//
val step3 = step2 // val step3 = step2
.join(paperFoS, step2("PaperId") === paperFoS("PaperId"), "left") // .join(paperFoS, step2("PaperId") === paperFoS("PaperId"), "left")
.select(step2("*"), paperFoS("FoS")) // .select(step2("*"), paperFoS("FoS"))
.cache() // .cache()
step3.count() // step3.count()
//
step2.unpersist() // step2.unpersist()
val journals = MagUtility val journals = MagUtility
.loadMagEntity(spark, "Journals", magBasePath) .loadMagEntity(spark, "Journals", magBasePath)
@ -143,20 +141,18 @@ class SparkCreateMagDenormalizedTable(propertyPath: String, args: Array[String],
$"JournalId", $"JournalId",
$"DisplayName".as("journalName"), $"DisplayName".as("journalName"),
$"Issn".as("journalIssn"), $"Issn".as("journalIssn"),
$"Publisher".as("journalPublisher"), $"Publisher".as("journalPublisher")
$"Webpage".as("journalWebpage")
) )
val step4 = step3 val step3 = step2
.join(journals, step3("JournalId") === journals("JournalId"), "left") .join(journals, step2("JournalId") === journals("JournalId"), "left")
.select( .select(
step3("*"), step2("*"),
journals("journalName"), journals("journalName"),
journals("journalIssn"), journals("journalIssn"),
journals("journalPublisher"), journals("journalPublisher")
journals("journalWebpage")
) )
.cache .cache
step4.count() step3.count()
val paper_urls = MagUtility val paper_urls = MagUtility
.loadMagEntity(spark, "PaperUrls", magBasePath) .loadMagEntity(spark, "PaperUrls", magBasePath)
@ -166,12 +162,11 @@ class SparkCreateMagDenormalizedTable(propertyPath: String, args: Array[String],
paper_urls.count paper_urls.count
step4 step3
.join(paper_urls, step4("PaperId") === paper_urls("PaperId")) .join(paper_urls, step3("PaperId") === paper_urls("PaperId"))
.select(step4("*"), paper_urls("urls")) .select(step3("*"), paper_urls("urls"))
.select( .select(
$"PaperId".as("paperId"), $"PaperId".as("paperId"),
$"Rank".as("rank"),
$"Doi".as("doi"), $"Doi".as("doi"),
$"DocType".as("docType"), $"DocType".as("docType"),
$"PaperTitle".as("paperTitle"), $"PaperTitle".as("paperTitle"),
@ -202,11 +197,9 @@ class SparkCreateMagDenormalizedTable(propertyPath: String, args: Array[String],
$"conferenceLocation".as("conferenceLocation"), $"conferenceLocation".as("conferenceLocation"),
$"conferenceStartDate".as("conferenceStartDate"), $"conferenceStartDate".as("conferenceStartDate"),
$"conferenceEndDate".as("conferenceEndDate"), $"conferenceEndDate".as("conferenceEndDate"),
$"FoS".as("fos"),
$"journalName".as("journalName"), $"journalName".as("journalName"),
$"journalIssn".as("journalIssn"), $"journalIssn".as("journalIssn"),
$"journalPublisher".as("journalPublisher"), $"journalPublisher".as("journalPublisher"),
$"journalWebpage".as("journalWebpage"),
$"urls" $"urls"
) )
.write .write

View File

@ -0,0 +1,849 @@
{
"URL": "http://dx.doi.org/10.1016/j.joca.2019.11.002",
"resource": {
"primary": {
"URL": "https://linkinghub.elsevier.com/retrieve/pii/S106345841931266X"
}
},
"member": "78",
"score": 0.0,
"created": {
"date-parts": [
[
2019,
11,
14
]
],
"date-time": "2019-11-14T18:08:38Z",
"timestamp": 1573754918000
},
"update-policy": "http://dx.doi.org/10.1016/elsevier_cm_policy",
"license": [
{
"start": {
"date-parts": [
[
2020,
5,
1
]
],
"date-time": "2020-05-01T00:00:00Z",
"timestamp": 1588291200000
},
"content-version": "tdm",
"delay-in-days": 0,
"URL": "https://www.elsevier.com/tdm/userlicense/1.0/"
},
{
"start": {
"date-parts": [
[
2021,
5,
1
]
],
"date-time": "2021-05-01T00:00:00Z",
"timestamp": 1619827200000
},
"content-version": "vor",
"delay-in-days": 365,
"URL": "http://www.elsevier.com/open-access/userlicense/1.0/"
}
],
"ISSN": [
"1063-4584"
],
"container-title": [
"Osteoarthritis and Cartilage"
],
"issued": {
"date-parts": [
[
2020,
5
]
]
},
"issue": "5",
"prefix": "10.1016",
"reference-count": 50,
"indexed": {
"date-parts": [
[
2024,
2,
27
]
],
"date-time": "2024-02-27T00:38:44Z",
"timestamp": 1708994324729
},
"author": [
{
"given": "N.",
"family": "Sharma",
"sequence": "first",
"affiliation": []
},
{
"given": "P.",
"family": "Drobinski",
"sequence": "additional",
"affiliation": []
},
{
"given": "A.",
"family": "Kayed",
"sequence": "additional",
"affiliation": []
},
{
"given": "Z.",
"family": "Chen",
"sequence": "additional",
"affiliation": []
},
{
"given": "C.F.",
"family": "Kjelgaard-Petersen",
"sequence": "additional",
"affiliation": []
},
{
"given": "T.",
"family": "Gantzel",
"sequence": "additional",
"affiliation": []
},
{
"given": "M.A.",
"family": "Karsdal",
"sequence": "additional",
"affiliation": []
},
{
"given": "M.",
"family": "Michaelis",
"sequence": "additional",
"affiliation": []
},
{
"given": "C.",
"family": "Ladel",
"sequence": "additional",
"affiliation": []
},
{
"given": "A.C.",
"family": "Bay-Jensen",
"sequence": "additional",
"affiliation": []
},
{
"given": "S.",
"family": "Lindemann",
"sequence": "additional",
"affiliation": []
},
{
"given": "C.S.",
"family": "Thudium",
"sequence": "additional",
"affiliation": []
}
],
"DOI": "10.1016/j.joca.2019.11.002",
"is-referenced-by-count": 27,
"funder": [
{
"DOI": "10.13039/501100001732",
"name": "Danmarks Grundforskningsfond",
"doi-asserted-by": "publisher"
}
],
"published": {
"date-parts": [
[
2020,
5
]
]
},
"published-print": {
"date-parts": [
[
2020,
5
]
]
},
"alternative-id": [
"S106345841931266X"
],
"subject": [
"Orthopedics and Sports Medicine",
"Biomedical Engineering",
"Rheumatology"
],
"content-domain": {
"domain": [
"clinicalkey.fr",
"clinicalkey.jp",
"clinicalkey.es",
"clinicalkey.com.au",
"oarsijournal.com",
"clinicalkey.com",
"elsevier.com",
"sciencedirect.com"
],
"crossmark-restriction": true
},
"reference": [
{
"key": "10.1016/j.joca.2019.11.002_bib1",
"series-title": "Priority Medicines for Europe and the World. 2013 Update",
"author": "Kaplan",
"year": "2013"
},
{
"issue": "August 2016",
"key": "10.1016/j.joca.2019.11.002_bib2",
"article-title": "Osteoarthritis: toward a comprehensive understanding of pathological mechanism",
"volume": "5",
"author": "Chen",
"year": "2017",
"journal-title": "Bone Res"
},
{
"doi-asserted-by": "crossref",
"first-page": "77",
"issue": "2",
"author": "Sokolove",
"key": "10.1016/j.joca.2019.11.002_bib3",
"DOI": "10.1177/1759720X12467868",
"article-title": "Role of inflammation in the pathogenesis of osteoarthritis: latest findings and interpretations",
"year": "2013",
"volume": "5",
"journal-title": "Ther Adv Musculoskelet Dis"
},
{
"doi-asserted-by": "crossref",
"issue": "5",
"author": "Alberton",
"key": "10.1016/j.joca.2019.11.002_bib4",
"DOI": "10.3390/ijms20051008",
"article-title": "Aggrecan hypomorphism compromises articular cartilage biomechanical properties and is associated with increased incidence of spontaneous osteoarthritis",
"year": "2019",
"volume": "20",
"journal-title": "Int J Mol Sci"
},
{
"doi-asserted-by": "crossref",
"issue": "6",
"author": "Miller",
"key": "10.1016/j.joca.2019.11.002_bib5",
"DOI": "10.1172/jci.insight.95704",
"article-title": "An aggrecan fragment drives osteoarthritis pain through Toll-like receptor 2",
"year": "2018",
"volume": "3",
"journal-title": "JCI Insight"
},
{
"doi-asserted-by": "crossref",
"first-page": "1240",
"issue": "5",
"author": "Lees",
"key": "10.1016/j.joca.2019.11.002_bib6",
"DOI": "10.1002/art.39063",
"article-title": "Bioactivity in an aggrecan 32-mer fragment is mediated via toll-like receptor 2",
"year": "2015",
"volume": "67",
"journal-title": "Arthritis Rheum"
},
{
"key": "10.1016/j.joca.2019.11.002_bib7",
"article-title": "Increased function of pronociceptive TRPV1 at the level of the joint in a rat model of osteoarthritis pain",
"author": "Kelly",
"year": "2013",
"journal-title": "Ann Rheum Dis"
},
{
"doi-asserted-by": "crossref",
"first-page": "580",
"issue": "10",
"author": "Robinson",
"key": "10.1016/j.joca.2019.11.002_bib8",
"DOI": "10.1038/nrrheum.2016.136",
"article-title": "Low-grade inflammation as a key mediator of the pathogenesis of osteoarthritis",
"year": "2016",
"volume": "12",
"journal-title": "Nat Rev Rheumatol"
},
{
"doi-asserted-by": "crossref",
"first-page": "625",
"issue": "11",
"author": "Sellam",
"key": "10.1016/j.joca.2019.11.002_bib9",
"DOI": "10.1038/nrrheum.2010.159",
"article-title": "The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis",
"year": "2010",
"volume": "6",
"journal-title": "Nat Rev Rheumatol"
},
{
"doi-asserted-by": "crossref",
"first-page": "249",
"issue": "2",
"author": "Scanzello",
"key": "10.1016/j.joca.2019.11.002_bib10",
"DOI": "10.1016/j.bone.2012.02.012",
"article-title": "The role of synovitis in osteoarthritis pathogenesis",
"year": "2012",
"volume": "51",
"journal-title": "Bone"
},
{
"doi-asserted-by": "crossref",
"first-page": "33",
"issue": "1",
"author": "Kapoor",
"key": "10.1016/j.joca.2019.11.002_bib11",
"DOI": "10.1038/nrrheum.2010.196",
"article-title": "Role of proinflammatory cytokines in the pathophysiology of osteoarthritis",
"year": "2011",
"volume": "7",
"journal-title": "Nat Rev Rheumatol"
},
{
"issue": "4",
"key": "10.1016/j.joca.2019.11.002_bib12",
"first-page": "1",
"volume": "18",
"author": "Roh",
"year": "2018",
"journal-title": "Origin and List of Damps"
},
{
"key": "10.1016/j.joca.2019.11.002_bib13",
"doi-asserted-by": "crossref",
"article-title": "Overexpression of toll-like receptors 3 and 4 in synovial tissue from patients with early rheumatoid arthritis: toll-like receptor expression in early and longstanding arthritis",
"author": "Ospelt",
"year": "2008",
"journal-title": "Arthritis Rheum",
"DOI": "10.1002/art.24140"
},
{
"doi-asserted-by": "crossref",
"first-page": "338",
"issue": "2",
"author": "Roelofs",
"key": "10.1016/j.joca.2019.11.002_bib14",
"DOI": "10.1002/art.23217",
"article-title": "The orchestra of toll-like receptors and their potential role in frequently occurring rheumatic conditions",
"year": "2008",
"volume": "58",
"journal-title": "Arthritis Rheum"
},
{
"doi-asserted-by": "crossref",
"first-page": "657",
"issue": "3",
"author": "Gondokaryono",
"key": "10.1016/j.joca.2019.11.002_bib15",
"DOI": "10.1189/jlb.1206730",
"article-title": "The extra domain A of fibronectin stimulates murine mast cells via Toll-like receptor 4",
"year": "2007",
"volume": "82",
"journal-title": "J Leukoc Biol"
},
{
"issue": "7",
"key": "10.1016/j.joca.2019.11.002_bib16",
"first-page": "2004",
"article-title": "Chondrocyte innate immune myeloid differentiation factor 88-dependent signaling drives procatabolic effects of the endogenous toll-like receptor 2/toll-like receptor 4 ligands low molecular weight hyaluronan and high mobility group box chromosomal protein",
"volume": "62",
"author": "Liu-Bryan",
"year": "2010",
"journal-title": "Arthritis Rheum"
},
{
"doi-asserted-by": "crossref",
"first-page": "774",
"issue": "7",
"author": "Midwood",
"key": "10.1016/j.joca.2019.11.002_bib17",
"DOI": "10.1038/nm.1987",
"article-title": "Tenascin-C is an endogenous activator of Toll-like receptor 4 that is essential for maintaining inflammation in arthritic joint disease",
"year": "2009",
"volume": "15",
"journal-title": "Nat Med"
},
{
"doi-asserted-by": "crossref",
"first-page": "16",
"author": "Berenbaum",
"key": "10.1016/j.joca.2019.11.002_bib18",
"DOI": "10.1016/j.joca.2012.11.012",
"article-title": "Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!)",
"year": "2013",
"volume": "21",
"journal-title": "Osteoarthr Cartil"
},
{
"doi-asserted-by": "crossref",
"first-page": "259",
"issue": "4",
"author": "Wang",
"key": "10.1016/j.joca.2019.11.002_bib19",
"DOI": "10.1186/1756-0500-2-259",
"article-title": "Suppression of MMP activity in bovine cartilage explants cultures has little if any effect on the release of aggrecanase-derived aggrecan fragments",
"year": "2009",
"volume": "2",
"journal-title": "BMC Res Notes"
},
{
"doi-asserted-by": "crossref",
"first-page": "423",
"issue": "56",
"author": "Bay-Jensen",
"key": "10.1016/j.joca.2019.11.002_bib20",
"DOI": "10.1016/j.clinbiochem.2011.01.001",
"article-title": "Enzyme-linked immunosorbent assay (ELISAs) for metalloproteinase derived type II collagen neoepitope, CIIM-Increased serum CIIM in subjects with severe radiographic osteoarthritis",
"year": "2011",
"volume": "44",
"journal-title": "Clin Biochem"
},
{
"doi-asserted-by": "crossref",
"first-page": "18789",
"issue": "10",
"author": "Gudmann",
"key": "10.1016/j.joca.2019.11.002_bib21",
"DOI": "10.3390/ijms151018789",
"article-title": "Cartilage turnover reflected by metabolic processing of type II collagen: a novel marker of anabolic function in chondrocytes",
"year": "2014",
"volume": "15",
"journal-title": "Int J Mol Sci"
},
{
"doi-asserted-by": "crossref",
"first-page": "899",
"issue": "1011",
"author": "Barascuk",
"key": "10.1016/j.joca.2019.11.002_bib22",
"DOI": "10.1016/j.clinbiochem.2010.03.012",
"article-title": "A novel assay for extracellular matrix remodeling associated with liver fibrosis: an enzyme-linked immunosorbent assay (ELISA) for a MMP-9 proteolytically revealed neo-epitope of type III collagen",
"year": "2010",
"volume": "43",
"journal-title": "Clin Biochem"
},
{
"doi-asserted-by": "crossref",
"first-page": "93",
"author": "Sun",
"key": "10.1016/j.joca.2019.11.002_bib23",
"DOI": "10.1186/1471-2474-15-93",
"article-title": "The active form of MMP-3 is a marker of synovial inflammation and cartilage turnover in inflammatory joint diseases",
"year": "2014",
"volume": "15",
"journal-title": "BMC Muscoskelet Disord"
},
{
"doi-asserted-by": "crossref",
"first-page": "103",
"issue": "1",
"author": "Chandrasekhar",
"key": "10.1016/j.joca.2019.11.002_bib24",
"DOI": "10.1016/0003-2697(87)90658-0",
"article-title": "Microdetermination of proteoglycans and glycosaminoglycans in the presence of guanidine hydrochloride",
"year": "1987",
"volume": "161",
"journal-title": "Anal Biochem"
},
{
"doi-asserted-by": "crossref",
"first-page": "8722",
"issue": "22",
"author": "Hankins",
"key": "10.1016/j.joca.2019.11.002_bib25",
"DOI": "10.1073/pnas.1201313109",
"article-title": "Amino acid addition to Vibrio cholerae LPS establishes a link between surface remodeling in Gram-positive and Gram-negative bacteria",
"year": "2012",
"volume": "109",
"journal-title": "Proc Natl Acad Sci"
},
{
"issue": "c",
"key": "10.1016/j.joca.2019.11.002_bib26",
"first-page": "1",
"volume": "2",
"author": "Bailie",
"year": "2010",
"journal-title": "Supporting Information"
},
{
"doi-asserted-by": "crossref",
"first-page": "1145",
"issue": "9",
"author": "Neogi",
"key": "10.1016/j.joca.2019.11.002_bib27",
"DOI": "10.1016/j.joca.2013.03.018",
"article-title": "The epidemiology and impact of pain in osteoarthritis",
"year": "2013",
"volume": "21",
"journal-title": "Osteoarthr Cartil"
},
{
"doi-asserted-by": "crossref",
"first-page": "623",
"issue": "3",
"author": "Hunter",
"key": "10.1016/j.joca.2019.11.002_bib28",
"DOI": "10.1016/j.rdc.2008.05.004",
"article-title": "The symptoms of osteoarthritis and the genesis of pain",
"year": "2008",
"volume": "34",
"journal-title": "Rheum Dis Clin N Am"
},
{
"doi-asserted-by": "crossref",
"first-page": "1326",
"issue": "11",
"author": "Xu",
"key": "10.1016/j.joca.2019.11.002_bib29",
"DOI": "10.1038/nm.3978",
"article-title": "Inhibition of mechanical allodynia in neuropathic pain by TLR5-mediated A-fiber blockade",
"year": "2015",
"volume": "21",
"journal-title": "Nat Med"
},
{
"doi-asserted-by": "crossref",
"first-page": "145",
"issue": "October 2017",
"author": "Lacagnina",
"key": "10.1016/j.joca.2019.11.002_bib30",
"DOI": "10.1016/j.pharmthera.2017.10.006",
"article-title": "Toll-like receptors and their role in persistent pain",
"year": "2018",
"volume": "184",
"journal-title": "Pharmacol Ther"
},
{
"key": "10.1016/j.joca.2019.11.002_bib31",
"first-page": "2016",
"article-title": "Blockade of toll-like receptors (TLR2, TLR4) attenuates pain and potentiates buprenorphine analgesia in a rat neuropathic pain model",
"author": "Jurga",
"year": "2016",
"journal-title": "Neural Plast"
},
{
"doi-asserted-by": "crossref",
"first-page": "357",
"issue": "5",
"author": "Huang",
"key": "10.1016/j.joca.2019.11.002_bib32",
"DOI": "10.1007/s11926-009-0051-z",
"article-title": "The role of toll-like receptors in rheumatoid arthritis",
"year": "2009",
"volume": "11",
"journal-title": "Curr Rheumatol Rep"
},
{
"issue": "3",
"key": "10.1016/j.joca.2019.11.002_bib33",
"first-page": "1",
"article-title": "TLR4 signalling in osteoarthritis-finding targets for candidate DMOADs",
"volume": "11",
"author": "Gómez",
"year": "2014",
"journal-title": "Nat Rev Rheumatol"
},
{
"issue": "5 Pt 1",
"key": "10.1016/j.joca.2019.11.002_bib34",
"first-page": "1432",
"article-title": "Procollagen types I and III aminoterminal propeptide levels during acute respiratory distress syndrome and in response to methylprednisolone treatment.[Erratum appears in Am J Respir Crit Care Med. 2013 Dec 15;188(12):1477]",
"volume": "158",
"author": "Meduri",
"year": "1998",
"journal-title": "Am J Respir Crit Care Med"
},
{
"doi-asserted-by": "crossref",
"first-page": "5",
"issue": "1",
"author": "Scheja",
"key": "10.1016/j.joca.2019.11.002_bib35",
"DOI": "10.3109/03009749209095054",
"article-title": "Serum levels of aminoterminal type III procollagen peptide and hyaluronan predict mortality in systemic sclerosis",
"year": "1992",
"volume": "21",
"journal-title": "Scand J Rheumatol"
},
{
"doi-asserted-by": "crossref",
"first-page": "547",
"issue": "8",
"author": "Kjelgaard-Petersen",
"key": "10.1016/j.joca.2019.11.002_bib36",
"DOI": "10.3109/1354750X.2015.1105497",
"article-title": "Synovitis biomarkers: ex vivo characterization of three biomarkers for identification of inflammatory osteoarthritis",
"year": "2015",
"volume": "20",
"journal-title": "Biomarkers"
},
{
"key": "10.1016/j.joca.2019.11.002_bib37",
"article-title": "Sensitization and serological biomarkers in knee osteoarthritis patients with different degrees of synovitis",
"author": "Petersen",
"year": "2015",
"journal-title": "Clin J Pain"
},
{
"doi-asserted-by": "crossref",
"first-page": "309",
"author": "He",
"key": "10.1016/j.joca.2019.11.002_bib38",
"DOI": "10.1186/1471-2474-15-309",
"article-title": "Type X collagen levels are elevated in serum from human osteoarthritis patients and associated with biomarkers of cartilage degradation and inflammation",
"year": "2014",
"volume": "15",
"journal-title": "BMC Muscoskelet Disord"
},
{
"doi-asserted-by": "crossref",
"first-page": "1",
"issue": "1",
"author": "Bay-Jensen",
"key": "10.1016/j.joca.2019.11.002_bib39",
"DOI": "10.1371/journal.pone.0054504",
"article-title": "Circulating protein fragments of cartilage and connective tissue degradation are diagnostic and prognostic markers of rheumatoid arthritis and ankylosing spondylitis",
"year": "2013",
"volume": "8",
"journal-title": "PLoS One"
},
{
"doi-asserted-by": "crossref",
"first-page": "44",
"issue": "1",
"author": "Siebuhr",
"key": "10.1016/j.joca.2019.11.002_bib40",
"DOI": "10.1016/j.joca.2013.10.020",
"article-title": "Identification and characterisation of osteoarthritis patients with inflammation derived tissue turnover",
"year": "2014",
"volume": "22",
"journal-title": "Osteoarthr Cartil"
},
{
"doi-asserted-by": "crossref",
"first-page": "221",
"issue": "3",
"author": "Yuji Yoshida",
"key": "10.1016/j.joca.2019.11.002_bib41",
"DOI": "10.1016/0306-9877(88)90147-8",
"article-title": "Interleukin 6 and rheumatoid arthritis",
"year": "1988",
"volume": "27",
"journal-title": "Med Hypotheses"
},
{
"doi-asserted-by": "crossref",
"first-page": "5173",
"issue": "22",
"author": "Suthaus",
"key": "10.1016/j.joca.2019.11.002_bib42",
"DOI": "10.1182/blood-2011-09-377705",
"article-title": "HHV8 encoded viral IL-6 collaborates with mouse IL-6 in MCD-like development in mice",
"year": "2012",
"volume": "119",
"journal-title": "Blood"
},
{
"doi-asserted-by": "crossref",
"first-page": "1",
"issue": "1",
"author": "Zhou",
"key": "10.1016/j.joca.2019.11.002_bib43",
"DOI": "10.1186/s12974-016-0607-6",
"article-title": "Interleukin-6: an emerging regulator of pathological pain",
"year": "2016",
"volume": "13",
"journal-title": "J Neuroinflammation"
},
{
"key": "10.1016/j.joca.2019.11.002_bib44",
"series-title": "Local Translation and Retrograde Axonal Transport of Creb Regulates Il-6-Induced Nociceptive Plasticity",
"first-page": "1",
"author": "Melemedjian",
"year": "2014"
},
{
"issue": "6",
"key": "10.1016/j.joca.2019.11.002_bib45",
"first-page": "1149",
"article-title": "Aggrecanase and Aggrecan degradation in osteoarthritis: a review",
"volume": "36",
"author": "Huang",
"year": "2008",
"journal-title": "J Int Med Res"
},
{
"doi-asserted-by": "crossref",
"first-page": "133",
"issue": "1",
"author": "Troeberg",
"key": "10.1016/j.joca.2019.11.002_bib46",
"DOI": "10.1016/j.bbapap.2011.06.020",
"article-title": "Proteases involved in cartilage matrix degradation in osteoarthritis",
"year": "2012",
"volume": "1824",
"journal-title": "Biochim Biophys Acta Protein Proteonomics"
},
{
"doi-asserted-by": "crossref",
"first-page": "1633",
"issue": "11",
"author": "Zhang",
"key": "10.1016/j.joca.2019.11.002_bib47",
"DOI": "10.1136/ard.2007.079574",
"article-title": "Differential Toll-like receptor-dependent collagenase expression in chondrocytes",
"year": "2008",
"volume": "67",
"journal-title": "Ann Rheum Dis"
},
{
"doi-asserted-by": "crossref",
"first-page": "2152",
"issue": "7",
"author": "Kim",
"key": "10.1016/j.joca.2019.11.002_bib48",
"DOI": "10.1002/art.21951",
"article-title": "The catabolic pathway mediated by toll-like receptors in human osteoarthritic chondrocytes",
"year": "2006",
"volume": "54",
"journal-title": "Arthritis Rheum"
},
{
"issue": "2018",
"key": "10.1016/j.joca.2019.11.002_bib49",
"article-title": "The anti-ADAMTS-5 nanobody®, M6495, protects against cartilage breakdown in cartilage and synovial joint tissue explant models",
"volume": "26",
"author": "Siebuhr",
"year": "2018",
"journal-title": "Osteoarthr Cartil"
},
{
"doi-asserted-by": "crossref",
"first-page": "2933",
"issue": "11",
"author": "Miller",
"key": "10.1016/j.joca.2019.11.002_bib50",
"DOI": "10.1002/art.39291",
"article-title": "Damage-associated molecular patterns generated in osteoarthritis directly excite murine nociceptive neurons through toll-like receptor 4",
"year": "2015",
"volume": "67",
"journal-title": "Arthritis Rheum"
}
],
"title": [
"Inflammation and joint destruction may be linked to the generation of cartilage metabolites of ADAMTS-5 through activation of toll-like receptors"
],
"link": [
{
"URL": "https://api.elsevier.com/content/article/PII:S106345841931266X?httpAccept=text/xml",
"content-type": "text/xml",
"content-version": "vor",
"intended-application": "text-mining"
},
{
"URL": "https://api.elsevier.com/content/article/PII:S106345841931266X?httpAccept=text/plain",
"content-type": "text/plain",
"content-version": "vor",
"intended-application": "text-mining"
}
],
"source": "Crossref",
"type": "journal-article",
"publisher": "Elsevier BV",
"journal-issue": {
"issue": "5",
"published-print": {
"date-parts": [
[
2020,
5
]
]
}
},
"volume": "28",
"references-count": 50,
"issn-type": [
{
"value": "2227-9717",
"type": "electronic"
},
{
"value": "VALUE",
"type": "PIPPO"
},
{
"value": "1063-4584",
"type": "pu"
}
],
"assertion": [
{
"value": "Elsevier",
"name": "publisher",
"label": "This article is maintained by"
},
{
"value": "Inflammation and joint destruction may be linked to the generation of cartilage metabolites of ADAMTS-5 through activation of toll-like receptors",
"name": "articletitle",
"label": "Article Title"
},
{
"value": "Osteoarthritis and Cartilage",
"name": "journaltitle",
"label": "Journal Title"
},
{
"value": "https://doi.org/10.1016/j.joca.2019.11.002",
"name": "articlelink",
"label": "CrossRef DOI link to publisher maintained version"
},
{
"value": "article",
"name": "content_type",
"label": "Content Type"
},
{
"value": "© 2019 Osteoarthritis Research Society International. Published by Elsevier Ltd.",
"name": "copyright",
"label": "Copyright"
}
],
"deposited": {
"date-parts": [
[
2022,
7,
9
]
],
"date-time": "2022-07-09T14:08:02Z",
"timestamp": 1657375682000
},
"language": "en",
"page": "658-668",
"short-container-title": [
"Osteoarthritis and Cartilage"
]
}

View File

@ -0,0 +1,41 @@
package eu.dnetlib.dhp.collection.crossref
import com.fasterxml.jackson.databind.ObjectMapper
import eu.dnetlib.dhp.aggregation.AbstractVocabularyTest
import org.apache.spark.sql.SparkSession
import org.junit.jupiter.api.{BeforeEach, Test}
import org.junit.jupiter.api.extension.ExtendWith
import org.mockito.junit.jupiter.MockitoExtension
import org.slf4j.{Logger, LoggerFactory}
import scala.io.Source
@ExtendWith(Array(classOf[MockitoExtension]))
class CrossrefMappingTest extends AbstractVocabularyTest {
val logger: Logger = LoggerFactory.getLogger(Crossref2Oaf.getClass)
val mapper = new ObjectMapper()
@BeforeEach
def setUp(): Unit = {
super.setUpVocabulary()
}
@Test
def testMapping(): Unit = {
val spark = SparkSession.builder().master("local[*]").appName("TransformCrossref").getOrCreate()
val s = new SparkMapDumpIntoOAF(null, null, null)
import spark.implicits._
s.transformCrossref(
spark,
sourcePath = "/home/sandro/Downloads/crossref",
targetPath = "/home/sandro/Downloads/crossref_transformed",
vocabularies = vocabularies
)
print(spark.read.text("/home/sandro/Downloads/crossref_transformed").count)
}
}

View File

@ -55,8 +55,6 @@ case object Crossref2Oaf {
.map(f => f.id) .map(f => f.id)
} }
def mappingResult(result: Result, json: JValue, instanceType: Qualifier, originalType: String): Result = { def mappingResult(result: Result, json: JValue, instanceType: Qualifier, originalType: String): Result = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
@ -349,7 +347,6 @@ case object Crossref2Oaf {
if (result == null) if (result == null)
return List() return List()
mappingResult(result, json, typology._1, typology._3) mappingResult(result, json, typology._1, typology._3)
if (result == null || result.getId == null) if (result == null || result.getId == null)
return List() return List()
@ -693,8 +690,7 @@ case object Crossref2Oaf {
val item = new Dataset val item = new Dataset
item.setResourcetype(objectType) item.setResourcetype(objectType)
return item return item
} } else if (objectType.getClassid.equalsIgnoreCase("software")) {
else if (objectType.getClassid.equalsIgnoreCase("software")){
val item = new Software val item = new Software
item.setResourcetype(objectType) item.setResourcetype(objectType)
return item return item

View File

@ -48,7 +48,6 @@ object SparkMapDumpIntoOAF {
val vocabularies = VocabularyGroup.loadVocsFromIS(isLookupService) val vocabularies = VocabularyGroup.loadVocsFromIS(isLookupService)
require(vocabularies != null) require(vocabularies != null)
spark.read spark.read
.load(parser.get("sourcePath")) .load(parser.get("sourcePath"))
.as[CrossrefDT] .as[CrossrefDT]

View File

@ -47,5 +47,4 @@ public abstract class AbstractVocabularyTest {
AbstractVocabularyTest.class.getResourceAsStream("/eu/dnetlib/dhp/doiboost/synonyms.txt"))); AbstractVocabularyTest.class.getResourceAsStream("/eu/dnetlib/dhp/doiboost/synonyms.txt")));
} }
} }

View File

@ -35,9 +35,14 @@ class CrossrefMappingTest extends AbstractVocabularyTest{
@Test @Test
def testMissingAuthorParser(): Unit = { def testMissingAuthorParser(): Unit = {
val json: String = Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/doiboost/crossref/s41567-022-01757-y.json")).mkString val json: String = Source
.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/doiboost/crossref/s41567-022-01757-y.json"))
.mkString
val result = Crossref2Oaf.convert(json, vocabularies) val result = Crossref2Oaf.convert(json, vocabularies)
result.filter(o => o.isInstanceOf[Publication]).map(p=> p.asInstanceOf[Publication]).foreach(p =>assertTrue(p.getAuthor.size()>0)) result
.filter(o => o.isInstanceOf[Publication])
.map(p => p.asInstanceOf[Publication])
.foreach(p => assertTrue(p.getAuthor.size() > 0))
} }
@Test @Test
@ -127,7 +132,6 @@ class CrossrefMappingTest extends AbstractVocabularyTest{
val items = resultList.filter(p => p.isInstanceOf[Result]) val items = resultList.filter(p => p.isInstanceOf[Result])
items.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p))) items.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p)))
} }
@ -169,7 +173,6 @@ class CrossrefMappingTest extends AbstractVocabularyTest{
assertEquals(doisReference.size, relationList.size) assertEquals(doisReference.size, relationList.size)
relationList.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p))) relationList.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p)))
} }
@ -190,7 +193,6 @@ class CrossrefMappingTest extends AbstractVocabularyTest{
val items = resultList.filter(p => p.isInstanceOf[Result]) val items = resultList.filter(p => p.isInstanceOf[Result])
items.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p))) items.foreach(p => println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(p)))
} }
@ -646,7 +648,6 @@ class CrossrefMappingTest extends AbstractVocabularyTest{
val item: Result = resultList.filter(p => p.isInstanceOf[Result]).head.asInstanceOf[Result] val item: Result = resultList.filter(p => p.isInstanceOf[Result]).head.asInstanceOf[Result]
println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(item)) println(mapper.writerWithDefaultPrettyPrinter().writeValueAsString(item))
assertTrue( assertTrue(