fix formtat problem

This commit is contained in:
Enrico Ottonello 2020-04-23 15:25:39 +02:00
parent bb6c9785b4
commit 4a6aea1a37
7 changed files with 1100 additions and 978 deletions

View File

@ -1,6 +1,8 @@
package eu.dnetlib.doiboost.crossref;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import java.io.ByteArrayOutputStream;
import java.util.zip.Inflater;
import org.apache.commons.codec.binary.Base64;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
@ -12,30 +14,29 @@ import org.apache.hadoop.io.Text;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.io.ByteArrayOutputStream;
import java.util.zip.Inflater;
public class CrossrefImporter {
public static void main(String[] args) throws Exception {
final ArgumentApplicationParser parser = new ArgumentApplicationParser(IOUtils.toString(CrossrefImporter.class.getResourceAsStream("/eu/dnetlib/dhp/doiboost/import_from_es.json")));
final ArgumentApplicationParser parser =
new ArgumentApplicationParser(
IOUtils.toString(
CrossrefImporter.class.getResourceAsStream(
"/eu/dnetlib/dhp/doiboost/import_from_es.json")));
Logger logger = LoggerFactory.getLogger(CrossrefImporter.class);
parser.parseArgument(args);
final String hdfsuri = parser.get("namenode");
logger.info("HDFS URI"+hdfsuri);
logger.info("HDFS URI" + hdfsuri);
Path hdfswritepath = new Path(parser.get("targetPath"));
logger.info("TargetPath: "+hdfsuri);
logger.info("TargetPath: " + hdfsuri);
final Long timestamp = StringUtils.isNotBlank(parser.get("timestamp"))?Long.parseLong(parser.get("timestamp")):-1;
if(timestamp>0)
logger.info("Timestamp added "+timestamp);
final Long timestamp =
StringUtils.isNotBlank(parser.get("timestamp"))
? Long.parseLong(parser.get("timestamp"))
: -1;
if (timestamp > 0) logger.info("Timestamp added " + timestamp);
// ====== Init HDFS File System Object
Configuration conf = new Configuration();
@ -45,16 +46,21 @@ public class CrossrefImporter {
conf.set("fs.hdfs.impl", org.apache.hadoop.hdfs.DistributedFileSystem.class.getName());
conf.set("fs.file.impl", org.apache.hadoop.fs.LocalFileSystem.class.getName());
ESClient client =
timestamp > 0
? new ESClient(
"ip-90-147-167-25.ct1.garrservices.it", "crossref", timestamp)
: new ESClient("ip-90-147-167-25.ct1.garrservices.it", "crossref");
ESClient client = timestamp>0?new ESClient("ip-90-147-167-25.ct1.garrservices.it", "crossref", timestamp):new ESClient("ip-90-147-167-25.ct1.garrservices.it", "crossref");
try (SequenceFile.Writer writer = SequenceFile.createWriter(conf,
SequenceFile.Writer.file(hdfswritepath), SequenceFile.Writer.keyClass(IntWritable.class),
SequenceFile.Writer.valueClass(Text.class))) {
try (SequenceFile.Writer writer =
SequenceFile.createWriter(
conf,
SequenceFile.Writer.file(hdfswritepath),
SequenceFile.Writer.keyClass(IntWritable.class),
SequenceFile.Writer.valueClass(Text.class))) {
int i = 0;
long start= System.currentTimeMillis();
long start = System.currentTimeMillis();
long end = 0;
final IntWritable key = new IntWritable(i);
final Text value = new Text();
@ -65,7 +71,10 @@ public class CrossrefImporter {
if (i % 1000000 == 0) {
end = System.currentTimeMillis();
final float time = (end - start) / 1000.0F;
logger.info(String.format("Imported %d records last 100000 imported in %f seconds", i, time));
logger.info(
String.format(
"Imported %d records last 100000 imported in %f seconds",
i, time));
start = System.currentTimeMillis();
}
}
@ -87,7 +96,7 @@ public class CrossrefImporter {
decompresser.end();
return new String(unzippeddata);
} catch (Throwable e) {
throw new RuntimeException("Wrong record:" + blob,e);
throw new RuntimeException("Wrong record:" + blob, e);
}
}
}

View File

@ -1,6 +1,9 @@
package eu.dnetlib.doiboost.crossref;
import com.jayway.jsonpath.JsonPath;
import java.io.IOException;
import java.util.Iterator;
import java.util.List;
import org.apache.commons.io.IOUtils;
import org.apache.http.client.methods.CloseableHttpResponse;
import org.apache.http.client.methods.HttpPost;
@ -10,18 +13,15 @@ import org.apache.http.impl.client.HttpClients;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.io.IOException;
import java.util.Iterator;
import java.util.List;
public class ESClient implements Iterator<String> {
private final static Logger logger = LoggerFactory.getLogger(ESClient.class);
private static final Logger logger = LoggerFactory.getLogger(ESClient.class);
final static String blobPath = "$.hits[*].hits[*]._source.blob";
final static String scrollIdPath = "$._scroll_id";
final static String JSON_NO_TS ="{\"size\":1000}";
final static String JSON_WITH_TS ="{\"size\":1000, \"query\":{\"range\":{\"timestamp\":{\"gte\":%d}}}}";
final static String JSON_SCROLL = "{\"scroll_id\":\"%s\",\"scroll\" : \"1m\"}";
static final String blobPath = "$.hits[*].hits[*]._source.blob";
static final String scrollIdPath = "$._scroll_id";
static final String JSON_NO_TS = "{\"size\":1000}";
static final String JSON_WITH_TS =
"{\"size\":1000, \"query\":{\"range\":{\"timestamp\":{\"gte\":%d}}}}";
static final String JSON_SCROLL = "{\"scroll_id\":\"%s\",\"scroll\" : \"1m\"}";
private final String scrollId;
@ -29,29 +29,34 @@ public class ESClient implements Iterator<String> {
private final String esHost;
public ESClient(final String esHost, final String esIndex) throws IOException {
this.esHost = esHost;
final String body =getResponse(String.format("http://%s:9200/%s/_search?scroll=1m", esHost, esIndex), JSON_NO_TS);
scrollId= getJPathString(scrollIdPath, body);
final String body =
getResponse(
String.format("http://%s:9200/%s/_search?scroll=1m", esHost, esIndex),
JSON_NO_TS);
scrollId = getJPathString(scrollIdPath, body);
buffer = getBlobs(body);
}
public ESClient(final String esHost, final String esIndex, final long timestamp) throws IOException {
public ESClient(final String esHost, final String esIndex, final long timestamp)
throws IOException {
this.esHost = esHost;
final String body =getResponse(String.format("http://%s:9200/%s/_search?scroll=1m", esHost, esIndex), String.format(JSON_WITH_TS, timestamp));
scrollId= getJPathString(scrollIdPath, body);
final String body =
getResponse(
String.format("http://%s:9200/%s/_search?scroll=1m", esHost, esIndex),
String.format(JSON_WITH_TS, timestamp));
scrollId = getJPathString(scrollIdPath, body);
buffer = getBlobs(body);
}
private String getResponse(final String url,final String json ) {
private String getResponse(final String url, final String json) {
CloseableHttpClient client = HttpClients.createDefault();
try {
HttpPost httpPost = new HttpPost(url);
if (json!= null) {
if (json != null) {
StringEntity entity = new StringEntity(json);
httpPost.setEntity(entity);
httpPost.setHeader("Accept", "application/json");
@ -61,22 +66,20 @@ public class ESClient implements Iterator<String> {
return IOUtils.toString(response.getEntity().getContent());
} catch (Throwable e) {
throw new RuntimeException("Error on executing request ",e);
throw new RuntimeException("Error on executing request ", e);
} finally {
try {
client.close();
} catch (IOException e) {
throw new RuntimeException("Unable to close client ",e);
throw new RuntimeException("Unable to close client ", e);
}
}
}
private String getJPathString(final String jsonPath, final String json) {
private String getJPathString(final String jsonPath, final String json) {
try {
Object o = JsonPath.read(json, jsonPath);
if (o instanceof String)
return (String) o;
if (o instanceof String) return (String) o;
return null;
} catch (Exception e) {
return "";
@ -84,14 +87,13 @@ public class ESClient implements Iterator<String> {
}
private List<String> getBlobs(final String body) {
final List<String > res = JsonPath.read(body, "$.hits.hits[*]._source.blob");
final List<String> res = JsonPath.read(body, "$.hits.hits[*]._source.blob");
return res;
}
@Override
public boolean hasNext() {
return (buffer!= null && !buffer.isEmpty());
return (buffer != null && !buffer.isEmpty());
}
@Override
@ -100,11 +102,12 @@ public class ESClient implements Iterator<String> {
if (buffer.isEmpty()) {
final String json_param = String.format(JSON_SCROLL, scrollId);
final String body =getResponse(String.format("http://%s:9200/_search/scroll", esHost), json_param);
final String body =
getResponse(String.format("http://%s:9200/_search/scroll", esHost), json_param);
try {
buffer = getBlobs(body);
} catch (Throwable e) {
logger.error("Error on get next page: body:"+body);
logger.error("Error on get next page: body:" + body);
}
}
return nextItem;

View File

@ -1,20 +1,18 @@
package eu.dnetlib.doiboost.orcid.json;
import com.google.gson.JsonObject;
import eu.dnetlib.doiboost.orcid.model.AuthorData;
public class JsonWriter {
public static String create(AuthorData authorData) {
JsonObject author = new JsonObject();
author.addProperty("oid", authorData.getOid());
author.addProperty("name", authorData.getName());
author.addProperty("surname", authorData.getSurname());
if (authorData.getCreditName()!=null) {
author.addProperty("creditname", authorData.getCreditName());
}
return author.toString();
}
public static String create(AuthorData authorData) {
JsonObject author = new JsonObject();
author.addProperty("oid", authorData.getOid());
author.addProperty("name", authorData.getName());
author.addProperty("surname", authorData.getSurname());
if (authorData.getCreditName() != null) {
author.addProperty("creditname", authorData.getCreditName());
}
return author.toString();
}
}

View File

@ -2,40 +2,49 @@ package eu.dnetlib.doiboost.orcid.model;
public class AuthorData {
private String oid;
private String name;
private String surname;
private String creditName;
private String errorCode;
public String getErrorCode() {
return errorCode;
}
public void setErrorCode(String errorCode) {
this.errorCode = errorCode;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
public String getSurname() {
return surname;
}
public void setSurname(String surname) {
this.surname = surname;
}
public String getCreditName() {
return creditName;
}
public void setCreditName(String creditName) {
this.creditName = creditName;
}
public String getOid() {
return oid;
}
public void setOid(String oid) {
this.oid = oid;
}
private String oid;
private String name;
private String surname;
private String creditName;
private String errorCode;
public String getErrorCode() {
return errorCode;
}
public void setErrorCode(String errorCode) {
this.errorCode = errorCode;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
public String getSurname() {
return surname;
}
public void setSurname(String surname) {
this.surname = surname;
}
public String getCreditName() {
return creditName;
}
public void setCreditName(String creditName) {
this.creditName = creditName;
}
public String getOid() {
return oid;
}
public void setOid(String oid) {
this.oid = oid;
}
}

View File

@ -10,19 +10,6 @@ import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.oaiIProvenance;
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.qualifier;
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.structuredProperty;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Date;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.apache.commons.lang3.StringUtils;
import org.dom4j.Document;
import org.dom4j.DocumentFactory;
import org.dom4j.DocumentHelper;
import org.dom4j.Node;
import eu.dnetlib.dhp.oa.graph.raw.common.MigrationConstants;
import eu.dnetlib.dhp.schema.oaf.Author;
import eu.dnetlib.dhp.schema.oaf.DataInfo;
@ -41,388 +28,439 @@ import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.schema.oaf.Result;
import eu.dnetlib.dhp.schema.oaf.Software;
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Date;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.apache.commons.lang3.StringUtils;
import org.dom4j.Document;
import org.dom4j.DocumentFactory;
import org.dom4j.DocumentHelper;
import org.dom4j.Node;
public abstract class AbstractMdRecordToOafMapper {
protected final Map<String, String> code2name;
protected static final Qualifier MAIN_TITLE_QUALIFIER =
qualifier("main title", "main title", "dnet:dataCite_title", "dnet:dataCite_title");
protected AbstractMdRecordToOafMapper(final Map<String, String> code2name) {
this.code2name = code2name;
}
public List<Oaf> processMdRecord(final String xml) {
try {
final Map<String, String> nsContext = new HashMap<>();
nsContext.put("dr", "http://www.driver-repository.eu/namespace/dr");
nsContext.put("dri", "http://www.driver-repository.eu/namespace/dri");
nsContext.put("oaf", "http://namespace.openaire.eu/oaf");
nsContext.put("oai", "http://www.openarchives.org/OAI/2.0/");
nsContext.put("prov", "http://www.openarchives.org/OAI/2.0/provenance");
nsContext.put("dc", "http://purl.org/dc/elements/1.1/");
nsContext.put("datacite", "http://datacite.org/schema/kernel-3");
DocumentFactory.getInstance().setXPathNamespaceURIs(nsContext);
final Document doc =
DocumentHelper.parseText(xml.replaceAll("http://datacite.org/schema/kernel-4", "http://datacite.org/schema/kernel-3"));
final String type = doc.valueOf("//dr:CobjCategory/@type");
final KeyValue collectedFrom =
keyValue(createOpenaireId(10, doc.valueOf("//oaf:collectedFrom/@id"), true), doc.valueOf("//oaf:collectedFrom/@name"));
final KeyValue hostedBy =
StringUtils.isBlank(doc.valueOf("//oaf:hostedBy/@id"))
? collectedFrom
: keyValue(createOpenaireId(10, doc.valueOf("//oaf:hostedBy/@id"), true), doc.valueOf("//oaf:hostedBy/@name"));
final DataInfo info = prepareDataInfo(doc);
final long lastUpdateTimestamp = new Date().getTime();
return createOafs(doc, type, collectedFrom, hostedBy, info, lastUpdateTimestamp);
} catch (final Exception e) {
throw new RuntimeException(e);
}
}
protected List<Oaf> createOafs(
final Document doc,
final String type,
final KeyValue collectedFrom,
final KeyValue hostedBy,
final DataInfo info,
final long lastUpdateTimestamp) {
final List<Oaf> oafs = new ArrayList<>();
switch (type.toLowerCase()) {
case "":
case "publication":
final Publication p = new Publication();
populateResultFields(p, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
p.setResulttype(MigrationConstants.PUBLICATION_RESULTTYPE_QUALIFIER);
p.setJournal(prepareJournal(doc, info));
oafs.add(p);
break;
case "dataset":
final Dataset d = new Dataset();
populateResultFields(d, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
d.setResulttype(MigrationConstants.DATASET_RESULTTYPE_QUALIFIER);
d.setStoragedate(prepareDatasetStorageDate(doc, info));
d.setDevice(prepareDatasetDevice(doc, info));
d.setSize(prepareDatasetSize(doc, info));
d.setVersion(prepareDatasetVersion(doc, info));
d.setLastmetadataupdate(prepareDatasetLastMetadataUpdate(doc, info));
d.setMetadataversionnumber(prepareDatasetMetadataVersionNumber(doc, info));
d.setGeolocation(prepareDatasetGeoLocations(doc, info));
oafs.add(d);
break;
case "software":
final Software s = new Software();
populateResultFields(s, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
s.setResulttype(MigrationConstants.SOFTWARE_RESULTTYPE_QUALIFIER);
s.setDocumentationUrl(prepareSoftwareDocumentationUrls(doc, info));
s.setLicense(prepareSoftwareLicenses(doc, info));
s.setCodeRepositoryUrl(prepareSoftwareCodeRepositoryUrl(doc, info));
s.setProgrammingLanguage(prepareSoftwareProgrammingLanguage(doc, info));
oafs.add(s);
break;
case "otherresearchproducts":
default:
final OtherResearchProduct o = new OtherResearchProduct();
populateResultFields(o, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
o.setResulttype(MigrationConstants.OTHER_RESULTTYPE_QUALIFIER);
o.setContactperson(prepareOtherResearchProductContactPersons(doc, info));
o.setContactgroup(prepareOtherResearchProductContactGroups(doc, info));
o.setTool(prepareOtherResearchProductTools(doc, info));
oafs.add(o);
break;
}
if (!oafs.isEmpty()) {
oafs.addAll(addProjectRels(doc, collectedFrom, info, lastUpdateTimestamp));
oafs.addAll(addOtherResultRels(doc, collectedFrom, info, lastUpdateTimestamp));
}
return oafs;
}
private List<Oaf> addProjectRels(
final Document doc,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp) {
final List<Oaf> res = new ArrayList<>();
final String docId = createOpenaireId(50, doc.valueOf("//dri:objIdentifier"), false);
for (final Object o : doc.selectNodes("//oaf:projectid")) {
final String projectId = createOpenaireId(40, ((Node) o).getText(), true);
final Relation r1 = new Relation();
r1.setRelType("resultProject");
r1.setSubRelType("outcome");
r1.setRelClass("isProducedBy");
r1.setSource(docId);
r1.setTarget(projectId);
r1.setCollectedfrom(Arrays.asList(collectedFrom));
r1.setDataInfo(info);
r1.setLastupdatetimestamp(lastUpdateTimestamp);
res.add(r1);
final Relation r2 = new Relation();
r2.setRelType("resultProject");
r2.setSubRelType("outcome");
r2.setRelClass("produces");
r2.setSource(projectId);
r2.setTarget(docId);
r2.setCollectedfrom(Arrays.asList(collectedFrom));
r2.setDataInfo(info);
r2.setLastupdatetimestamp(lastUpdateTimestamp);
res.add(r2);
}
return res;
}
protected abstract List<Oaf> addOtherResultRels(
final Document doc,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp);
private void populateResultFields(
final Result r,
final Document doc,
final KeyValue collectedFrom,
final KeyValue hostedBy,
final DataInfo info,
final long lastUpdateTimestamp) {
r.setDataInfo(info);
r.setLastupdatetimestamp(lastUpdateTimestamp);
r.setId(createOpenaireId(50, doc.valueOf("//dri:objIdentifier"), false));
r.setOriginalId(Arrays.asList(doc.valueOf("//dri:objIdentifier")));
r.setCollectedfrom(Arrays.asList(collectedFrom));
r.setPid(prepareListStructProps(doc, "//oaf:identifier", "@identifierType", "dnet:pid_types", "dnet:pid_types", info));
r.setDateofcollection(doc.valueOf("//dr:dateOfCollection"));
r.setDateoftransformation(doc.valueOf("//dr:dateOfTransformation"));
r.setExtraInfo(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setOaiprovenance(prepareOAIprovenance(doc));
r.setAuthor(prepareAuthors(doc, info));
r.setLanguage(prepareLanguages(doc));
r.setCountry(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setSubject(prepareSubjects(doc, info));
r.setTitle(prepareTitles(doc, info));
r.setRelevantdate(prepareRelevantDates(doc, info));
r.setDescription(prepareDescriptions(doc, info));
r.setDateofacceptance(prepareField(doc, "//oaf:dateAccepted", info));
r.setPublisher(preparePublisher(doc, info));
r.setEmbargoenddate(prepareField(doc, "//oaf:embargoenddate", info));
r.setSource(prepareSources(doc, info));
r.setFulltext(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setFormat(prepareFormats(doc, info));
r.setContributor(prepareContributors(doc, info));
r.setResourcetype(prepareResourceType(doc, info));
r.setCoverage(prepareCoverages(doc, info));
r.setContext(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setExternalReference(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setInstance(prepareInstances(doc, info, collectedFrom, hostedBy));
}
protected abstract Qualifier prepareResourceType(Document doc, DataInfo info);
protected abstract List<Instance> prepareInstances(
Document doc,
DataInfo info,
KeyValue collectedfrom,
KeyValue hostedby);
protected abstract List<Field<String>> prepareSources(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareRelevantDates(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareCoverages(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareContributors(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareFormats(Document doc, DataInfo info);
protected abstract Field<String> preparePublisher(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareDescriptions(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareTitles(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareSubjects(Document doc, DataInfo info);
protected abstract Qualifier prepareLanguages(Document doc);
protected abstract List<Author> prepareAuthors(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductTools(
Document doc,
DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactGroups(
Document doc,
DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactPersons(
Document doc,
DataInfo info);
protected abstract Qualifier prepareSoftwareProgrammingLanguage(Document doc, DataInfo info);
protected abstract Field<String> prepareSoftwareCodeRepositoryUrl(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareSoftwareLicenses(
Document doc,
DataInfo info);
protected abstract List<Field<String>> prepareSoftwareDocumentationUrls(
Document doc,
DataInfo info);
protected abstract List<GeoLocation> prepareDatasetGeoLocations(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetMetadataVersionNumber(
Document doc,
DataInfo info);
protected abstract Field<String> prepareDatasetLastMetadataUpdate(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetVersion(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetSize(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetDevice(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetStorageDate(Document doc, DataInfo info);
private Journal prepareJournal(final Document doc, final DataInfo info) {
final Node n = doc.selectSingleNode("//oaf:journal");
if (n != null) {
final String name = n.getText();
final String issnPrinted = n.valueOf("@issn");
final String issnOnline = n.valueOf("@eissn");
final String issnLinking = n.valueOf("@lissn");
final String ep = n.valueOf("@ep");
final String iss = n.valueOf("@iss");
final String sp = n.valueOf("@sp");
final String vol = n.valueOf("@vol");
final String edition = n.valueOf("@edition");
if (StringUtils.isNotBlank(name)) { return journal(name, issnPrinted, issnOnline, issnLinking, ep, iss, sp, vol, edition, null, null, info); }
}
return null;
}
protected Qualifier prepareQualifier(
final Node node,
final String xpath,
final String schemeId,
final String schemeName) {
final String classId = node.valueOf(xpath);
final String className = code2name.get(classId);
return qualifier(classId, className, schemeId, schemeName);
}
protected List<StructuredProperty> prepareListStructProps(
final Node node,
final String xpath,
final String xpathClassId,
final String schemeId,
final String schemeName,
final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
final String classId = n.valueOf(xpathClassId);
final String className = code2name.get(classId);
res.add(structuredProperty(n.getText(), classId, className, schemeId, schemeName, info));
}
return res;
}
protected List<StructuredProperty> prepareListStructProps(
final Node node,
final String xpath,
final Qualifier qualifier,
final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
res.add(structuredProperty(n.getText(), qualifier, info));
}
return res;
}
protected List<StructuredProperty> prepareListStructProps(
final Node node,
final String xpath,
final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
res.add(structuredProperty(n.getText(), n.valueOf("@classid"), n.valueOf("@classname"), n.valueOf("@schemeid"), n.valueOf("@schemename"), info));
}
return res;
}
protected OAIProvenance prepareOAIprovenance(final Document doc) {
final Node n =
doc.selectSingleNode("//*[local-name()='provenance']/*[local-name()='originDescription']");
if (n == null) { return null; }
final String identifier = n.valueOf("./*[local-name()='identifier']");
final String baseURL = n.valueOf("./*[local-name()='baseURL']");;
final String metadataNamespace = n.valueOf("./*[local-name()='metadataNamespace']");;
final boolean altered = n.valueOf("@altered").equalsIgnoreCase("true");
final String datestamp = n.valueOf("./*[local-name()='datestamp']");;
final String harvestDate = n.valueOf("@harvestDate");;
return oaiIProvenance(identifier, baseURL, metadataNamespace, altered, datestamp, harvestDate);
}
protected DataInfo prepareDataInfo(final Document doc) {
final Node n = doc.selectSingleNode("//oaf:datainfo");
if (n == null) { return dataInfo(false, null, false, false, MigrationConstants.REPOSITORY_PROVENANCE_ACTIONS, "0.9"); }
final String paClassId = n.valueOf("./oaf:provenanceaction/@classid");
final String paClassName = n.valueOf("./oaf:provenanceaction/@classname");
final String paSchemeId = n.valueOf("./oaf:provenanceaction/@schemeid");
final String paSchemeName = n.valueOf("./oaf:provenanceaction/@schemename");
final boolean deletedbyinference =
Boolean.parseBoolean(n.valueOf("./oaf:deletedbyinference"));
final String inferenceprovenance = n.valueOf("./oaf:inferenceprovenance");
final Boolean inferred = Boolean.parseBoolean(n.valueOf("./oaf:inferred"));
final String trust = n.valueOf("./oaf:trust");
return dataInfo(deletedbyinference, inferenceprovenance, inferred, false, qualifier(paClassId, paClassName, paSchemeId, paSchemeName), trust);
}
protected Field<String> prepareField(final Node node, final String xpath, final DataInfo info) {
return field(node.valueOf(xpath), info);
}
protected List<Field<String>> prepareListFields(
final Node node,
final String xpath,
final DataInfo info) {
return listFields(info, prepareListString(node, xpath));
}
protected List<String> prepareListString(final Node node, final String xpath) {
final List<String> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final String s = ((Node) o).getText().trim();
if (StringUtils.isNotBlank(s)) {
res.add(s);
}
}
return res;
}
protected final Map<String, String> code2name;
protected static final Qualifier MAIN_TITLE_QUALIFIER =
qualifier("main title", "main title", "dnet:dataCite_title", "dnet:dataCite_title");
protected AbstractMdRecordToOafMapper(final Map<String, String> code2name) {
this.code2name = code2name;
}
public List<Oaf> processMdRecord(final String xml) {
try {
final Map<String, String> nsContext = new HashMap<>();
nsContext.put("dr", "http://www.driver-repository.eu/namespace/dr");
nsContext.put("dri", "http://www.driver-repository.eu/namespace/dri");
nsContext.put("oaf", "http://namespace.openaire.eu/oaf");
nsContext.put("oai", "http://www.openarchives.org/OAI/2.0/");
nsContext.put("prov", "http://www.openarchives.org/OAI/2.0/provenance");
nsContext.put("dc", "http://purl.org/dc/elements/1.1/");
nsContext.put("datacite", "http://datacite.org/schema/kernel-3");
DocumentFactory.getInstance().setXPathNamespaceURIs(nsContext);
final Document doc =
DocumentHelper.parseText(
xml.replaceAll(
"http://datacite.org/schema/kernel-4",
"http://datacite.org/schema/kernel-3"));
final String type = doc.valueOf("//dr:CobjCategory/@type");
final KeyValue collectedFrom =
keyValue(
createOpenaireId(10, doc.valueOf("//oaf:collectedFrom/@id"), true),
doc.valueOf("//oaf:collectedFrom/@name"));
final KeyValue hostedBy =
StringUtils.isBlank(doc.valueOf("//oaf:hostedBy/@id"))
? collectedFrom
: keyValue(
createOpenaireId(10, doc.valueOf("//oaf:hostedBy/@id"), true),
doc.valueOf("//oaf:hostedBy/@name"));
final DataInfo info = prepareDataInfo(doc);
final long lastUpdateTimestamp = new Date().getTime();
return createOafs(doc, type, collectedFrom, hostedBy, info, lastUpdateTimestamp);
} catch (final Exception e) {
throw new RuntimeException(e);
}
}
protected List<Oaf> createOafs(
final Document doc,
final String type,
final KeyValue collectedFrom,
final KeyValue hostedBy,
final DataInfo info,
final long lastUpdateTimestamp) {
final List<Oaf> oafs = new ArrayList<>();
switch (type.toLowerCase()) {
case "":
case "publication":
final Publication p = new Publication();
populateResultFields(p, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
p.setResulttype(MigrationConstants.PUBLICATION_RESULTTYPE_QUALIFIER);
p.setJournal(prepareJournal(doc, info));
oafs.add(p);
break;
case "dataset":
final Dataset d = new Dataset();
populateResultFields(d, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
d.setResulttype(MigrationConstants.DATASET_RESULTTYPE_QUALIFIER);
d.setStoragedate(prepareDatasetStorageDate(doc, info));
d.setDevice(prepareDatasetDevice(doc, info));
d.setSize(prepareDatasetSize(doc, info));
d.setVersion(prepareDatasetVersion(doc, info));
d.setLastmetadataupdate(prepareDatasetLastMetadataUpdate(doc, info));
d.setMetadataversionnumber(prepareDatasetMetadataVersionNumber(doc, info));
d.setGeolocation(prepareDatasetGeoLocations(doc, info));
oafs.add(d);
break;
case "software":
final Software s = new Software();
populateResultFields(s, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
s.setResulttype(MigrationConstants.SOFTWARE_RESULTTYPE_QUALIFIER);
s.setDocumentationUrl(prepareSoftwareDocumentationUrls(doc, info));
s.setLicense(prepareSoftwareLicenses(doc, info));
s.setCodeRepositoryUrl(prepareSoftwareCodeRepositoryUrl(doc, info));
s.setProgrammingLanguage(prepareSoftwareProgrammingLanguage(doc, info));
oafs.add(s);
break;
case "otherresearchproducts":
default:
final OtherResearchProduct o = new OtherResearchProduct();
populateResultFields(o, doc, collectedFrom, hostedBy, info, lastUpdateTimestamp);
o.setResulttype(MigrationConstants.OTHER_RESULTTYPE_QUALIFIER);
o.setContactperson(prepareOtherResearchProductContactPersons(doc, info));
o.setContactgroup(prepareOtherResearchProductContactGroups(doc, info));
o.setTool(prepareOtherResearchProductTools(doc, info));
oafs.add(o);
break;
}
if (!oafs.isEmpty()) {
oafs.addAll(addProjectRels(doc, collectedFrom, info, lastUpdateTimestamp));
oafs.addAll(addOtherResultRels(doc, collectedFrom, info, lastUpdateTimestamp));
}
return oafs;
}
private List<Oaf> addProjectRels(
final Document doc,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp) {
final List<Oaf> res = new ArrayList<>();
final String docId = createOpenaireId(50, doc.valueOf("//dri:objIdentifier"), false);
for (final Object o : doc.selectNodes("//oaf:projectid")) {
final String projectId = createOpenaireId(40, ((Node) o).getText(), true);
final Relation r1 = new Relation();
r1.setRelType("resultProject");
r1.setSubRelType("outcome");
r1.setRelClass("isProducedBy");
r1.setSource(docId);
r1.setTarget(projectId);
r1.setCollectedfrom(Arrays.asList(collectedFrom));
r1.setDataInfo(info);
r1.setLastupdatetimestamp(lastUpdateTimestamp);
res.add(r1);
final Relation r2 = new Relation();
r2.setRelType("resultProject");
r2.setSubRelType("outcome");
r2.setRelClass("produces");
r2.setSource(projectId);
r2.setTarget(docId);
r2.setCollectedfrom(Arrays.asList(collectedFrom));
r2.setDataInfo(info);
r2.setLastupdatetimestamp(lastUpdateTimestamp);
res.add(r2);
}
return res;
}
protected abstract List<Oaf> addOtherResultRels(
final Document doc,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp);
private void populateResultFields(
final Result r,
final Document doc,
final KeyValue collectedFrom,
final KeyValue hostedBy,
final DataInfo info,
final long lastUpdateTimestamp) {
r.setDataInfo(info);
r.setLastupdatetimestamp(lastUpdateTimestamp);
r.setId(createOpenaireId(50, doc.valueOf("//dri:objIdentifier"), false));
r.setOriginalId(Arrays.asList(doc.valueOf("//dri:objIdentifier")));
r.setCollectedfrom(Arrays.asList(collectedFrom));
r.setPid(
prepareListStructProps(
doc,
"//oaf:identifier",
"@identifierType",
"dnet:pid_types",
"dnet:pid_types",
info));
r.setDateofcollection(doc.valueOf("//dr:dateOfCollection"));
r.setDateoftransformation(doc.valueOf("//dr:dateOfTransformation"));
r.setExtraInfo(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setOaiprovenance(prepareOAIprovenance(doc));
r.setAuthor(prepareAuthors(doc, info));
r.setLanguage(prepareLanguages(doc));
r.setCountry(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setSubject(prepareSubjects(doc, info));
r.setTitle(prepareTitles(doc, info));
r.setRelevantdate(prepareRelevantDates(doc, info));
r.setDescription(prepareDescriptions(doc, info));
r.setDateofacceptance(prepareField(doc, "//oaf:dateAccepted", info));
r.setPublisher(preparePublisher(doc, info));
r.setEmbargoenddate(prepareField(doc, "//oaf:embargoenddate", info));
r.setSource(prepareSources(doc, info));
r.setFulltext(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setFormat(prepareFormats(doc, info));
r.setContributor(prepareContributors(doc, info));
r.setResourcetype(prepareResourceType(doc, info));
r.setCoverage(prepareCoverages(doc, info));
r.setContext(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setExternalReference(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setInstance(prepareInstances(doc, info, collectedFrom, hostedBy));
}
protected abstract Qualifier prepareResourceType(Document doc, DataInfo info);
protected abstract List<Instance> prepareInstances(
Document doc, DataInfo info, KeyValue collectedfrom, KeyValue hostedby);
protected abstract List<Field<String>> prepareSources(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareRelevantDates(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareCoverages(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareContributors(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareFormats(Document doc, DataInfo info);
protected abstract Field<String> preparePublisher(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareDescriptions(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareTitles(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareSubjects(Document doc, DataInfo info);
protected abstract Qualifier prepareLanguages(Document doc);
protected abstract List<Author> prepareAuthors(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductTools(
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactGroups(
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactPersons(
Document doc, DataInfo info);
protected abstract Qualifier prepareSoftwareProgrammingLanguage(Document doc, DataInfo info);
protected abstract Field<String> prepareSoftwareCodeRepositoryUrl(Document doc, DataInfo info);
protected abstract List<StructuredProperty> prepareSoftwareLicenses(
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareSoftwareDocumentationUrls(
Document doc, DataInfo info);
protected abstract List<GeoLocation> prepareDatasetGeoLocations(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetMetadataVersionNumber(
Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetLastMetadataUpdate(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetVersion(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetSize(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetDevice(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetStorageDate(Document doc, DataInfo info);
private Journal prepareJournal(final Document doc, final DataInfo info) {
final Node n = doc.selectSingleNode("//oaf:journal");
if (n != null) {
final String name = n.getText();
final String issnPrinted = n.valueOf("@issn");
final String issnOnline = n.valueOf("@eissn");
final String issnLinking = n.valueOf("@lissn");
final String ep = n.valueOf("@ep");
final String iss = n.valueOf("@iss");
final String sp = n.valueOf("@sp");
final String vol = n.valueOf("@vol");
final String edition = n.valueOf("@edition");
if (StringUtils.isNotBlank(name)) {
return journal(
name,
issnPrinted,
issnOnline,
issnLinking,
ep,
iss,
sp,
vol,
edition,
null,
null,
info);
}
}
return null;
}
protected Qualifier prepareQualifier(
final Node node, final String xpath, final String schemeId, final String schemeName) {
final String classId = node.valueOf(xpath);
final String className = code2name.get(classId);
return qualifier(classId, className, schemeId, schemeName);
}
protected List<StructuredProperty> prepareListStructProps(
final Node node,
final String xpath,
final String xpathClassId,
final String schemeId,
final String schemeName,
final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
final String classId = n.valueOf(xpathClassId);
final String className = code2name.get(classId);
res.add(
structuredProperty(
n.getText(), classId, className, schemeId, schemeName, info));
}
return res;
}
protected List<StructuredProperty> prepareListStructProps(
final Node node, final String xpath, final Qualifier qualifier, final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
res.add(structuredProperty(n.getText(), qualifier, info));
}
return res;
}
protected List<StructuredProperty> prepareListStructProps(
final Node node, final String xpath, final DataInfo info) {
final List<StructuredProperty> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final Node n = (Node) o;
res.add(
structuredProperty(
n.getText(),
n.valueOf("@classid"),
n.valueOf("@classname"),
n.valueOf("@schemeid"),
n.valueOf("@schemename"),
info));
}
return res;
}
protected OAIProvenance prepareOAIprovenance(final Document doc) {
final Node n =
doc.selectSingleNode(
"//*[local-name()='provenance']/*[local-name()='originDescription']");
if (n == null) {
return null;
}
final String identifier = n.valueOf("./*[local-name()='identifier']");
final String baseURL = n.valueOf("./*[local-name()='baseURL']");
;
final String metadataNamespace = n.valueOf("./*[local-name()='metadataNamespace']");
;
final boolean altered = n.valueOf("@altered").equalsIgnoreCase("true");
final String datestamp = n.valueOf("./*[local-name()='datestamp']");
;
final String harvestDate = n.valueOf("@harvestDate");
;
return oaiIProvenance(
identifier, baseURL, metadataNamespace, altered, datestamp, harvestDate);
}
protected DataInfo prepareDataInfo(final Document doc) {
final Node n = doc.selectSingleNode("//oaf:datainfo");
if (n == null) {
return dataInfo(
false,
null,
false,
false,
MigrationConstants.REPOSITORY_PROVENANCE_ACTIONS,
"0.9");
}
final String paClassId = n.valueOf("./oaf:provenanceaction/@classid");
final String paClassName = n.valueOf("./oaf:provenanceaction/@classname");
final String paSchemeId = n.valueOf("./oaf:provenanceaction/@schemeid");
final String paSchemeName = n.valueOf("./oaf:provenanceaction/@schemename");
final boolean deletedbyinference =
Boolean.parseBoolean(n.valueOf("./oaf:deletedbyinference"));
final String inferenceprovenance = n.valueOf("./oaf:inferenceprovenance");
final Boolean inferred = Boolean.parseBoolean(n.valueOf("./oaf:inferred"));
final String trust = n.valueOf("./oaf:trust");
return dataInfo(
deletedbyinference,
inferenceprovenance,
inferred,
false,
qualifier(paClassId, paClassName, paSchemeId, paSchemeName),
trust);
}
protected Field<String> prepareField(final Node node, final String xpath, final DataInfo info) {
return field(node.valueOf(xpath), info);
}
protected List<Field<String>> prepareListFields(
final Node node, final String xpath, final DataInfo info) {
return listFields(info, prepareListString(node, xpath));
}
protected List<String> prepareListString(final Node node, final String xpath) {
final List<String> res = new ArrayList<>();
for (final Object o : node.selectNodes(xpath)) {
final String s = ((Node) o).getText().trim();
if (StringUtils.isNotBlank(s)) {
res.add(s);
}
}
return res;
}
}

View File

@ -6,17 +6,28 @@ import eu.dnetlib.dhp.schema.oaf.Qualifier;
public class MigrationConstants {
public static final Qualifier PUBLICATION_RESULTTYPE_QUALIFIER =
qualifier("publication", "publication", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier DATASET_RESULTTYPE_QUALIFIER =
qualifier("dataset", "dataset", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier SOFTWARE_RESULTTYPE_QUALIFIER =
qualifier("software", "software", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier OTHER_RESULTTYPE_QUALIFIER =
qualifier("other", "other", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier REPOSITORY_PROVENANCE_ACTIONS =
qualifier("sysimport:crosswalk:repository", "sysimport:crosswalk:repository", "dnet:provenanceActions", "dnet:provenanceActions");
public static final Qualifier ENTITYREGISTRY_PROVENANCE_ACTION =
qualifier("sysimport:crosswalk:entityregistry", "sysimport:crosswalk:entityregistry", "dnet:provenanceActions", "dnet:provenanceActions");
public static final Qualifier PUBLICATION_RESULTTYPE_QUALIFIER =
qualifier(
"publication",
"publication",
"dnet:result_typologies",
"dnet:result_typologies");
public static final Qualifier DATASET_RESULTTYPE_QUALIFIER =
qualifier("dataset", "dataset", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier SOFTWARE_RESULTTYPE_QUALIFIER =
qualifier("software", "software", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier OTHER_RESULTTYPE_QUALIFIER =
qualifier("other", "other", "dnet:result_typologies", "dnet:result_typologies");
public static final Qualifier REPOSITORY_PROVENANCE_ACTIONS =
qualifier(
"sysimport:crosswalk:repository",
"sysimport:crosswalk:repository",
"dnet:provenanceActions",
"dnet:provenanceActions");
public static final Qualifier ENTITYREGISTRY_PROVENANCE_ACTION =
qualifier(
"sysimport:crosswalk:entityregistry",
"sysimport:crosswalk:entityregistry",
"dnet:provenanceActions",
"dnet:provenanceActions");
}