sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

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
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/crossref/CrossrefImporter.java
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),
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);
}
}
}

61
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/crossref/ESClient.java
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) {
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;

4
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/json/JsonWriter.java
View File

@ -1,10 +1,8 @@
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) {
@ -12,7 +10,7 @@ public class JsonWriter {
author.addProperty("oid", authorData.getOid());
author.addProperty("name", authorData.getName());
author.addProperty("surname", authorData.getSurname());
if (authorData.getCreditName()!=null) {
if (authorData.getCreditName() != null) {
author.addProperty("creditname", authorData.getCreditName());
}
return author.toString();

9
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/model/AuthorData.java
View File

@ -11,30 +11,39 @@ public class AuthorData {
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;
}

156
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java
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;
@ -41,6 +28,17 @@ 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 {
@ -66,15 +64,22 @@ public abstract class AbstractMdRecordToOafMapper {
DocumentFactory.getInstance().setXPathNamespaceURIs(nsContext);
final Document doc =
DocumentHelper.parseText(xml.replaceAll("http://datacite.org/schema/kernel-4", "http://datacite.org/schema/kernel-3"));
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"));
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"));
: 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();
@ -204,7 +209,14 @@ public abstract class AbstractMdRecordToOafMapper {
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.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
@ -233,10 +245,7 @@ public abstract class AbstractMdRecordToOafMapper {
protected abstract Qualifier prepareResourceType(Document doc, DataInfo info);
protected abstract List<Instance> prepareInstances(
Document doc,
DataInfo info,
KeyValue collectedfrom,
KeyValue hostedby);
Document doc, DataInfo info, KeyValue collectedfrom, KeyValue hostedby);
protected abstract List<Field<String>> prepareSources(Document doc, DataInfo info);
@ -261,34 +270,28 @@ public abstract class AbstractMdRecordToOafMapper {
protected abstract List<Author> prepareAuthors(Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductTools(
Document doc,
DataInfo info);
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactGroups(
Document doc,
DataInfo info);
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareOtherResearchProductContactPersons(
Document doc,
DataInfo info);
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);
Document doc, DataInfo info);
protected abstract List<Field<String>> prepareSoftwareDocumentationUrls(
Document doc,
DataInfo info);
Document doc, DataInfo info);
protected abstract List<GeoLocation> prepareDatasetGeoLocations(Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetMetadataVersionNumber(
Document doc,
DataInfo info);
Document doc, DataInfo info);
protected abstract Field<String> prepareDatasetLastMetadataUpdate(Document doc, DataInfo info);
@ -312,16 +315,27 @@ public abstract class AbstractMdRecordToOafMapper {
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); }
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 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);
@ -339,16 +353,15 @@ public abstract class AbstractMdRecordToOafMapper {
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));
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 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;
@ -358,37 +371,58 @@ public abstract class AbstractMdRecordToOafMapper {
}
protected List<StructuredProperty> prepareListStructProps(
final Node node,
final String xpath,
final DataInfo info) {
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));
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']");
doc.selectSingleNode(
"//*[local-name()='provenance']/*[local-name()='originDescription']");
if (n == null) { return null; }
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 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");;
final String datestamp = n.valueOf("./*[local-name()='datestamp']");
;
final String harvestDate = n.valueOf("@harvestDate");
;
return oaiIProvenance(identifier, baseURL, metadataNamespace, altered, datestamp, 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"); }
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");
@ -401,7 +435,13 @@ public abstract class AbstractMdRecordToOafMapper {
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);
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) {
@ -409,9 +449,7 @@ public abstract class AbstractMdRecordToOafMapper {
}
protected List<Field<String>> prepareListFields(
final Node node,
final String xpath,
final DataInfo info) {
final Node node, final String xpath, final DataInfo info) {
return listFields(info, prepareListString(node, xpath));
}

144
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplication.java
View File

@ -10,23 +10,6 @@ import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.listKeyValues;
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.qualifier;
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.structuredProperty;
import java.io.Closeable;
import java.io.IOException;
import java.sql.Array;
import java.sql.ResultSet;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Date;
import java.util.List;
import java.util.function.Consumer;
import java.util.function.Function;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.oa.graph.raw.common.AbstractMigrationApplication;
import eu.dnetlib.dhp.oa.graph.raw.common.DbClient;
@ -48,6 +31,21 @@ 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.io.Closeable;
import java.io.IOException;
import java.sql.Array;
import java.sql.ResultSet;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Date;
import java.util.List;
import java.util.function.Consumer;
import java.util.function.Function;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
implements Closeable {
@ -61,8 +59,9 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
public static void main(final String[] args) throws Exception {
final ArgumentApplicationParser parser =
new ArgumentApplicationParser(
IOUtils.toString(MigrateDbEntitiesApplication.class
.getResourceAsStream("/eu/dnetlib/dhp/oa/graph/migrate_db_entities_parameters.json")));
IOUtils.toString(
MigrateDbEntitiesApplication.class.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/migrate_db_entities_parameters.json")));
parser.parseArgument(args);
@ -91,7 +90,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
smdbe.execute("queryOrganizations.sql", smdbe::processOrganization);
log.info("Processing relations ds <-> orgs ...");
smdbe.execute("queryDatasourceOrganization.sql", smdbe::processDatasourceOrganization);
smdbe.execute(
"queryDatasourceOrganization.sql", smdbe::processDatasourceOrganization);
log.info("Processing projects <-> orgs ...");
smdbe.execute("queryProjectOrganization.sql", smdbe::processProjectOrganization);
@ -117,7 +117,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
public void execute(final String sqlFile, final Function<ResultSet, List<Oaf>> producer)
throws Exception {
final String sql =
IOUtils.toString(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/sql/" + sqlFile));
IOUtils.toString(
getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/sql/" + sqlFile));
final Consumer<ResultSet> consumer = rs -> producer.apply(rs).forEach(oaf -> emitOaf(oaf));
@ -134,14 +135,18 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
ds.setId(createOpenaireId(10, rs.getString("datasourceid"), true));
ds.setOriginalId(Arrays.asList(rs.getString("datasourceid")));
ds.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
ds.setCollectedfrom(
listKeyValues(
createOpenaireId(10, rs.getString("collectedfromid"), true),
rs.getString("collectedfromname")));
ds.setPid(new ArrayList<>());
ds.setDateofcollection(asString(rs.getDate("dateofcollection")));
ds.setDateoftransformation(null); // Value not returned by the SQL query
ds.setExtraInfo(new ArrayList<>()); // Values not present in the DB
ds.setOaiprovenance(null); // Values not present in the DB
ds.setDatasourcetype(prepareQualifierSplitting(rs.getString("datasourcetype")));
ds.setOpenairecompatibility(prepareQualifierSplitting(rs.getString("openairecompatibility")));
ds.setOpenairecompatibility(
prepareQualifierSplitting(rs.getString("openairecompatibility")));
ds.setOfficialname(field(rs.getString("officialname"), info));
ds.setEnglishname(field(rs.getString("englishname"), info));
ds.setWebsiteurl(field(rs.getString("websiteurl"), info));
@ -174,7 +179,11 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
ds.setPidsystems(field(rs.getString("pidsystems"), info));
ds.setCertificates(field(rs.getString("certificates"), info));
ds.setPolicies(new ArrayList<>()); // The sql query returns an empty array
ds.setJournal(prepareJournal(rs.getString("officialname"), rs.getString("journal"), info)); // Journal
ds.setJournal(
prepareJournal(
rs.getString("officialname"),
rs.getString("journal"),
info)); // Journal
ds.setDataInfo(info);
ds.setLastupdatetimestamp(lastUpdateTimestamp);
@ -193,7 +202,10 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
p.setId(createOpenaireId(40, rs.getString("projectid"), true));
p.setOriginalId(Arrays.asList(rs.getString("projectid")));
p.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
p.setCollectedfrom(
listKeyValues(
createOpenaireId(10, rs.getString("collectedfromid"), true),
rs.getString("collectedfromname")));
p.setPid(new ArrayList<>());
p.setDateofcollection(asString(rs.getDate("dateofcollection")));
p.setDateoftransformation(asString(rs.getDate("dateoftransformation")));
@ -209,7 +221,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
p.setKeywords(field(rs.getString("keywords"), info));
p.setDuration(field(Integer.toString(rs.getInt("duration")), info));
p.setEcsc39(field(Boolean.toString(rs.getBoolean("ecsc39")), info));
p.setOamandatepublications(field(Boolean.toString(rs.getBoolean("oamandatepublications")), info));
p.setOamandatepublications(
field(Boolean.toString(rs.getBoolean("oamandatepublications")), info));
p.setEcarticle29_3(field(Boolean.toString(rs.getBoolean("ecarticle29_3")), info));
p.setSubjects(prepareListOfStructProps(rs.getArray("subjects"), info));
p.setFundingtree(prepareListFields(rs.getArray("fundingtree"), info));
@ -245,7 +258,10 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
o.setId(createOpenaireId(20, rs.getString("organizationid"), true));
o.setOriginalId(Arrays.asList(rs.getString("organizationid")));
o.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
o.setCollectedfrom(
listKeyValues(
createOpenaireId(10, rs.getString("collectedfromid"), true),
rs.getString("collectedfromname")));
o.setPid(new ArrayList<>());
o.setDateofcollection(asString(rs.getDate("dateofcollection")));
o.setDateoftransformation(asString(rs.getDate("dateoftransformation")));
@ -259,10 +275,17 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
o.setEclegalbody(field(Boolean.toString(rs.getBoolean("eclegalbody")), info));
o.setEclegalperson(field(Boolean.toString(rs.getBoolean("eclegalperson")), info));
o.setEcnonprofit(field(Boolean.toString(rs.getBoolean("ecnonprofit")), info));
o.setEcresearchorganization(field(Boolean.toString(rs.getBoolean("ecresearchorganization")), info));
o.setEchighereducation(field(Boolean.toString(rs.getBoolean("echighereducation")), info));
o.setEcinternationalorganizationeurinterests(field(Boolean.toString(rs.getBoolean("ecinternationalorganizationeurinterests")), info));
o.setEcinternationalorganization(field(Boolean.toString(rs.getBoolean("ecinternationalorganization")), info));
o.setEcresearchorganization(
field(Boolean.toString(rs.getBoolean("ecresearchorganization")), info));
o.setEchighereducation(
field(Boolean.toString(rs.getBoolean("echighereducation")), info));
o.setEcinternationalorganizationeurinterests(
field(
Boolean.toString(
rs.getBoolean("ecinternationalorganizationeurinterests")),
info));
o.setEcinternationalorganization(
field(Boolean.toString(rs.getBoolean("ecinternationalorganization")), info));
o.setEcenterprise(field(Boolean.toString(rs.getBoolean("ecenterprise")), info));
o.setEcsmevalidated(field(Boolean.toString(rs.getBoolean("ecsmevalidated")), info));
o.setEcnutscode(field(Boolean.toString(rs.getBoolean("ecnutscode")), info));
@ -282,7 +305,9 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
final String orgId = createOpenaireId(20, rs.getString("organization"), true);
final String dsId = createOpenaireId(10, rs.getString("datasource"), true);
final List<KeyValue> collectedFrom =
listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname"));
listKeyValues(
createOpenaireId(10, rs.getString("collectedfromid"), true),
rs.getString("collectedfromname"));
final Relation r1 = new Relation();
r1.setRelType("datasourceOrganization");
@ -316,7 +341,9 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
final String orgId = createOpenaireId(20, rs.getString("resporganization"), true);
final String projectId = createOpenaireId(40, rs.getString("project"), true);
final List<KeyValue> collectedFrom =
listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname"));
listKeyValues(
createOpenaireId(10, rs.getString("collectedfromid"), true),
rs.getString("collectedfromname"));
final Relation r1 = new Relation();
r1.setRelType("projectOrganization");
@ -347,7 +374,17 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
public List<Oaf> processClaims(final ResultSet rs) {
final DataInfo info =
dataInfo(false, null, false, false, qualifier("user:claim", "user:claim", "dnet:provenanceActions", "dnet:provenanceActions"), "0.9");
dataInfo(
false,
null,
false,
false,
qualifier(
"user:claim",
"user:claim",
"dnet:provenanceActions",
"dnet:provenanceActions"),
"0.9");
final List<KeyValue> collectedFrom =
listKeyValues(createOpenaireId(10, "infrastruct_::openaire", true), "OpenAIRE");
@ -379,9 +416,11 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
return Arrays.asList(r);
} else {
final String sourceId =
createOpenaireId(rs.getString("source_type"), rs.getString("source_id"), false);
createOpenaireId(
rs.getString("source_type"), rs.getString("source_id"), false);
final String targetId =
createOpenaireId(rs.getString("target_type"), rs.getString("target_id"), false);
createOpenaireId(
rs.getString("target_type"), rs.getString("target_id"), false);
final Relation r1 = new Relation();
final Relation r2 = new Relation();
@ -438,11 +477,19 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
final String inferenceprovenance = rs.getString("inferenceprovenance");
final Boolean inferred = rs.getBoolean("inferred");
final String trust = rs.getString("trust");
return dataInfo(deletedbyinference, inferenceprovenance, inferred, false, MigrationConstants.ENTITYREGISTRY_PROVENANCE_ACTION, trust);
return dataInfo(
deletedbyinference,
inferenceprovenance,
inferred,
false,
MigrationConstants.ENTITYREGISTRY_PROVENANCE_ACTION,
trust);
}
private Qualifier prepareQualifierSplitting(final String s) {
if (StringUtils.isBlank(s)) { return null; }
if (StringUtils.isBlank(s)) {
return null;
}
final String[] arr = s.split("@@@");
return arr.length == 4 ? qualifier(arr[0], arr[1], arr[2], arr[3]) : null;
}
@ -458,19 +505,22 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
}
private StructuredProperty prepareStructProp(final String s, final DataInfo dataInfo) {
if (StringUtils.isBlank(s)) { return null; }
if (StringUtils.isBlank(s)) {
return null;
}
final String[] parts = s.split("###");
if (parts.length == 2) {
final String value = parts[0];
final String[] arr = parts[1].split("@@@");
if (arr.length == 4) { return structuredProperty(value, arr[0], arr[1], arr[2], arr[3], dataInfo); }
if (arr.length == 4) {
return structuredProperty(value, arr[0], arr[1], arr[2], arr[3], dataInfo);
}
}
return null;
}
private List<StructuredProperty> prepareListOfStructProps(
final Array array,
final DataInfo dataInfo) throws SQLException {
final Array array, final DataInfo dataInfo) throws SQLException {
final List<StructuredProperty> res = new ArrayList<>();
if (array != null) {
for (final String s : (String[]) array.getArray()) {
@ -489,10 +539,14 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
final String[] arr = sj.split("@@@");
if (arr.length == 3) {
final String issn = StringUtils.isNotBlank(arr[0]) ? arr[0] : null;
final String eissn = StringUtils.isNotBlank(arr[1]) ? arr[1] : null;;
final String lissn = StringUtils.isNotBlank(arr[2]) ? arr[2] : null;;
final String eissn = StringUtils.isNotBlank(arr[1]) ? arr[1] : null;
;
final String lissn = StringUtils.isNotBlank(arr[2]) ? arr[2] : null;
;
if (issn != null || eissn != null || lissn != null) {
return journal(name, issn, eissn, eissn, null, null, null, null, null, null, null, info);
return journal(
name, issn, eissn, eissn, null, null, null, null, null, null, null,
info);
}
}
}

19
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/common/MigrationConstants.java
View File

@ -7,7 +7,11 @@ 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");
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 =
@ -15,8 +19,15 @@ public class MigrationConstants {
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");
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");
qualifier(
"sysimport:crosswalk:entityregistry",
"sysimport:crosswalk:entityregistry",
"dnet:provenanceActions",
"dnet:provenanceActions");
}