forked from D-Net/dnet-hadoop
[AffiliationInference]Extended the affiliation ingestion from OpenAIRE to include also the links derived from web crawl. Changed the provenance from BIP! to OpenAIRE
This commit is contained in:
parent
3ee8a7d18a
commit
4dbce39237
|
|
@ -9,6 +9,7 @@ import java.util.List;
|
|||
|
||||
import org.apache.commons.io.IOUtils;
|
||||
import org.apache.hadoop.io.Text;
|
||||
import org.apache.hadoop.io.compress.BZip2Codec;
|
||||
import org.apache.hadoop.io.compress.GzipCodec;
|
||||
import org.apache.hadoop.mapred.SequenceFileOutputFormat;
|
||||
import org.apache.spark.SparkConf;
|
||||
|
|
@ -40,9 +41,9 @@ public class PrepareAffiliationRelations implements Serializable {
|
|||
private static final Logger log = LoggerFactory.getLogger(PrepareAffiliationRelations.class);
|
||||
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
|
||||
private static final String ID_PREFIX = "50|doi_________::";
|
||||
public static final String BIP_AFFILIATIONS_CLASSID = "result:organization:bipinference";
|
||||
public static final String BIP_AFFILIATIONS_CLASSNAME = "Affiliation relation inferred by BIP!";
|
||||
public static final String BIP_INFERENCE_PROVENANCE = "bip:affiliation:crossref";
|
||||
public static final String BIP_AFFILIATIONS_CLASSID = "result:organization:openaireinference";
|
||||
public static final String BIP_AFFILIATIONS_CLASSNAME = "Affiliation relation inferred by OpenAIRE";
|
||||
public static final String BIP_INFERENCE_PROVENANCE = "openaire:affiliation";
|
||||
|
||||
public static <I extends Result> void main(String[] args) throws Exception {
|
||||
|
||||
|
|
@ -70,6 +71,9 @@ public class PrepareAffiliationRelations implements Serializable {
|
|||
final String dataciteInputPath = parser.get("dataciteInputPath");
|
||||
log.info("dataciteInputPath: {}", dataciteInputPath);
|
||||
|
||||
final String webcrawlInputPath = parser.get("webCrawlInputPath");
|
||||
log.info("webcrawlInputPath: {}", webcrawlInputPath);
|
||||
|
||||
final String outputPath = parser.get("outputPath");
|
||||
log.info("outputPath: {}", outputPath);
|
||||
|
||||
|
|
@ -101,12 +105,18 @@ public class PrepareAffiliationRelations implements Serializable {
|
|||
JavaPairRDD<Text, Text> dataciteRelations = prepareAffiliationRelations(
|
||||
spark, dataciteInputPath, collectedFromDatacite);
|
||||
|
||||
List<KeyValue> collectedFromWebCrawl = OafMapperUtils
|
||||
.listKeyValues(Constants.WEB_CRAWL_ID, Constants.WEB_CRAWL_NAME);
|
||||
JavaPairRDD<Text, Text> webCrawlRelations = prepareAffiliationRelations(
|
||||
spark, webcrawlInputPath, collectedFromWebCrawl);
|
||||
|
||||
crossrefRelations
|
||||
.union(pubmedRelations)
|
||||
.union(openAPCRelations)
|
||||
.union(dataciteRelations)
|
||||
.union(webCrawlRelations)
|
||||
.saveAsHadoopFile(
|
||||
outputPath, Text.class, Text.class, SequenceFileOutputFormat.class, GzipCodec.class);
|
||||
outputPath, Text.class, Text.class, SequenceFileOutputFormat.class, BZip2Codec.class);
|
||||
|
||||
});
|
||||
}
|
||||
|
|
|
|||
|
|
@ -28,7 +28,13 @@
|
|||
"paramLongName": "dataciteInputPath",
|
||||
"paramDescription": "the path to get the input data from Datacite",
|
||||
"paramRequired": true
|
||||
},
|
||||
},{
|
||||
"paramName": "wip",
|
||||
"paramLongName": "webCrawlInputPath",
|
||||
"paramDescription": "the path to get the input data from Web Crawl",
|
||||
"paramRequired": true
|
||||
}
|
||||
,
|
||||
{
|
||||
"paramName": "o",
|
||||
"paramLongName": "outputPath",
|
||||
|
|
|
|||
|
|
@ -35,5 +35,6 @@ crossrefInputPath=/data/bip-affiliations/crossref-data.json
|
|||
pubmedInputPath=/data/bip-affiliations/pubmed-data.json
|
||||
openapcInputPath=/data/bip-affiliations/openapc-data.json
|
||||
dataciteInputPath=/data/bip-affiliations/datacite-data.json
|
||||
webCrawlInputPath=/data/bip-affiliations/webCrawl/
|
||||
|
||||
outputPath=/tmp/crossref-affiliations-output-v5
|
||||
|
|
|
|||
|
|
@ -17,6 +17,10 @@
|
|||
<name>dataciteInputPath</name>
|
||||
<description>the path where to find the inferred affiliation relations from Datacite</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>webCrawlInputPath</name>
|
||||
<description>the path where to find the inferred affiliation relations from webCrawl</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>outputPath</name>
|
||||
<description>the path where to store the actionset</description>
|
||||
|
|
@ -112,7 +116,7 @@
|
|||
<arg>--pubmedInputPath</arg><arg>${pubmedInputPath}</arg>
|
||||
<arg>--openapcInputPath</arg><arg>${openapcInputPath}</arg>
|
||||
<arg>--dataciteInputPath</arg><arg>${dataciteInputPath}</arg>
|
||||
|
||||
<arg>--webCrawlInputPath</arg><arg>${webCrawlInputPath}</arg>
|
||||
<arg>--outputPath</arg><arg>${outputPath}</arg>
|
||||
</spark>
|
||||
<ok to="End"/>
|
||||
|
|
|
|||
|
|
@ -88,6 +88,7 @@ public class PrepareAffiliationRelationsTest {
|
|||
"-pubmedInputPath", crossrefAffiliationRelationPath,
|
||||
"-openapcInputPath", crossrefAffiliationRelationPath,
|
||||
"-dataciteInputPath", crossrefAffiliationRelationPath,
|
||||
"-webCrawlInputPath", crossrefAffiliationRelationPath,
|
||||
"-outputPath", outputPath
|
||||
});
|
||||
|
||||
|
|
@ -104,7 +105,7 @@ public class PrepareAffiliationRelationsTest {
|
|||
// );
|
||||
// }
|
||||
// count the number of relations
|
||||
assertEquals(80, tmp.count());
|
||||
assertEquals(120, tmp.count());
|
||||
|
||||
Dataset<Relation> dataset = spark.createDataset(tmp.rdd(), Encoders.bean(Relation.class));
|
||||
dataset.createOrReplaceTempView("result");
|
||||
|
|
@ -115,7 +116,7 @@ public class PrepareAffiliationRelationsTest {
|
|||
// verify that we have equal number of bi-directional relations
|
||||
Assertions
|
||||
.assertEquals(
|
||||
40, execVerification
|
||||
60, execVerification
|
||||
.filter(
|
||||
"relClass='" + ModelConstants.HAS_AUTHOR_INSTITUTION + "'")
|
||||
.collectAsList()
|
||||
|
|
@ -123,7 +124,7 @@ public class PrepareAffiliationRelationsTest {
|
|||
|
||||
Assertions
|
||||
.assertEquals(
|
||||
40, execVerification
|
||||
60, execVerification
|
||||
.filter(
|
||||
"relClass='" + ModelConstants.IS_AUTHOR_INSTITUTION_OF + "'")
|
||||
.collectAsList()
|
||||
|
|
|
|||
|
|
@ -4,4 +4,6 @@
|
|||
{"DOI":"10.1142\/s021821650200186x","Matchings":[{"RORid":"https:\/\/ror.org\/035xkbk20","Confidence":1},{"RORid":"https:\/\/ror.org\/05apxxy63","Confidence":1}]}
|
||||
{"DOI":"10.1061\/(asce)0733-9372(2002)128:7(575)","Matchings":[{"RORid":"https:\/\/ror.org\/04j198w64","Confidence":0.82}]}
|
||||
{"DOI":"10.1061\/(asce)0733-9372(2002)128:7(588)","Matchings":[{"RORid":"https:\/\/ror.org\/03m8km719","Confidence":0.8660254038},{"RORid":"https:\/\/ror.org\/02aze4h65","Confidence":0.87}]}
|
||||
{"DOI":"10.1161\/hy0202.103001","Matchings":[{"RORid":"https:\/\/ror.org\/057xtrt18","Confidence":0.7071067812}]}
|
||||
{"DOI":"10.1161\/hy0202.103001","Matchings":[{"RORid":"https:\/\/ror.org\/057xtrt18","Confidence":0.7071067812}]}
|
||||
{"DOI": "10.1080/13669877.2015.1042504", "Matchings": [{"Confidence": 1.0, "RORid": "https://ror.org/03265fv13"}]}
|
||||
{"DOI": "10.1007/3-540-47984-8_14", "Matchings": [{"Confidence": 1.0, "RORid": "https://ror.org/00a0n9e72"}]}
|
||||
Loading…
Reference in New Issue