From 27af5122d2d7f7d6d549dc23a5bf79a0c0030d1c Mon Sep 17 00:00:00 2001 From: Alessia Bardi Date: Mon, 12 Sep 2022 14:25:23 +0200 Subject: [PATCH] logs for non well formed XML files --- .../raw/AbstractMdRecordToOafMapper.java | 62 +++++++++------- .../dnetlib/dhp/oa/graph/raw/MappersTest.java | 12 ++++ .../dhp/oa/graph/raw/oaf_notwellformed.xml | 70 +++++++++++++++++++ .../src/test/resources/log4j.properties | 8 +++ 4 files changed, 125 insertions(+), 27 deletions(-) create mode 100644 dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/oaf_notwellformed.xml create mode 100644 dhp-workflows/dhp-graph-mapper/src/test/resources/log4j.properties diff --git a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java index a05faab80..e263cffa8 100644 --- a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java +++ b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java @@ -22,6 +22,8 @@ import java.util.*; import org.apache.commons.lang3.StringUtils; import org.dom4j.*; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; import com.google.common.collect.Lists; import com.google.common.collect.Sets; @@ -77,6 +79,8 @@ public abstract class AbstractMdRecordToOafMapper { protected static final Map nsContext = new HashMap<>(); + private static final Logger log = LoggerFactory.getLogger(AbstractMdRecordToOafMapper.class); + static { nsContext.put("dr", "http://www.driver-repository.eu/namespace/dr"); nsContext.put("dri", "http://www.driver-repository.eu/namespace/dri"); @@ -106,37 +110,41 @@ public abstract class AbstractMdRecordToOafMapper { public List processMdRecord(final String xml) throws DocumentException { DocumentFactory.getInstance().setXPathNamespaceURIs(nsContext); + try { + final Document doc = DocumentHelper + .parseText( + xml + .replaceAll(DATACITE_SCHEMA_KERNEL_4, DATACITE_SCHEMA_KERNEL_3) + .replaceAll(DATACITE_SCHEMA_KERNEL_4_SLASH, DATACITE_SCHEMA_KERNEL_3) + .replaceAll(DATACITE_SCHEMA_KERNEL_3_SLASH, DATACITE_SCHEMA_KERNEL_3)); - final Document doc = DocumentHelper - .parseText( - xml - .replaceAll(DATACITE_SCHEMA_KERNEL_4, DATACITE_SCHEMA_KERNEL_3) - .replaceAll(DATACITE_SCHEMA_KERNEL_4_SLASH, DATACITE_SCHEMA_KERNEL_3) - .replaceAll(DATACITE_SCHEMA_KERNEL_3_SLASH, DATACITE_SCHEMA_KERNEL_3)); + final KeyValue collectedFrom = getProvenanceDatasource( + doc, "//oaf:collectedFrom/@id", "//oaf:collectedFrom/@name"); - final KeyValue collectedFrom = getProvenanceDatasource( - doc, "//oaf:collectedFrom/@id", "//oaf:collectedFrom/@name"); + if (collectedFrom == null) { + return Lists.newArrayList(); + } - if (collectedFrom == null) { - return Lists.newArrayList(); + final KeyValue hostedBy = StringUtils.isBlank(doc.valueOf("//oaf:hostedBy/@id")) + ? collectedFrom + : getProvenanceDatasource(doc, "//oaf:hostedBy/@id", "//oaf:hostedBy/@name"); + + if (hostedBy == null) { + return Lists.newArrayList(); + } + + final DataInfo info = prepareDataInfo(doc, invisible); + final long lastUpdateTimestamp = new Date().getTime(); + + final List instances = prepareInstances(doc, info, collectedFrom, hostedBy); + + final String type = getResultType(doc, instances); + + return createOafs(doc, type, instances, collectedFrom, info, lastUpdateTimestamp); + } catch (DocumentException e) { + log.error("Error with record:\n" + xml); + throw e; } - - final KeyValue hostedBy = StringUtils.isBlank(doc.valueOf("//oaf:hostedBy/@id")) - ? collectedFrom - : getProvenanceDatasource(doc, "//oaf:hostedBy/@id", "//oaf:hostedBy/@name"); - - if (hostedBy == null) { - return Lists.newArrayList(); - } - - final DataInfo info = prepareDataInfo(doc, invisible); - final long lastUpdateTimestamp = new Date().getTime(); - - final List instances = prepareInstances(doc, info, collectedFrom, hostedBy); - - final String type = getResultType(doc, instances); - - return createOafs(doc, type, instances, collectedFrom, info, lastUpdateTimestamp); } protected String getResultType(final Document doc, final List instances) { diff --git a/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java b/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java index b75f626d6..c4e34a9a8 100644 --- a/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java +++ b/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java @@ -12,6 +12,7 @@ import java.util.Objects; import java.util.Optional; import java.util.stream.Collectors; +import kotlin.jvm.Throws; import org.apache.commons.io.IOUtils; import org.apache.commons.lang3.StringUtils; import org.dom4j.DocumentException; @@ -926,6 +927,17 @@ class MappersTest { // assertTrue(StringUtils.isNotBlank(p.getTitle().get(0).getValue())); } + @Test + void testNotWellFormed() throws IOException, DocumentException { + final String xml = IOUtils + .toString(Objects.requireNonNull(getClass().getResourceAsStream("oaf_notwellformed.xml"))); + final DocumentException generalEx = new DocumentException(); + + DocumentException exception = assertThrows(DocumentException.class, () -> { + new OafToOafMapper(vocs, false, true).processMdRecord(xml); + }); + } + private void assertValidId(final String id) { // System.out.println(id); diff --git a/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/oaf_notwellformed.xml b/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/oaf_notwellformed.xml new file mode 100644 index 000000000..09384054e --- /dev/null +++ b/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/oaf_notwellformed.xml @@ -0,0 +1,70 @@ + + +
+ jairo_______::000012e58ed836576ef2a0d38b0f726f + oai:irdb.nii.ac.jp:01221:0000010198 + + + + + + 2021-05-10T11:31:09.424Z + 2021-06-03T01:45:42.536Z + jairo_______ +
+ + 多項式GCDを用いた復号法に関する研究 + 上原, 剛 + 甲斐, 博 + 野田, 松太郎 + application/pdf + http://hdl.handle.net/2433/25934 + jpn + 京都大学数理解析研究所 + 410 + Departmental Bulletin Paper + 0014 + 2004-10-01 + + openaire____::554c7c2873 + OPEN + + + 2433/25934 + AN00061013 + http://hdl.handle.net/2433/25934 + http://repository.kulib.kyoto-u.ac.jp/dspace/bitstream/2433/25934/1/1395-16.pdf + 数理解析研究所講究録 + + + + + https%3A%2F%2Firdb.nii.ac.jp%2Foai + oai:irdb.nii.ac.jp:01221:0000010198 + 2021-04-13T13:36:29Z + + + http://repository.kulib.kyoto-u.ac.jp/dspace-oai/request + oai:repository.kulib.kyoto-u.ac.jp:2433/25934 + 2012-07-12T14:15:41Z + http://irdb.nii.ac.jp/oai + + + + + false + false + 0.9 + + + + +
\ No newline at end of file diff --git a/dhp-workflows/dhp-graph-mapper/src/test/resources/log4j.properties b/dhp-workflows/dhp-graph-mapper/src/test/resources/log4j.properties new file mode 100644 index 000000000..71255bb77 --- /dev/null +++ b/dhp-workflows/dhp-graph-mapper/src/test/resources/log4j.properties @@ -0,0 +1,8 @@ +# Root logger option +log4j.rootLogger=DEBUG, stdout + +# Direct log messages to stdout +log4j.appender.stdout=org.apache.log4j.ConsoleAppender +log4j.appender.stdout.Target=System.out +log4j.appender.stdout.layout=org.apache.log4j.PatternLayout +log4j.appender.stdout.layout.ConversionPattern=%d{yyyy-MM-dd HH:mm:ss} %-5p %c{1}:%L - %m%n \ No newline at end of file