Upgrade de.gwdg.metadataqa:metadata-qa-api from 0.9.4 to 0.9.5 #497

pkiraly · Aug 21, 2024 · ad323cb · ad323cb
1 parent 668af5c
commit ad323cb
Show file tree

Hide file tree

Showing 8 changed files with 28 additions and 4 deletions.
diff --git a/README.md b/README.md
@@ -2166,7 +2166,7 @@ Parameters:
 * `-C <file>`, `--shaclConfigurationFile <file>`: specify the SHACL like configuration file
 * `-O <file>`, `--shaclOutputFile <file>`: output file (default: `shacl4bib.csv`)
 * `-P <type>`, `--shaclOutputType <type>`: specify what the output files should contain. Possible values:
-  * `STATUS`: status only, where the following values appear:
+  * `STATUS`: status only (default), where the following values appear:
     * `1` the criteria met,
     * `0` the criteria have not met,
     * `NA`: the data element is not available in the record),

diff --git a/pom.xml b/pom.xml
@@ -110,7 +110,7 @@
 		<dependency>
 			<groupId>de.gwdg.metadataqa</groupId>
 			<artifactId>metadata-qa-api</artifactId>
-			<version>0.9.4</version>
+			<version>0.9.5</version>
 		</dependency>
 		<dependency>
 			<groupId>junit</groupId>

diff --git a/scripts/shacl4bib/shacl4bib.R b/scripts/shacl4bib/shacl4bib.R
@@ -1,3 +1,8 @@
+#' it creates a statistics from the status columns counting the number
+#' of distinct values from all possible values (that are 0, 1, NA).
+#' The header of the file are
+# id,0,1,NA
+# the rows are the individual rules identified by their IDs
 library(tidyverse)
 
 args = commandArgs(trailingOnly=TRUE)

diff --git a/src/main/java/de/gwdg/metadataqa/marc/cli/utils/MarcSpecSelector.java b/src/main/java/de/gwdg/metadataqa/marc/cli/utils/MarcSpecSelector.java
@@ -1,5 +1,6 @@
 package de.gwdg.metadataqa.marc.cli.utils;
 
+import de.gwdg.metadataqa.api.json.DataElement;
 import de.gwdg.metadataqa.api.model.XmlFieldInstance;
 import de.gwdg.metadataqa.marc.dao.record.BibliographicRecord;
 import de.gwdg.metadataqa.marc.utils.marcspec.legacy.MarcSpec;
@@ -20,6 +21,10 @@ public List<XmlFieldInstance> get(String path) {
     return transformTags(extract(path));
   }
 
+  public List<XmlFieldInstance> get(DataElement dataElement) {
+    return get(dataElement.getPath());
+  }
+
   public List<String> extract(String path) {
     return record.select(getMarcSpec(path));
   }

diff --git a/src/main/java/de/gwdg/metadataqa/marc/cli/utils/PicaPathSelector.java b/src/main/java/de/gwdg/metadataqa/marc/cli/utils/PicaPathSelector.java
@@ -1,5 +1,6 @@
 package de.gwdg.metadataqa.marc.cli.utils;
 
+import de.gwdg.metadataqa.api.json.DataElement;
 import de.gwdg.metadataqa.api.model.XmlFieldInstance;
 import de.gwdg.metadataqa.marc.dao.record.BibliographicRecord;
 import de.gwdg.metadataqa.marc.utils.pica.path.PicaSpec;
@@ -16,6 +17,10 @@ public List<XmlFieldInstance> get(String path) {
     return transformTags(extract(path));
   }
 
+  public List<XmlFieldInstance> get(DataElement dataElement) {
+    return get(dataElement.getPath());
+  }
+
   public List<String> extract(String path) {
     return record.select(new PicaSpec(path));
   }

diff --git a/src/test/java/de/gwdg/metadataqa/marc/cli/CompletenessTest.java b/src/test/java/de/gwdg/metadataqa/marc/cli/CompletenessTest.java
@@ -341,7 +341,7 @@ public void completeness_pica_groupBy_file() throws Exception {
         assertTrue(line.contains("\"pica\":true,"));
         assertTrue(line.contains("\"replacementInControlFields\":null,"));
         assertTrue(line.contains("\"marc21\":false,"));
-        assertTrue(line.contains("\"mqaf.version\":\"0.9.4\","));
+        assertTrue(line.contains("\"mqaf.version\":\"0.9.5\","));
         assertTrue(line.contains("\"qa-catalogue.version\":\"0.8.0-SNAPSHOT\""));
         assertTrue(line.contains("\"duration\":\"00:00:00\""));
         assertTrue(line.contains("\"numberOfprocessedRecords\":10"));

diff --git a/src/test/java/de/gwdg/metadataqa/marc/cli/ValidatorCliTest.java b/src/test/java/de/gwdg/metadataqa/marc/cli/ValidatorCliTest.java
@@ -247,7 +247,7 @@ public void validate_pica_groupBy() throws Exception {
         assertTrue(line.contains("\"pica\":true,"));
         assertTrue(line.contains("\"replacementInControlFields\":null,"));
         assertTrue(line.contains("\"marc21\":false,"));
-        assertTrue(line.contains("\"mqaf.version\":\"0.9.4\","));
+        assertTrue(line.contains("\"mqaf.version\":\"0.9.5\","));
         assertTrue(line.contains("\"qa-catalogue.version\":\"0.8.0-SNAPSHOT\""));
         assertTrue(line.contains("\"duration\":\"00:00:00\""));
         assertTrue(line.contains("\"numberOfprocessedRecords\":10"));

diff --git a/validate b/validate
@@ -63,15 +63,20 @@ fi
 SHORT_OPTIONS="m:hnl:o:i:d:qabpxyt:rz:v:f:s:g:1:2:u:j:w:k:c:e:3:4:G:SHF:R:WTI:"
 LONG_OPTIONS="marcVersion:,help,nolog,limit:,offset:,id:,defaultRecordType:,fixAlephseq,fixAlma,fixKbr,alephseq,marcxml,lineSeparated,outputDir:,trimId,ignorableFields:,ignorableRecords:,marcFormat:,dataSource:,defaultEncoding:,alephseqLineType:,picaIdField:,picaSubfieldSeparator:,picaSchemaFile:,schemaType:,picaRecordType:,allowableRecords:,groupBy:,groupListFile:,solrForScoresUrl:,summaryFileName:,summary,details,detailsFileName:,format:,emptyLargeCollectors,collectAllErrors,ignorableIssueTypes:"
 
+echo "@: ${@}"
+
 GETOPT=$(getopt \
   -o ${SHORT_OPTIONS} \
   --long ${LONG_OPTIONS} \
   -n ${ME} -- "$@")
 eval set -- "${GETOPT}"
 
+echo "GETOPT: ${GETOPT}"
+
 PARAMS=""
 HELP=0
 while true ; do
+  echo "process $1"
   case "$1" in
     -m|--marcVersion)            PARAMS="$PARAMS --marcVersion $2" ;           shift 2 ;;
     -h|--help)                   PARAMS="$PARAMS --help" ; HELP=1;             shift   ;;
@@ -122,5 +127,9 @@ fi
 
 CMD="/usr/bin/java -Xmx8g -cp $JAR de.gwdg.metadataqa.marc.cli.ValidatorCli"
 
+echo 'CMD: ' $CMD
+echo 'PARAMS: ' $PARAMS
+echo 'REST: ' "$@"
 echo $CMD $PARAMS "$@"
+exit
 $CMD $PARAMS "$@"