#1430

Committed 12 Mar 2025 12:23PM UTC coverage: 90.395% (-0.06%) from 90.457%

Build # #1430

Build Type

push

Committed by

Commit Message

Merge branch 'main' of github.com:pkiraly/qa-catalogue

Coverage Stats

36517 of 40397 relevant lines covered (90.4%)

0.9 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

78.26

/src/main/java/de/gwdg/metadataqa/marc/cli/utils/ignorablerecords/Marc21Filter.java

package de.gwdg.metadataqa.marc.cli.utils.ignorablerecords;

import de.gwdg.metadataqa.marc.MarcSubfield;
import de.gwdg.metadataqa.marc.dao.DataField;
import de.gwdg.metadataqa.marc.dao.record.BibliographicRecord;
import org.apache.commons.lang3.StringUtils;

import java.util.ArrayList;
import java.util.List;
import java.util.regex.Pattern;

public class Marc21Filter {

  protected List<Condition> conditions;

  protected void parseInput(String input) {
    if (StringUtils.isNotBlank(input)) {
      conditions = new ArrayList<>();
      for (String field : input.split(",")) {
        Condition f = parseFieldCondition(field);
        if (f != null)
          conditions.add(f);
      }
    }
  }

  public List<Condition> getConditions() {
    return conditions;
  }

  public boolean isEmpty() {
    return conditions == null || conditions.isEmpty();
  }

  /**
   * Parses the given string field condition which is in a format of "tag$subfield=value" (e.g. "001$0=123456").
   * @param field The field condition to parse
   * @return The parsed field condition, which is essentially a new DataField object
   */
  protected Condition parseFieldCondition(String field) {
    var pattern = Pattern.compile(
      "^(.{3})(?:\\$(.))?\\s?(=~|!=|!~|==|=\\^|=\\$|\\?|=|not in|in)\\s?'?(.*?)'?$"
    );
    var matcher = pattern.matcher(field);
    if (matcher.matches()) {
      String tag = matcher.group(1);
      String subfield = matcher.group(2);
      String operator = matcher.group(3);
      String value = matcher.group(4);
      try {
        return new Condition(tag, subfield, Operator.byCode(operator), value);
      } catch (Exception e) {
        throw new RuntimeException(e);
      }
    }
    return null;
  }

  /**
   * Checks if the given record meets the set of criteria. If only one of the criteria is met, the method returns
   * true. If none of the criteria is met, the method returns false.
   * @param marcRecord The record to check
   * @return True if the record meets the criteria, false otherwise
   */
  protected boolean met(BibliographicRecord marcRecord) {
    for (Condition condition : conditions) {
      List<DataField> recordFields = marcRecord.getDatafieldsByTag(condition.getTag());
      if (recordFields == null || recordFields.isEmpty())
        continue;
      boolean passed = metCondition(condition, recordFields);
      if (passed)
        return true;
    }
    return false;
  }

  /**
   * Checks if the given fields meet the given condition. If the condition is met, the method returns true.
   * @param condition Condition to check the record for, for now in the form of a DataField object
   * @param recordFields Fields of the record to check
   * @return True if the fields meet the condition, false otherwise
   */
  private boolean metCondition(Condition condition, List<DataField> recordFields) {
    for (DataField recordField : recordFields) {
      String code = condition.getSubfield();
      List<MarcSubfield> recordSubfields = recordField.getSubfield(code);

      if (recordSubfields == null || recordSubfields.isEmpty())
        continue;

      for (MarcSubfield recordSubfield : recordSubfields) {
        if (condition.isUseEqual()) {
          if (recordSubfield.getValue().equals(condition.getValue()))
            return condition.isNegate() ? false : true;
        } else if (condition.isUsePattern()) {
          var matcher = condition.getPattern().matcher(recordSubfield.getValue());
          if (matcher.find())
            return condition.isNegate() ? false : true;
        }
      }
    }
    return false;
  }
}

1	package de.gwdg.metadataqa.marc.cli.utils.ignorablerecords;
2
3	import de.gwdg.metadataqa.marc.MarcSubfield;
4	import de.gwdg.metadataqa.marc.dao.DataField;
5	import de.gwdg.metadataqa.marc.dao.record.BibliographicRecord;
6	import org.apache.commons.lang3.StringUtils;
7
8	import java.util.ArrayList;
9	import java.util.List;
10	import java.util.regex.Pattern;
11
12	public class Marc21Filter {	1✔
13
14	protected List<Condition> conditions;
15
16	protected void parseInput(String input) {
17	if (StringUtils.isNotBlank(input)) {	1✔
18	conditions = new ArrayList<>();	1✔
19	for (String field : input.split(",")) {	1✔
20	Condition f = parseFieldCondition(field);	1✔
21	if (f != null)	1✔
22	conditions.add(f);	1✔
23	}
24	}
25	}	1✔
26
27	public List<Condition> getConditions() {
28	return conditions;	1✔
29	}
30
31	public boolean isEmpty() {
32	return conditions == null \|\| conditions.isEmpty();	1✔
33	}
34
35	/**
36	* Parses the given string field condition which is in a format of "tag$subfield=value" (e.g. "001$0=123456").
37	* @param field The field condition to parse
38	* @return The parsed field condition, which is essentially a new DataField object
39	*/
40	protected Condition parseFieldCondition(String field) {
41	var pattern = Pattern.compile(	1✔
42	"^(.{3})(?:\\$(.))?\\s?(=~\|!=\|!~\|==\|=\\^\|=\\$\|\\?\|=\|not in\|in)\\s?'?(.*?)'?$"
43	);
44	var matcher = pattern.matcher(field);	1✔
45	if (matcher.matches()) {	1✔
46	String tag = matcher.group(1);	1✔
47	String subfield = matcher.group(2);	1✔
48	String operator = matcher.group(3);	1✔
49	String value = matcher.group(4);	1✔
50	try {
51	return new Condition(tag, subfield, Operator.byCode(operator), value);	1✔
52	} catch (Exception e) {	×
53	throw new RuntimeException(e);	×
54	}
55	}
56	return null;	×
57	}
58
59	/**
60	* Checks if the given record meets the set of criteria. If only one of the criteria is met, the method returns
61	* true. If none of the criteria is met, the method returns false.
62	* @param marcRecord The record to check
63	* @return True if the record meets the criteria, false otherwise
64	*/
65	protected boolean met(BibliographicRecord marcRecord) {
66	for (Condition condition : conditions) {	1✔
67	List<DataField> recordFields = marcRecord.getDatafieldsByTag(condition.getTag());	1✔
68	if (recordFields == null \|\| recordFields.isEmpty())	1✔
69	continue;	×
70	boolean passed = metCondition(condition, recordFields);	1✔
71	if (passed)	1✔
72	return true;	1✔
73	}	×
74	return false;	×
75	}
76
77	/**
78	* Checks if the given fields meet the given condition. If the condition is met, the method returns true.
79	* @param condition Condition to check the record for, for now in the form of a DataField object
80	* @param recordFields Fields of the record to check
81	* @return True if the fields meet the condition, false otherwise
82	*/
83	private boolean metCondition(Condition condition, List<DataField> recordFields) {
84	for (DataField recordField : recordFields) {	1✔
85	String code = condition.getSubfield();	1✔
86	List<MarcSubfield> recordSubfields = recordField.getSubfield(code);	1✔
87
88	if (recordSubfields == null \|\| recordSubfields.isEmpty())	1✔
89	continue;	×
90
91	for (MarcSubfield recordSubfield : recordSubfields) {	1✔
92	if (condition.isUseEqual()) {	1✔
93	if (recordSubfield.getValue().equals(condition.getValue()))	1✔
94	return condition.isNegate() ? false : true;	1✔
95	} else if (condition.isUsePattern()) {	1✔
96	var matcher = condition.getPattern().matcher(recordSubfield.getValue());	1✔
97	if (matcher.find())	1✔
98	return condition.isNegate() ? false : true;	1✔
99	}
100	}	×
101	}	×
102	return false;	×
103	}
104	}

pkiraly / metadata-qa-marc / #1430

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous