18  Shacl4Bib

since v0.7.0. Note: This is an experimental feature.

The Shapes Constraint Language (SHACL) is a formal language for validating Resource Description Framework (RDF) graphs against a set of conditions (expressed also in RDF). Following this idea and implementing a subset of the language, the Metadata Quality Assessment Framework provides a mechanism to define SHACL-like rules for data sources in non-RDF based formats, such as XML, CSV and JSON (SHACL validates only RDF graphs). Shacl4Bib is the extension enabling the validation of bibliographic records. The rules can be defined either with YAML or JSON configuration files or with Java code. SCHACL uses RDF notation to specify or “address” the data element about which the constraints are set. Shacl4Bib supports Carsten Klee’s MARCspec for MARC records, and PICApath for PICA. You can find more information and full definition of the implemented subset of SHACL here: Defining schema with a configuration file

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 (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),
  • SCORE: score only. Its value is calculated the following way:
    • if the criteria met it returns the value of successScore property (or 0 if no such property)
    • if the criteria have not met, it returns the value of failureScore property (or 0 if no such property)
  • BOTH: both status and score

Here is a simple example for setting up rules against a MARC subfield:

format: MARC
fields:
- name: 040$a
  path: 040$a
  rules:
  - id: 040$a.minCount
    minCount: 1
  - id: 040$a.pattern
    pattern: ^BE-KBR00

• format represents the format of the input data. It can be either MARC or PICA
• fields: the list of fields we would like to investigate. Since it is a YAMPL example, the - and indentation denotes child elements. Here there is only one child, so we analyse here a single subfield.
• name is how the data element is called within the rule set. It could be a machine or a human readable string.
• path is the “address” of the metadata element. It should be expressed in an addressing language such as MARCSpec or PICAPath (040$a contains the original cataloging agency)
• rules: the parent element of the set of rules. Here we have two rules.
• id: the identifier of the rule. This will be the header of the column in CSV, and it could be references elsewhere in the SHACL configuration file.
• mintCount: this specify the minimum number of instances of the data element in the record
• pattern: a regular expression which should match the values of all instances of the data element

The output contains an extra column, the record identifier, so it looks like something like this:

id,040$a.minCount,040$a.pattern
17529680,1,1
18212975,1,1
18216050,1,1
18184955,1,1
18184431,1,1
9550740,NA,NA
19551181,NA,NA
118592844,1,1
18592704,1,1
18592557,1,1