Blank Nodes in YARRRML trigger TypeError: argument of type 'NoneType' is not iterable

Stiksels commented 7 months ago

Describe the bug Following the YARRRML subjects spec, I use blank nodes with s: null . This is not supported by yatter

To Reproduce Steps to reproduce the behavior (and resources): with the files below, run the yatter library python3 -m yatter -i yarrrml/yatter-test.yml -o rml/yatter-test_rml.ttl

yatter-test.yml

# Define namespace prefixes for commonly used ontologies
prefixes:
  exp: "http://example.org/participant/"
  ex: "http://example.org/#"
  rdf: "http://www.w3.org/1999/02/22-rdf-syntax-ns#"
  rdfs: "http://www.w3.org/2000/01/rdf-schema#"
  xsd: "http://www.w3.org/2001/XMLSchema#"
  dcterms: "http://purl.org/dc/terms/"
  locn: "http://www.w3.org/ns/locn#"
  grel: "http://users.ugent.be/~bjdmeest/function/grel.ttl#"
  skos: "http://www.w3.org/2004/02/skos/core#"
  wegwijs: "https://data.vlaanderen.be/id/organisatie/"
  adms: "http://www.w3.org/ns/adms#"

# Define the mapping rules
mappings:

  # Mapping for 'participants' data
  participants:
    # Define sources to extract data from JSON
    sources:
      - ['yatter-test.csv~csv']
    # Define subject mapping
    s: exp:$(participantUuid)
    # Define predicate-object mappings
    po:
      - [a, ex:Participant]
      - [dcterms:created, $(created), xsd:dateTime]
      - [dcterms:modified, $(modified), xsd:dateTime]
      - p: adms:identifier
        o: 
          type: blank
          mapping: identificator
          condition:
            function: equal
            parameters:
              - [str1, $(participantUuid), s]
              - [str2, $(participantUuid), o]

  # Mapping for 'identificator' data
  identificator:
    sources:
      - ['yatter-test.csv~csv']
    # Define subject as null, indicating a blank node will be generated
    s: null
    # Define predicate-object mappings
    po:
      - [a, adms:Identifier] 
      - [skos:notation, exp:$(participantUuid), xsd:anyURI]
      - [dcterms:creator, wegwijs:OVO026559~iri]

yatter-test.csv

participantUuid,created,modified,birthYear,placeHolderStatSect
493b78b8-efa7-45b8-af5f-ca08ec90179e,2023-06-06T15:17:32.954+0000,2023-06-06T15:19:14.358+0000,1964,Wilrijk 
ea9ba26f-e283-42a2-b57a-55ea3f5571ff,2022-12-28T14:46:03.682+0000,2022-12-28T14:48:23.17+0000,1957,Marcinelle
e0f5d6b0-984d-4192-b00f-8a07d2189712,2023-06-06T17:36:44.39+0000,2023-06-06T17:36:44.396+0000,1963,Zingem
20a15803-ca0e-42f9-9abe-93dc5d593a31,2022-12-16T11:39:39.481+0000,2022-12-16T11:43:58.073+0000,1951,Gent

Current behavior TypeError stacktrace

Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/yatter/__main__.py", line 68, in <module>
    mapping_content = translate(mapping_data, mapping_format)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/yatter/__init__.py", line 15, in translate
    list_initial_sources = get_initial_sources(yarrrml_data)
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/yatter/source.py", line 9, in get_initial_sources
    if YARRRML_SOURCES in data:
       ^^^^^^^^^^^^^^^^^^^^^^^
TypeError: argument of type 'NoneType' is not iterable

Expected behavior an RML file is produced like the one by YARRRML's matey: output.rml.ttl.txt

output turtle output (1).txt

dachafra commented 7 months ago

Hi @Stiksels, As there is no consensus in the W3C CG about supporting BN without data references, I would not consider implementing this feature for now. However, I can provide support for BN with data reference on the subject. As there is no example in the spec, I will do it similarly as it's indicated for ObjectMaps, as both (OM and SM) are term maps.

It will be something like:

identificator:
    sources:
      - ['yatter-test.csv~csv']
    # Define subject as null, indicating a blank node will be generated
    s: 
      value: $(participantUuid) # could be template, reference or constant
      type: blank
    # Define predicate-object mappings
    po:
      - [a, adms:Identifier] 
      - [skos:notation, exp:$(participantUuid), xsd:anyURI]
      - [dcterms:creator, wegwijs:OVO026559~iri]

dachafra commented 7 months ago

BN are now supported in subject and object, please see the test case as an example https://github.com/oeg-upm/yatter/blob/main/test/r2rml/YARRRMLTC-r2rml-0016/mapping.yml

oeg-upm / yatter

Blank Nodes in YARRRML trigger TypeError: argument of type 'NoneType' is not iterable #69