BUG: JSON path not finding nested object that contain empty lists

[Manoe-K]

What Happens?

When iterating over a list using JSON path, objects that happen to have an empty list as an attribute will not be matched.
(I am very not sure about how I define this issue in the above phrase sorry, but I tried to make an example as clear as possible.)

In my example, I'm iterating over a list of persons who may have a list of friends. Those friends may have nicknames.
If I try to create a triple of the form "<person_A> foaf:knows <person_A's_friend>". The triple will not be created if that person's friend doesn't have nicknames. To be precise, it will not work if it has an empty list as value of one of its attributes. Even though said attribute may not even be used in the mapping.

I am doing that example on the python library of morph-kgc.

To Reproduce

data:

{
    "persons": [
        {
            "name": "Alice",
            "friends": [
                {
                    "name": "Bob",
                    "nicknames": ["Bobby"]
                },
                {
                    "name": "Curtis",
                    "nicknames": []
                }
            ]
        }
    ]
}

mapping:

@prefix ql: <http://semweb.mmlab.be/ns/ql#> .
@prefix rml: <http://semweb.mmlab.be/ns/rml#> .
@prefix rr: <http://www.w3.org/ns/r2rml#> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .


<#example> a rr:TriplesMap;
	
	rml:logicalSource [
		rml:source "data.json";
		rml:referenceFormulation ql:JSONPath;
		rml:iterator "$.persons[*]";
	];

	rr:subjectMap [
 		rr:template "http://example.com/{name}";
 	];
    
	rr:predicateObjectMap [
		rr:predicate foaf:knows;
  		rr:objectMap [ rml:reference "friends.name"; ];
  	];
.

python code:

import morph_kgc

config = """
            [DataSource1]
            mappings=mapping.ttl
         """

g = morph_kgc.materialize_oxigraph(config)
g.dump(output='data.ttl', mime_type='text/turtle')

results:

<http://example.com/Alice> <http://xmlns.com/foaf/0.1/knows> "Bob" .

Expected results:

<http://example.com/Alice> <http://xmlns.com/foaf/0.1/knows> "Bob" .
<http://example.com/Alice> <http://xmlns.com/foaf/0.1/knows> "Curtis" .

Environment:

• OS: Ventura 13.2
• Python version: 3.10.8
• Morph-KGC version: 2.1.1

This issue happens while using version 2.1.1, with that version my current example returns 1 triple instead of my expected 2.
I first had another issue using version 2.3.1, with that version my current example returns 0 triple.
But that second issue might just be issue 132.

[arenas-guerrero-julian]

Hi @Manoe-K ,

Thanks for your detailed issue.

There is the following problem when converting from json to dataframe:

JSON:

[{'ID': 10, 'Sport': 100, 'Name': 'Venus Williams'}, {'ID': 20, 'Name': 'Demi Moore'}]

DataFrame

   ID  Sport            Name
0  10  100.0  Venus Williams
1  20    NaN      Demi Moore

Because of the missing values, pandas uses a NaN, and the integers are read as doubles. This is related to your issue because morph-kgc has a hack to fix this that prevents fixing this issue.

I am checking how to make everything work.