Python rdflib解析JSON-LD为何仅提取顶层三元组？

我用Python的rdflib库解析JSON-LD数据，但解析后只输出了顶层的rdf:type三元组，像WebPage的name这类属性都没被解析出来。

我的代码：

from rdflib import Dataset
from rdflib.namespace import Namespace, NamespaceManager

local_input="""
[
    {
        "@context": "https://schema.org/"
    },
    {
        "@type": "WebPage",
        "@id": "https://example.com/glossary/term/",
        "name": "My Glossary Term",
        "abstract": "Just my glossary Term.",
        "datePublished": "2024-03-08T07:52:13+02:00",
        "dateModified": "2024-03-08T14:54:13+02:00",
        "url": "https://example.com/glossary/term/",
        "author": {
            "@type": "Person",
            "@id": "https://example.com/",
            "name": "John Doe"
        },
        "about": [
            {
                "@type": "DefinedTerm",
                "@id": "https://example.com/glossary/term/#definedTerm"
            }
        ]
    },
    {
        "@type": "DefinedTerm",
        "@id": "https://example.com/glossary/term/#definedTerm",
        "name": "My Term",
        "description": "Just my Term",
        "inDefinedTermSet": {
            "@type": "DefinedTermSet",
            "@id": "https://example.com/glossary/#definedTermSet"
        }
    }
]
"""

SCH = Namespace('https://schema.org/')
namespace_manager = NamespaceManager(Dataset(), bind_namespaces='none')
namespace_manager.bind('', SCH, override=True)
g = Dataset()
g.namespace_manager = namespace_manager
g.parse(data=local_input, format='json-ld', publicID="http://schema.org/")

print(len(g))

import pprint
for stmt in g:
    pprint.pprint(stmt)

输出结果：

2
(rdflib.term.URIRef('https://example.com/glossary/term/'),
 rdflib.term.URIRef('http://www.w3.org/1999/02/22-rdf-syntax-ns#type'),
 rdflib.term.URIRef('http://schema.org/WebPage'),
 rdflib.term.URIRef('urn:x-rdflib:default'))
(rdflib.term.URIRef('https://example.com/glossary/term/#definedTerm'),
 rdflib.term.URIRef('http://www.w3.org/1999/02/22-rdf-syntax-ns#type'),
 rdflib.term.URIRef('http://schema.org/DefinedTerm'),
 rdflib.term.URIRef('urn:x-rdflib:default'))

核心问题1：JSON-LD结构不规范 #

把@context作为数组的单独元素，导致后续节点无法正确关联上下文。JSON-LD规范中，数组形式的图应该将@context放在包含@graph的顶层对象中，或者每个节点都携带@context。

核心问题2：命名空间与publicID不匹配 #

代码中publicID设为http://schema.org/，但schema.org的标准命名空间是https://schema.org/，同时命名空间绑定和解析时的上下文不统一，导致属性无法被正确映射。

修复后的代码 #

from rdflib import Dataset
from rdflib.namespace import Namespace, NamespaceManager

# 修正JSON-LD结构：用@graph包裹节点，@context放在顶层
local_input="""
{
    "@context": "https://schema.org/",
    "@graph": [
        {
            "@type": "WebPage",
            "@id": "https://example.com/glossary/term/",
            "name": "My Glossary Term",
            "abstract": "Just my glossary Term.",
            "datePublished": "2024-03-08T07:52:13+02:00",
            "dateModified": "2024-03-08T14:54:13+02:00",
            "url": "https://example.com/glossary/term/",
            "author": {
                "@type": "Person",
                "@id": "https://example.com/",
                "name": "John Doe"
            },
            "about": [
                {
                    "@type": "DefinedTerm",
                    "@id": "https://example.com/glossary/term/#definedTerm"
                }
            ]
        },
        {
            "@type": "DefinedTerm",
            "@id": "https://example.com/glossary/term/#definedTerm",
            "name": "My Term",
            "description": "Just my Term",
            "inDefinedTermSet": {
                "@type": "DefinedTermSet",
                "@id": "https://example.com/glossary/#definedTermSet"
            }
        }
    ]
}
"""

SCH = Namespace('https://schema.org/')
namespace_manager = NamespaceManager(Dataset(), bind_namespaces='none')
namespace_manager.bind('', SCH, override=True)
g = Dataset()
g.namespace_manager = namespace_manager
# 去掉错误的publicID，或设置为正确的https地址
g.parse(data=local_input, format='json-ld')

print(len(g))

import pprint
for stmt in g:
    pprint.pprint(stmt)

修复后效果 #

运行代码会输出所有三元组，包括WebPage的name、abstract，Author的name等属性，总长度不再是2，而是会包含所有节点的完整属性。

内容的提问来源于stack exchange，提问作者Kaj Kandler