Closed nikit-srivastava closed 2 years ago
Huh, can you show me the json line where it happens? I wasn't aware of a new field (/CC @xixi019 )
I do not know if this is it but these, Xi, should also be removed from https://raw.githubusercontent.com/KGQA/QALD_10/main/data/qald_10/qald_10.json
{
"id": 42,
"aggregation": false,
"question": [
{
"language": "en",
"string": "How long did the Han dynasty last?"
},
{
"language": "zh-cn",
"string": "汉朝持续了多长时间?"
},
{
"language": "de",
"string": "Wie lange bestand die Han-Dynastie?"
},
{
"language": "ru",
"string": "Как долго просуществовала династия Хань ?"
}
],
"answers": [
{
"head": {
"vars": [
"result"
]
},
"results": {
"bindings": [
{}
]
}
}
],
"query": {
"sparql": "SELECT ?result WHERE {wd:Q7209 wdt:P571 ?st; wdt:P576 ?et. BIND((?et - ?st) AS ?result)}"
}
},
Huh, can you show me the json line where it happens? I wasn't aware of a new field (/CC @xixi019 )
For example (do not mind the wrongly predicted sparql ;) ):
{
"id": 71,
"question": [
{
"language": "en",
"string": "How many students does the Free University of Amsterdam have?"
}
],
"query": {
"sparql": "SELECT (COUNT(DISTINCT ?selvar) AS ?c) WHERE { <http://dbpedia.org/resource/Amsterdam_University_College> <http://dbpedia.org/ontology/country> ?selvar . }"
},
"answers": [
{
"head": {
"link": [],
"vars": [
"c"
]
},
"results": {
"distinct": false,
"ordered": true,
"bindings": [
{
"c": {
"type": "typed-literal",
"datatype": "http://www.w3.org/2001/XMLSchema#integer",
"value": "1"
}
}
]
}
}
]
}
Ah, that happens in the training file (/CC @Perevalov )I would say, we can safely ignore it. Aleksandr, what do you think?
Ah, that happens in the training file (/CC @Perevalov )I would say, we can safely ignore it. Aleksandr, what do you think?
From the example by @nikit91, I understood that the:
"distinct": false,
"ordered": true,
and
"datatype": "http://www.w3.org/2001/XMLSchema#integer",
is something that should be ignored?
If yes, then I agree. It does not affect anything.
Perfect, so let's go ahead here and resolve the QALD-10 test issue (empty answers in the repo).
It would be nice to have this fix for GERBIL QA. :wink: Do you foresee to release and deploy a new version of qa.commons, soon? I would also be fine with a SNAPSHOT version. Where will you deploy it? On the AKSW archiva or somewhere else? :thinking:
Erm...no, I am not foreseeing working on this repo anytime soon actually. But as far as I know, you could also mvn deploy it to archiva (honestly, I do not even have the keys anymore due to laptop changes)
Hi @RicardoUsbeck, Is it fine if we just ignore the unknown fields in the QALD format?
I recently faced a problem where the query results from the dbpedia endpoint introduced new fields which were not part of the QALD format previously. However, these newly introduce fields do not affect the existing expected functionality and can be ignored. Below is the stacktrace for the problem:
What is your opinion of this?