Hi all, I'm trying to get a better sense for what relatedness actually means (e.g., "/relatedness?node1=/c/en/cat&node2=/c/en/pet" returns a value of 0.583, but I'm not sure where that value comes from).
Is there a paper I can read somewhere that gives a clear definition?
The conceptnet seems to have 500 internal server error... when you would fix it ?
I have the entire english part of the conceptnet semantic written in binary - if anyone needs it, you need to be able to decode it using python and the python module pickle
Hello everyone,
I am developing a ConceptNet Datasource on the Triple Pattern Fragment Server (TPF). It allow us to make simple queries like the ones in the https://api.conceptnet.io/query API. But the coolest thing about having this datasource is that it makes it possible to use SPARQL queries using the Comunica query engine. The TPF is a RDF interface that by design handle the latency and downtime in the servers (problem reported here) by delegating greater complexity to the client side (see https://linkeddatafragments.org/).
Having said that, I'd like to ask two questions: (1) How difficult is it to add other parameters in the query API. For example, currently it is possible to filter using start
to reference a node in the start
field of an edge
, as well as end
and rel
, but we cannot filter by, for instance, dataset
and language
fields. (2) Is it possible to know the total number of edges given a filter like the ones mentioned above (e.g., totalEdges: 5500)? If not, how difficult would it be to add that count since we have the pagination?
These features would help a lot in the development of the datasource and reduce the number of requests to the server.
Thank you very much in advance.
@ROAM447
Where exactly is this code? From what I understand the query part is done here in this line: https://github.com/commonsense/conceptnet5/blob/fda1b39607ea6f07e41ddce22d9d5f8360ded581/conceptnet5/api.py#L244
A search is performed considering limit + 1 to capture the existence of more values. It doesn't have a query that captures all the edges (in fact it would be quite unoptimized). I'm looking for something that I do this count directly in the DB (probably here).