Cross linked IDs between senses

In the previous version of the dictionary website, there were HTML links on some synonyms (or some words in definition / examples) which were pointing to appropriate word sense.
The new version removed the links, but some are still highlighted in bold.

e.g. extraterrestrial synonym should point to alien sense 3, not base 1

I guess there was a lot of effort that went into linking the words with their appropriate synonym sense/sub-sense; if the link id's are still available in the database (for synonyms/definition / examples) will it be possible to make it available for the new version of the APIs? This is very useful for NLP.

Regards
Simone

Comments

  • SimoneSimone Administrator admin

    Hi @simonesmith
    Thanks for the suggestion - as usual, I can't comment on it myself, but I am going to alert my colleagues who can! :)

  • preimannpreimann Member, Administrator, Moderator admin

    Hi @simonesmith

    You are correct, in the old version of the website it was possible to click on synonyms and be taken to the correct sense of that synonym for the word. Having rebuilt our content into an RDF linked database (which powers the new website), these links do still exist but we are yet to surface them either for the website or through the API. I'll log this as a feature request as I can see the value in adding it. Thanks for raising it!

    Phil

  • simonesmithsimonesmith Member

    That would be great, perhaps can link under: results.lexicalEntries.entries.(senses/senses.subsenses).crossReference and crossReferenceMarkers (Id should be the linked sense/sub-sense ID not high level lemma, currently it seems that CrossReferencesList model 6.text = model6.id )

    I'm wondering how the words links can be added for identified words IDs in definitions/examples (because are phrases which might contains more than one crossReferenceID)

  • preimannpreimann Member, Administrator, Moderator admin

    Hi @simonesmith. This is really useful detail, thank you. I have passed this detail on to the technical team so that it can be considered as part of our general exploration into surfacing our link data via the API.

  • simonesmithsimonesmith Member

    Thank you very much Phil, that is very helpful.

    Any idea when the new version will be out?

    Have a good one
    Simone

  • preimannpreimann Member, Administrator, Moderator admin

    @simonesmith We schedule releases as frequently as possible in order to add more content and address bugs, but our schedule is quite flexible. Our next release is June 6, which will launch two new datasets, but beyond that I don't have definite dates to pass on at the moment.

  • simonesmithsimonesmith Member

    Hi @preimann

    That's nice; what will be the new available databases?

    Is there a link with the a release version notes? Or perhaps we need a new category in the forum for the release notes.

    Cheers
    Simone

  • SimoneSimone Administrator admin

    Hi @simonesmith,
    How are you?
    Before @preimann can complement his reply, I just wanted to direct you to this area on the API portal where you can keep up-to-date on all the new releases:
    Supported languages
    (we also have a link to it from here forum here: Which languages are currently available via the Oxford Dictionaries API?)
    I hope this is helpful!

  • preimannpreimann Member, Administrator, Moderator admin

    @simonesmith We are aiming for the release to include our comprehensive German and Portuguese bilinguals, but as they are still in the final stages of processing I don't want to be too committal until closer to the date.

    You can keep up to date with all our releases at our updates page on the portal.

    Thanks,

    Phil

  • simonesmithsimonesmith Member

    Hi @Simone

    I'm fine thank you. Just trying to understand if I can link the main English dictionaries senses with the other dictionaries senses using IDs.

    Thanks for the links, I was not aware of the update page.

    Will the url for API call be different or will be the same link?

    Cheers
    Simone

  • joughtredjoughtred Member, Administrator, Moderator admin

    Hi @simonesmith,

    You can use the same link for this.

    Thanks,

    Joanne

  • simonesmithsimonesmith Member

    Hi @preimann

    Phil, I would like to check something with you.

    I noticed on thesaurus.com website that the synonyms are highlighter with different colors.
    If you guys are adding synonyms ID to sub-senses, will it be possible to differentiate between "same meaning" vs "similar meaning"? Perhaps it could be a different JSON tag. I remember some synonyms on the oxford web page were in bold font; not sure if that represented a "same meaning synonym"

    Regards
    Simone

  • preimannpreimann Member, Administrator, Moderator admin

    Hi @simonesmith. I'll pass the feedback to our editorial team, but we currently distinguish between 'core' and 'non-core' synonyms rather than same or similar meaning. Core synonyms are what you typically see in bold on our site, and are listed as first in the JSON response within each list of synonyms per subsense. This offers a very granular, accurate, depiction of the synonym relationship to the specific sense of the word, rather than providing potentially misleading information that the meaning is exactly the same.

  • simonesmithsimonesmith Member
    edited June 6

    Hi @preimann

    Thanks Phil, It would be great if we could differentiate between core and non-core synonyms IDs.

    Regards
    Simone

  • @preimann

    Hi Phil

    How are you lately. Any updates regarding the synonyms IDs?
    Will be great if can manage to add them.

    Regards
    Simone

  • AmosDuveenAmosDuveen Member, Administrator, Moderator admin

    Hi @simonesmith,

    Unfortunately it is a matter of prioritisation and this is fairly low down the list. We do however have some goodies in the pipeline, which will go live over the coming months, including more languages and LexiStats, access to our corpus-derived statistical data.

  • simonesmithsimonesmith Member
    edited September 25

    Hi @AmosDuveen

    Thanks for letting me know. How far down the list it is, any ETA for that one? I need to inform my boss if my R&D proposal is doable or not. If is more that a year down the road then I'll loose all the budget and support.

    Sincerely
    Simone

  • AmosDuveenAmosDuveen Member, Administrator, Moderator admin

    Hi @simonesmith,

    I'll be honest, I wouldn't want to entirely rule out the possibility that this won't get done within the next 12 months, but I would probably highlight it as a significant risk in any analysis of the issue.

    There are people looking at known losses (which probably includes this issue) in the data conversion process which prepares content for the API, but they are more focused on the dictionary data, though they will eventually get to the thesaurus content, I'm sure. Even when they do eventually find a satisfactory resolution, there would need to be a new export file produced that then needs converting and loading to the API, which is not an insignificant task.

    Can I perhaps offer you a different solution? If this work is part of a research project, then maybe we could sign a research agreement (i.e. a non-commercial license) that would allow us to share the bulk XML data, which will make the 'core'/'non-core' distinctions you are looking for.

    If that solution interests you, please email api@oxforddictionaries.com, where my colleague, Jo, will be better able to direct you. If you could please make sure to include the usual details (name, institution, purpose and time frame), then we can put them into a draft agreement and email it to you for review and signature.

  • Hi @AmosDuveen

    It sounds quite complicated reverting the thesaurus links, keeping my finger crossed.

    In the meanwhile I'll talk with my boss about the XML approach. Is there a license upgrade from research to commercial in case there are commercial spin-offs from a R&D project?

    Regards
    Simone

  • AmosDuveenAmosDuveen Member, Administrator, Moderator admin

    Hi @simonesmith,

    That is a long way beyond my remit, or even knowledge but, as a personal opinion, I very much doubt it would be a problem. I imagine you would have to come to an agreement over commercial terms in a new licensing deal rather than 'upgrade' the existing ones, but that is a fairly minor quibble.

    If it is of interest to you, when you make contact, I could arrange for Jo to bring in one of our Global Business Development (GBD) reps who will know more and be better able to advise on the commercial aspects of licensing, however, any detailed discussion would be better left until you know you have some commercial use for the data.

Sign In or Register to comment.