Nature translator broken?

iosonofabio
I get an error when trying to scrap any of these:

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature12053.html
http://www.nature.com/nature/journal/v473/n7346/full/nature09944.html

but it works with e.g. this:

http://www.nature.com/nature/journal/v477/n7363/full/nature10341.html

No idea what the issue is. Any clues?
  • iosonofabio
    edited April 5, 2013
    Actually, some deeper research shows problems in the Nature translator with non-human authors, such as

    NISC Comparative Sequencing Program

    It assumes "Program" is a last name, and this creates all sorts of problems. In particular, in the JS function

    scrape(doc, url)

    you get a shift of 1 in position between item and items[1]. This means that the lengths of items[1] and item differ and when the last author is reached things turn bad.
  • adamsmith
    thanks - I see that, yes. We'll take a look (@aurimas - can you take this?)
  • DWL-SDCA
    I have recently noticed this when using the magic wand and ISBNs when the item includes institional authors. Also, sometimes the book's author field in the right column is populated as

    "Institution Name, (first author)"

    The simple editing tasks required are only a minor inconvenience to me but if this could be fixed It would be nice.

    As an aside, for the past year or so, I've noticed that imports from PubMed can have a similar problem with items having institutional authors -- especially when the instutional authors are at the end of a list of person's names. This appears to be unrelated to Zotero because the PubMed xml omits the institution names. However, sometimes the tagged format includes the institution.
  • adamsmith
    DWL - are you saying ISBN import breaks with institutional authors? (that's what's happening for Nature). Importing institutional authors wrong is unfortunately often unavoidable - though it shouldn't happen for ISBNs where we get data from the LoC.
  • iosonofabio
    In the Nature case:

    - scrapeEM gets the authors right and creates items[0] --> item
    - scrapeRIS gets the authors wrong and creates items[1]

    items[1] is one too short and this makes a mess. I.e., the problem is in scrapeRIS.
  • DWL-SDCA
    @adamsmith "are you saying that ISBN breaks..." Yes and no.I've noticed that this occurs mostly with books published in the mid-1980s and before. It also seems related to the complexity of the institution name ("Transportation Research Board. National academy of Sciences." or "Institute of Medicine. National Academy of Sciences." or "Division of Unintentional Injuries. National Center for Injury Prevention and Control. Centers for Disease Control and Prevention."

    I will find specific examples and reply with them soon. I'll also look at and share how the LoC presents the metadata for the problem items.
  • iosonofabio
    @DWL-SDCA: please make a new bug report for that issue and stop posting here. It is inconsistent with the title of this one and very likely to involve a totally different part of the code. Thanks.
  • adamsmith
    (yeah, fabio is right - the Nature translator issue is very specific because of the way it tries to merge data from two different sources, best to open new threads for any other institutional author issues).
  • aurimas
    @adamsmith I'll fix this asap. would like to merge this with the other nature changes if I can, do it may take just a little bit longer. hopefully today though
  • aurimas
    This should be fixed as soon as https://github.com/zotero/translators/pull/549 is accepted
  • iosonofabio
    Thanks!
  • iosonofabio
    Ups... the first of the websites above now works, but the second still does not, for a related reason:

    http://www.nature.com/nature/journal/v473/n7346/full/nature09944.html

    The institutional author, in this case, is not in the RIS file at all. The consequence is the same as before: the lengths of items[0] and items[1] differ, and the translator fails.
  • aurimas
    The patch has not been accepted yet. I'll let you know once it goes through.
  • aurimas
    The patch was merged. You can update your translators via Preferences... -> General -> Update Now or wait up to 24 hours for them to update automatically.

    Also, this update introduced ability to automatically download supplementary data (or just attach links to it). It will also automatically attach links to GEO, PDB, and other relevant entries. We hope you find these changes useful.

    See http://www.zotero.org/support/preferences/hidden_preferences#translator_preferences

    Supplementary data attachment support is on its way for other websites as well
Sign In or Register to comment.