Optimal Import from CiteULike?

[iandol]

I use citeulike when I'm in my browser to quickly save and tag references (for which it does the job perfectly). Of course I would love some auto sync between citeulike and BE (this used to work wonderfully for Mendelay before they rewrote their API), but I understand this would take some development time for BE from which only a few users may gain.

However there are a few problems with using regular imports (assuming I export my whole citeulike library):

• Bibtex export is more flexible from citeulike, but citeulike bibtex is incompatible with BE.
• RIS export is more compatible, but BE doesn't exclude duplicates during import automatically, wasting time as I have to manually remove dups afterward each import.
• RIS imported references have no PMID or DOI, but the URL field is a DOI. I'd like BE to automatically fill the DOI field if it finds a http://dx.doi.org/... style URL.

So really out of this I'd minimise my friction with duplicate detection on import and DOI detection of URL fields.

[Jon]

Why is their BibTeX incompatible? You can edit the Bookends BibTeX filter to deal with it. Remember that for BibTeX you have to account for all fields, even if you just tell Bookends to ignore them.

Jon
Sonny Software

[iandol]

They seem to use double {{}} for titles which I cannot overcome in the import filter manager AFAIK (is this valid bibtex?). Also for page numbers they use a double -- bookends somehow to combines the subsequent field into page numbers (so I get 2114--2122},pmid = {25057148 for example — perhaps that is the account for all fields issue you mention?

I'll try to account for all fields and see if thia improves things, though being able to control {{}} would be helpful.

And I still am left with problems of having to do more post-processing than I'd like for a regular action — sometimes there is a doi link but no doi field and duplicates need to be dealt with.

Example Bibtex from CiteULike:

Code: Select all

@article{Oleskiw2014Spectral,
    author = {Oleskiw, Timothy D. and Pasupathy, Anitha and Bair, Wyeth},
    citeulike-article-id = {13296354},
    citeulike-linkout-0 = {http://dx.doi.org/10.1152/jn.00250.2014},
    citeulike-linkout-1 = {http://jn.physiology.org/content/early/2014/07/18/jn.00250.2014.abstract},
    citeulike-linkout-2 = {http://jn.physiology.org/content/early/2014/07/18/jn.00250.2014.full.pdf},
    citeulike-linkout-3 = {http://view.ncbi.nlm.nih.gov/pubmed/25057148},
    citeulike-linkout-4 = {http://www.hubmed.org/display.cgi?uids=25057148},
    day = {01},
    doi = {10.1152/jn.00250.2014},
    issn = {1522-1598},
    journal = {Journal of Neurophysiology},
    keywords = {v4},
    month = nov,
    number = {9},
    pages = {2114--2122},
    pmid = {25057148},
    posted-at = {2014-11-03 00:30:05},
    priority = {2},
    publisher = {American Physiological Society},
    title = {{Spectral receptive fields do not explain tuning for boundary curvature in V4}},
    url = {http://dx.doi.org/10.1152/jn.00250.2014},
    volume = {112},
    year = {2014}
}

[Jon]

The double dash is the way BibTeX indicates and en dash. If you enable BibTeX in Bookends preferences and tell it to convert BibTeX escaped chars to unicode equivalents it will import that as –. And yes, you should tell the Bookends import filter what to do with pmid.

Jon
Sonny Software