Search PubMed

It would be most helpful to be able to search PubMed (or other similar libraries, I'm most concerned by PubMed) by keywords or authors etc directly from Zotero and be able to include references selected from the results of that search in the library. That's an important feature when one is writing a paper or is doing a bibliographic search and it would contribute greatly to generalize the use of Zotero. Sorry if that request have already been formulated and I've missed it...
Jean-Paul Herman
  • I disagree. Zotero works differently than other reference managers; it imports content directly from web pages. As a result, if there isn't one already, the focus should be a translator for pubmed gateways.

    Just as a practical matter I (who couldn't care less about pubmed) would find it annoying to have privileged support for it, in the same way that you would be find it annoying if Zotero privileged my sources.
  • Zotero can already scrape PubMed. (And PubMed already has search plugins for Firefox, RSS search feeds, etc.)

    I agree that the benefit of having an in-Zotero search would be fairly small, but I personally wouldn't be opposed to having a universal search across multiple data sources (similar to the firefox searchbox or LibX, either of which might provide Jean-Paul with something he wants). I also agree that, if development costs are used to make it, it should be generic (to allow many search providers, using OpenSearch or similar).
  • Scraping pubmed is much, much more time-consuming than typing keywords into Zotero. Right now I have a list of about 200 references to enter into Zotero. You want me to do a separate web search for each one, click on the link, then ask Zotero to scrape it? Forget it. Endnote makes it easy by letting you search PubMed directly. I'm not going to switch to Zotero if it can't do that.
  • What format are these references in? Do you have the PubMed IDs? Zotero can import PubMed's XML format, so if you use any existing tool to get a file with all the references, you can then import that file into Zotero quite quickly. I believe PubMed has a way to look up large numbers of PMIDs at one time, which is precisely what you could use in this case.
  • edited November 13, 2011
    They are typed out, in Word documents, according to different style standards, eg Chicago, APA. No, I would have to search pubmed to get the pubmed IDs.

    What I'd really like is to give Zotero a word document or text file full of such references, and have it parse them and import them, and then insert a bibliography into a document that had them alphabetized. That would be awesome. I'm combining 4 of my progress reports into a thesis, and the references are in a variety of formats.
  • edited November 13, 2011
    That's not really possible, not with any reference manager. See here for more:
    http://www.zotero.org/support/kb/importing_formatted_bibliographies
    (though I think WizFolio actually queries pubmed, so it might work especially well for this).

    I don't know the Endnote search interface for pubmed, but I don't really see how it would/could be much faster than going through the web interface - could you describe how that actually works?
  • With Endnote, you use the tool to search your library; but you can tell it to search PubMed instead of your library of references. So then, e.g., you type in an author name and a year, and it retrieves all the references from pubmed that match, and you choose the correct one and add it to your library, and go on to the next.
  • I just don't see how that's different from searching pubmed online, clicking the Zotero folder in the search results, selecting the one(s) that you want, and going on to the next search?
    I mean, if you're really wedded to how this is done in Endnote, by all means, stay with Endnote, but have you actually tried doing this in Zotero?
  • The Zotero way has the same effect, but is slower. It's doable, though.
  • EndNote includes a Z39.50 implementation that allows it to search various providers for items, using basic fields like author and title. It would be a significant undertaking to build Z39.50 support into Zotero, but it would make it possible to retrieve items using a non-scraper interface. Since this would be only loosely tied into Zotero, it should probably be done as a plugin. If written sufficiently flexibly, it could support other query protocols like OAI-PMH. These non-web interfaces to databases can indeed be significantly faster.
Sign In or Register to comment.