Parse txt bibliography automatically

kak · January 28, 2007

Often, EndNote item or BibTex item is not available.
If zotero can parse a text such as

H. Balakrishnan, V. N. Padmanabhan, S. Seshan, and R. H. Katz. A comparison of mechanisms for improving TCP performance over wireless links. IEEE/ACM Transactions on Networking, December 1997.

and extracts Authors and Title, et al automatically, it could be great.

I know there are several formats, but they are similar and it is not necessary to be perfect. It will still help us.

I am currently just putting the text into "Notes", but the text is not appeared on the list view, so I anyway have to type at least Title by my hand.

noksagt · January 29, 2007

Please see:
http://forums.zotero.org/discussion/179/import-from-wordprocessed-apamla

kak · January 29, 2007

I still want this feature.
c2bib could be a good software, but I want to finish in zotero.
It need not to be perfect.
Even if zotero supports only one popular format for paper references, it still should be greately useful.

bhusky · March 31, 2007

I am new to Zotero (for which, compliments and many thanks), but I agree that a text parser is also essential. Often references appear in plain text, not generated by one of the bibliographical engines or even database structures like Amazon. (The obvious case is an online PDF with a standard bibliography.) I imagine a function that would allow highlighting of a single text reference, which Zotero would grab and parse, something like cb2Bib, referred to in another thread. I would emphasize that the parser needn't be perfect or as elaborate as cb2Bib -- it could even just break up the chunks of the entry and then with pull-down menus let the user identify type of reference, author, etc.
Seems like this wouldn't be out of proportion to the really extraordinary work the programming gang has already done.
Thanks, J Leiderman, MD

Matthias · March 31, 2007

I don't think this would work. Parsing of any free-text references is not a trivial task, and even if a lot of smartness is applied by the developer, the parsing would probably fail *miserably* in at least half of the cases. And I'm pretty sure users would complain about this endlessly (remember, there are hundreds of different formats, not to speak of individual formatting errors), and get frustrated quickly with Zotero's poor free-text parsing capabilities.

Personally, I am not convinced that pulling drop down menus for each of the parsed parts (usually 7 or more parts) would be faster than actually manually selecting the parts with your mouse and copy & pasting. And doing this manually will actually greatly increase accuracy. Currently, a machine cannot beat a human w.r.t. to free-text parsing.

Also, there's cb2Bib, why re-invent the wheel? IMHO, it's much better if the Zotero developers devote their precious time to making the core application better, and to further improve their standards support.

Speaking of the case of an "online PDF with a standard bibliography", it may be worth trying to contact the author of the bibliography and to kindly ask for an EndNote file of the bibliography (or the like). It might spare you from lots of typing (or pulling of drop down menus).

noksagt · March 31, 2007

Also, regarding the specific case of a PDF bibliography: Zotero has no official PDF parsing. The devs have said they'll use pdftotext to do fulltext indexing, but this process will lose formatting information needed to give contextual information about how to extract bibliographic information.