3

Is is possible to get references from a paper? For example, given that paper A cites paper B, C and D, i wanna to extract Bibtex of B, C and D from A. Is is possible to do this conveniently? Thank you for any help!

EDIT: A more concrete example: I have paper A which cites many other papers such as B, C and D. I wanna some ways to directly get bibtex of all these cited papers in paper A.

ytutow
  • 133
  • 1
    Welcome to TeX.SE. Please clarify what exactly you mean by "extract Bibtex of B, C and D from A". – Mico May 26 '17 at 11:45
  • 1
    Some database systems record the references in a paper, sometimes they can be extracted from the database. An example where this definite works is http://adswww.harvard.edu/. http://www.ams.org/mathscinet/ provides references but no means to download them as .bib directly, so does http://apps.webofknowledge.com/. – moewe May 26 '17 at 11:45
  • What I mean is that I have paper A which cites many other papers such as B, C and D. I wanna some ways to directly get bibtex of all these cited papers in paper A. – ytutow May 26 '17 at 11:48
  • Presumably some of the known citation database programmes such as Mendeley, Zotero and EndNote also have a similar option. – moewe May 26 '17 at 11:48
  • @ytutow Can you tell us which paper A you have in mind? – samcarter_is_at_topanswers.xyz May 26 '17 at 11:53
  • @samcarter It is 'gradient-free hamiltonian monte carlo with efficient kernel exponential families' – ytutow May 26 '17 at 11:54
  • @moewe I have tried http://adswww.harvard.edu/ and found that this database does not contain complete ref – ytutow May 26 '17 at 11:56
  • @ytutow You can get the source code from https://arxiv.org/format/1506.02564. Unfortunately this does not contain the .bib file, but you can extract the information from the bbl file and convert it to .bib via http://text2bib.economics.utoronto.ca/ – samcarter_is_at_topanswers.xyz May 26 '17 at 11:57
  • 2
    The information is not in the PDF or, even, the .tex, if that's what you're asking. So what a database provides will be the best you can get unless something like Zotero can parse the PDF directly, or you can get the .bib or .bbl as above. – cfr May 26 '17 at 11:58
  • @samcarter Thank you! You are really clever! – ytutow May 26 '17 at 12:05
  • A big thank to everyone here for helping me! – ytutow May 26 '17 at 12:06
  • I've marked this question as a duplicate. The earlier query actually states that the potential input is a bbl file, rather than the formatted pdf file. However, the main issues -- all linked to the loss of meta information -- are exactly the same, regardless of whether the source is a bbl file or a pdf file. – Mico May 26 '17 at 12:46

1 Answers1

3

Step 1:

The source code of the article is available from arxiv.org/format/1506.02564. Unfortunately this does not contain the .bib file, but "only" the .bbl file.

Step 2:

To extract the desired information from the .bbl file, one can try to convert it back to a .bib file. This can for example be done with text2bib.economics.utoronto.ca

The result will be something like 

@article{neal2011mcmc,
author = {R. M. Neal},
journal = {Handbook of Markov Chain Monte Carlo},
title = {{MCMC using Hamiltonian dynamics}},
volume = {2},
year = {2011},
}

Step 3:

Check the result and correct the output of this automatic conversion. For example the entry shown above should probably changed to @inbook.

samcarter_is_at_topanswers.xyz
  • 158,329
  • 1
    This example of a "quasi-automated approach" is, I'm afraid, both well-intentioned and an excellent illustration of some of the things that can go badly wrong with this approach. First, @article is definitely not correct as the entry type; it should be either @inbook or @incollection. Second, the journal field name should be changed to booktitle and the volume field name should be changed to edition. Third, the fields editor, address, publisher, and pages should all be supplied as well. I strongly suspect one would be better off building the entire entry from scratch. – Mico May 26 '17 at 12:34