Export of special chars to bibtex problem

async

I have exported my collection to bibtex and now have a problem when compiling the latex document.

One author is named "Wätjen". The ä is exported very strangely. In LaTeX it should be exported as \"a. Some goes for ö and ü. It even looks strange if I open the file in a simple text editor.

Furthermore URLs should be surrounded with the \url{} tag making links clickable in pdf documents.

Are there somewhere some settings for the bibtex module to correct the behaviour?

noksagt

Set your export character encoding to anything other than utf-8.

Ringelkrat

This issue is still not solved completely! There is a problem when exporting a Website from Zotero to Bibtex. A url = {...} entry will be generated in the bibtex containing the url. But if the url has special characters like _ or & then bibtex/latex will prompt an error "Missing & inserted"!
For example code #1:

@misc{bar,
title = {bar},
url = {http://foo_feature_transform&printable=yes},
howpublished = {{http://foo\_feature\_transform\&printable=yes}}
},

Here there is a problem with the url because of the _ and & !
The problem could be avoided, if the special characters would be translated like it is done in the "howpublished" entry (see code #1)! Using \_ and \& .

Or one could use the \url or \href tag like in code #2:

@misc{foo,
title = {foo},
url = {\url{http://foo_feature_transform&printable=yes}},
howpublished = {{http://foo\_feature\_transform\&printable=yes}}
},

But the latter method has been disapproved by the community because it is no standard. See http://forums.zotero.org/discussion/4331/bibtex-importexport-problems/

So the solution would be just to parse the url as it is done for "howpublished"? Am I right? And if so, can somebody correct the code in Zotero, please? This error prevents the exported library to compile.

I use Zotero 2.1.6 and Firefox 3.6.17.

noksagt

So the solution would be just to parse the url as it is done for "howpublished"? Am I right?

Sadly, no: character escaping was intentionally omitted from the url field because some bst files use \url or \href & the links are broken when the characters are escaped.

Ringelkrat

Thank you noksagt for the fast response! I see, that it makes sense to omit character escaping in some cases. To resolve this conflict it may be a good idea to provide a setting so that the user can decide whether the special characters should be escaped or not.

Unfortunately I am bounded to a specific bst file by my publisher. Is there any other way to deal with this problem, apart from escaping the characters in my bib file manually?

noksagt

We've historically avoided offering too many choices on export, particularly ones that are for a specific format. You can customize the Zotero BibTeX translator yourself. If I were in your position, I'd probably just remove all url lines & remap howpublished to url if it was needed.

Ringelkrat

Is there somewhere a tutorial or so how to cutomize the Zotero BibTeX translator?

noksagt

http://www.zotero.org/support/dev/translators

Ringelkrat

Thank you noksagt for the hyperlink!

I have copied and modified the original BibTeX translator by Simon Kornblith and Richard Karnesky in order to escape special characters in urls. I would like to share my experience in case somebody else encounters this issue:

The easy steps are:

1. create a new Translator: http://www.zotero.org/support/dev/translator_development_outside_scaffold#create_or_modify_a_translator_file and save it at the appropriate location (see http://www.zotero.org/support/dev/translator_development_outside_scaffold#translator_files)


2. Edit the metadata: Change the label to the new filename of the translator without .js: "label":"BibTeX (modified URLs)", in line 4 and set a new translator-ID: "translatorID":"aea153a0-812b-11e0-a9da-0002a5d5c51b", (you can use this service to generate one: http://www.itu.int/ITU-T/asn1/uuid.html)


3. Change the code at line 1830 and following: // url field is preserved, for use with \href and \url
// Other fields (DOI?) may need similar treatment
if(!((field == "url") || (field == "doi") | (field == "file"))) { to // url field is preserved, for use with \href and \url
// Other fields (DOI?) may need similar treatment
if(!((field == "doi") | (field == "file"))) { //edited! Original was: if(!((field == "url") || (field == "doi") | (field == "file"))) {This way "url" is not exempted from the escaping any more.


4. restart Firefox

It works so far. I hope this helps anybody. Feel free to criticize or correct me.

ajlyon

This looks like exactly the right solution. Hopefully this will help future people in your situation.