Converter EndNote XML to MODS or RDF

ben58

Getting data from EndNote into Zotero is a long standing problem, as RIS export from EndNote is broken, Zotero team can do nothing to improve it, and Thomson couldn't be less interested in fixing the issue. EndNote XML export is apparently better: at least after I imported a test data set back to EndNote (X1), they looked OK. As was discussed (by noksagt ?) their XML is also flawed - but for getting data out of EndNote, it's probably second best to reading the (closed) native format.

There exist a ticket https://www.zotero.org/trac/ticket/472 - priority "major", opened three years ago!

Wouldn't be a stand-alone or a server based solution probably easier to implement? Convert the "dirty" EndNote XML into a clean MODS or RDF XML, which could be read by Zotero without problems. Write an XML converter not necessary in JavaScript, but in the language and on the platform of your choice, using any suitable frameworks.

For the user this solution may involve a little additional effort in installing a separate software, even maybe on a different machine, but it's nothing as compared to manually checking hundreds or even thousands data sets (see e.g. http://forums.zotero.org/discussion/8607/advocacy-for-switching-endnote-vs-zotero/#Item_11 )

bdarcus

I think XSLT would be a sensible candidate for such conversion, but would note that I'm pretty sure the bibutils code can do this already (at least to MODS).

ben58

Thanks!

http://www.scripps.edu/~cdputnam/software/bibutils/#endx2xml

"endx2xml converts a EndNote-XML exported reference file to a MODS XML-intermediate bibliography file. "

Will give it a try

ben58

Well, far from perfect yet :(

In several cases the information contained in the MODS got lost - I can't judge whether it was poor bibutils MODS output or Zotero import failure (and going to start a separate thread). In any case there is hope for the issue being sorted out.

But then another thing. I didn't find any information whether MODS specification supports information on attachment file paths. Does it?

Would Zotero attach files if it finds something like
<url>file:///the-path-to-PDFs/PDFPruefPDF-4163337984/PruefPDF.pdf</url> ?

If yes, replacing the string
<url>internal-pdf://PruefPDF-4163337984/PruefPDF.pdf</url>
by one like one above would be more-or-less trivial.

noksagt

I can't judge whether it was poor bibutils MODS output or Zotero import failure

The problem may also be with the Endnote XML: there are a ton of variants & many of them aren't even valid XML, so it is hard to read.

In my experience:

• Bibutils output is very clean.
• Zotero's importer usually imports things correctly but
• Zotero's importer does not import everything

But then another thing. I didn't find any information whether MODS specification supports information on attachment file paths. Does it?

Nothing formal. However, based on the examples given by the LoC, an ad hoc method is in zotero and refbase and probably other programs. If the MODS contains something like:<location><url access="raw object">URL_GOES_HERE</url></location>zotero will import it.

ben58

nosagt, if you are able to judge the quality of an XML file, it would be nice if you could look at the files I have uploaded, see the separate thread
http://forums.zotero.org/discussion/8726/data-loss-on-import-of-bibutils-generated-mods-file/#Item_3

As the EndNote XML file could be put through two conversions to produce a nicely readable RIS - in this case it is probably not the EndNote fault.

awiklendt

http://endnote.com/downloads/style/bibtex-export

Download above, and add to EndNote.

Export using EndNote to txt with Bibtex filter.
Rename txt as bst.
Import into Zotero. Have 100s of references from Endnote in Zotero but have not had time to check accuracy/reliability/etc.

very quick.

aurimas

Zotero's support for RIS when exported from EndNote has very much improved and is now probably the most reliable and complete way of transferring metadata from EndNote to Zotero.

See http://www.zotero.org/support/kb/importing_records_from_endnote

awiklendt

oh my goodness, thank you aurimas! the bibtex proved to be less than reliable (transfered all references to a limited 5 or so types, so webpages only converted to "book" and lost their URL...!) i'll have a look at the RIS and hope it's better!