Allow importing of records without a DOI

[michael]

Currently we only allow importing of bibliographic records that have a DOI.

See: https://github.com/substance/texture/pull/480#discussion_r181580524

To add support for entries without DOI's we'd need to make sure we don't end up with a lot of duplicates. Some options:

• Be okay with duplicates being created for entries without a DOI
• Implement a checksum over JSON object so we can check if an entry has exactly the same content and if yes skip it.

[Melissa37]

Hi there

Why is this so important? Is it because you are still trying to build a database of references? :-)

Many journals do not have DOIs yet, and some journals with DOIs did not update their archives and add DOIs to them. Some books have DOIs, but not all.

We have to allow for all reference types to not have DOIs.

eLife has schematron that detects duplicate references I think @JGilbert-eLife can you explain how we do it?

Cheers M

[obuchtala]

Yeah 😀 no, it isn't because of that. As you said there will be a lot of references without DOI. Also we are allowing to Import from reference Managers, such as Zotero. We are just concerned to introduce problems by allowing to import the same record twice.

[Melissa37]

Cool! I hope James might be able to share how we do the check via Schematron if that helps?

[JGilbert-eLife]

Hi everyone. I'm afraid we don't have a rule that says 'references must occur only once' per se. We have rules that look at references with the same author information for the purposes of correctly differentiating between Smith et al., 2015a and Smith et al., 2015b, for example. These basically take the content of the author list and the year and check them against every other reference for matches up to a point (e.g., are there two references where the first author and the year will be displayed as Smith et al., 2015, in which case one will need to become 2015a and the other 2015b). These rules can be found at the end of https://github.com/elifesciences/reference-schematron/blob/master/element-citation-general.sch.

What we don't have is a rule that compares the whole content of one reference against every other reference to check for exact matches. My instinct is that this should be possible with schematron but I'm not sure exactly how to do it, since it would involve doing this across multiple different kinds of references. It's got to be a top-level check and that wouldn't catch mistakes like a reference entered once as a book and a second time as a journal article.

A more plausible step might be to compare just the titles, say, in conjunction with the author names, but that brings us up against the case where two separate references have the same author and title details for legitimate reasons (a preprint and the final article being cited separately, for example). Of course, that needn't be a problem if it's just a warning rather than a error.

Are you planning on including schematron in texture or did you just want to know the logic we were following so you could replicate it in something else?

[robsalasco]

any news about this feature? I'm using texture for a spanish journal but it's near impossible to have all the doi codes

[obuchtala]

Are you planning on including schematron in texture or did you just want to know the logic we were following so you could replicate it in something else?

When integrating Texture into XPub we would like to create an interface between Texture and Schematron. In addition to that, we are adding a simple validator into Texture that runs checks inside the editor on-the-fly.

[eternoinvisible]

Yeah no, it isn't because of that. As you said there will be a lot of references without DOI. Also we are allowing to Import from reference Managers, such as Zotero. We are just concerned to introduce problems by allowing to import the same record twice.

I tried to import the csl jason (generated with zotero) file of citations without DOI and it didn´t allow me to do it because of lack of DOI error