Electronic catalog of National Library of Belarus

r.douhi

Hello! Is it possible to make a translator for unicat.nlb.by?
For example this record http://unicat.nlb.by/opac/pls/pages.view_doc?qid=57314&nn=2&id=286787&siz=10&off=0&format=full (catalog has session time and this link may be expired)

Also there is problem with another electronic catalog on website of National Library of Belarus (record http://www.nlb.by/portal/page/portal/index/resources/expandedsearch?classId=B33E739B22884D82AACAC24EBFB1DA89&submitR=empty&_piref73_180746_73_34794_34794.biId=386521&_piref73_180746_73_34794_34794.strutsAction=biblinfoaction.do&lang=en)

I have read documentation, but there is no metadata on these electronic catalogues and i can't understand how to connect it with Zotero.

Thank you in advance. I will be pleased with any answer.

zuphilip

I have found BELMARC in both catalogues. Do you know whether this is similar the other dialects like to MARC21 or UNIMARC? This should be possible to try out...

r.douhi

Yes, BELMARC is based on UNIMARC. Do you mean to try to get record from the MARC fields? Sadly, exporting from the BELMARC files (*.mrc) doesn't work properly in Zotero.

zuphilip

Yes, I would try to get BELMARC from the HTML representation (which is not so nice for that). UNIMARC should be handled reasonable well in Zotero. I haven't seen that one can export to BELMARC from your catalogue, is this possible? Can you elaborate what exactly didn't worked properly in Zotero?

r.douhi

Catalogue from www.nlb.by can't export to mrc file, but union electronic catalog http://unicat.nlb.by can. Here some screenshots
http://i.szoter.com/054e10ae84cb5649.jpg - exporting BELMARC record to file
http://i.szoter.com/85a71df86e82d0ef.jpg - problem with importing it in Zotero. All fields are empty except Author (it's replaced with keywords) and Title.

Sorry it all in russian, thank you for your help.

zuphilip

Okay, thank you for the information. I will try to scrape the MARC from the national library catalogue first and then we have to see if we can tweak the MARC import a little to deal better with BELMARC.

I tried to come up with a mapping:

101 -> language
200 -> title
210 -> publicationPlace, publisher, year
215 -> numPages
225 -> seriesTitle + seriesNumber
606 -> keywords
675 -> DDC
700 -> author
899 -> callNumber

Can you check if this is reasonable or if there is something important missing?

r.douhi

These main fields are correct, nothing is missing but 675 is UDC in BELMARC if it's important.

zuphilip

Here is a first version of a translator for the catalogue of the national library of belarus: [...edited...]

Can you test and report your results here?

r.douhi

Thank you for the quick response and your help. It works great! There are only a few questions:

How to change "et al." in case of many authors to russian "и др."?
Initials of the authors repeated twice and also i think there is no need in russian letter "c" in the Number of pages field. Can it be fixed?

All this i marked in the screenshot
http://funkyimg.com/i/LXkw.png

Once more question, what about unicat.nlb.by? This search system used by many libraries in Belarus. Can it work with Zotero?

Thanks again.

zuphilip

1) The et-al in the middle pane of zotero is independent of everything else here in this discussion. Maybe that is hardcoded and cannot localized. @Dan : Any thoughts on that?

2) The doubling of the initials are a result of the MARC data, e.g.

702 #1 $3 BY-SEK-125257 $a Демиденко $b О. М. $g Олег Михайлович $c доктор технических наук $f род. 1960 $b О. М. $c доктор технических наук $f род. 1960 $g Олег Михайлович $4 340

There is twice the subfield b, which will lead to the doubling of initials in zotero. Does this make any sense in BELMARC?

3) We can try to just extract the first number. However, normally in Marc field 215 there can be more fuzzy information (volumes, sheets vs. pages, roman numbers). Thus, it is for me not clear if this is always an improvement. What do you think?

4) The Belarian union catalogue is also possible to do. Maybe I find some time next week(s). BTW the MARC import in zotero should give the same amount of data as for the national library, because in the background both are using the same mechanism.

r.douhi

Points 1 and 3 in your last message are not critical for work, it is necessary to check and correct all records for me manually anyway. Much more important ability of Zotero is to grab records automatically.

The doubling of the initials is a mistake of our catalog developers, i found out. There is no such problem in the union electronic catalog.

I would be very grateful if you did translator for unicat.nlb.by some time in future.

Thanks for the solution to our problems, zuphilip.

adamsmith

@zuphilip--you'll put this up with a pull request I assume?

zuphilip

@adamsmith: Yes, done.

@r.douhi: I will try to find some time. Alternatively, you or your staff can write that as well (that is the nice thing in open source software and coding at github). Maybe you can convince one or your "catalog developer" to spend some time in writing a translator and share it with the rest of the world. I am pretty sure that there are much more ressources from Belarus, Russian etc. which Zotero doesn't cover at the moment.

aurimas

1) The et-al in the middle pane of zotero is independent of everything else here in this discussion. Maybe that is hardcoded and cannot localized. @Dan : Any thoughts on that?

It is localized, it's just not translated into Russian (looks like @r.douhi is using a Russian interface). In case someone is thinking of contributing translations, see here

aurimas

2) The doubling of the initials are a result of the MARC data, e.g.

702 #1 $3 BY-SEK-125257 $a Демиденко $b О. М. $g Олег Михайлович $c доктор технических наук $f род. 1960 $b О. М. $c доктор технических наук $f род. 1960 $g Олег Михайлович $4 340

There is twice the subfield b, which will lead to the doubling of initials in zotero. Does this make any sense in BELMARC?

We can actually improve on this. MARC is quite specific on which fields and subfields may and may not be repeated. $a and $b happen to not be repeatable (NR = not repeatable)

aurimas

NLB translator is now live btw