SFX Documentation

Google Books: a half-baked target parser

In response to a Hackfest proposal at this year's Access conference in Canada, I whipped together a target for Google Books. Here is some weak attempt to record what I did:

- Created a target called GOOGLE_BOOKS_LOCAL
- For this target, added the getWebSearch service, using the existing Google Scholar target and target parser as a template
- Modified the threshold set for Google Scholar to scan the incoming OpenURL for a book title - ($obj->need('bookTitle'))
- Copied the Title.pm target parser, modified a few lines to direct it to the right URL, use book title to populate the title token in the outbound URL, and added the little hook as_brr=1 which restricts a Google Books search to free full text only (i.e.- no snippet or limited preview titles)

That's about it. I did this in test, but thanks to Marjorie's tireless efforts to get the test to local copy cron job to work, it rolled quietly into production this morning. I'd go deactivate it, but it does work and is tucked at the bottom of the menu anyway. Incidentally, I declared--I believe correctly--that K-State now holds the Hackfest record for quickest concept to production curve!

LEXIS parser edit

The issue Judy pointed out and that is correctable is that the default date range for searching was going to the LN default, i.e.- last six months, when 'all available dates' is more appropriate. Our links from the SerSol ejournal lists were loading in LN with 'all dates' selected. Looking at the URL, I found that the snippet that sets that is after=0:ALL.

To correct the issue in SFX, I edited the target parser for Lexis (in the target parser folder it's /LEXIS/NEXIS.pm) on the global instance and added this snippet for the service getFullTxt. I need to check with EL how to make sure this isn't overwritten by an update. The only edit I made was to add that snippet to the URL mask:

$url = "$host1/sourceselect/returnToSearch.asp?after=0:ALL&csisr
c=$jkey&srcpdn=academic&cc=&spn=&product=universe&unix=$host2

Not sure why there are two getFullTxt stanzas, since they appear identical, but to be sure, I edited both.

The issue Judy pointed out and that is correctable is that the default date range for searching was going to the LN default, i.e.- last six months, when 'all available dates' is more appropriate. Our links from the SerSol ejournal lists were loading in LN with 'all dates' selected. Looking at the URL, I found that the snippet that sets that is after=0:ALL.

To correct the issue in SFX, I edited the target parser for Lexis (in the target parser folder it's /LEXIS/NEXIS.pm) on the global instance and added this snippet for the service getFullTxt. I need to check with EL how to make sure this isn't overwritten by an update. The only edit I made was to add that snippet to the URL mask:

$url = "$host1/sourceselect/returnToSearch.asp?after=0:ALL&csisr
c=$jkey&srcpdn=academic&cc=&spn=&product=universe&unix=$host2

Not sure why there are two getFullTxt stanzas, since they appear identical, but to be sure, I edited both.