SlacSciDoc overview

SciDoc or TechPubs are interchangeable terms to describe the process of managing SLAC published papers. This is done by Maggie Loera, and aided by Arsella Raman on the Spires side. Simply, SLAC documents are submitted through a registry: http://www-public.slac.stanford.edu/sciDoc/register-document.asp and managed in a Filemaker database.

File definitions:

• SciDoc The filemaker database of metadata for SLAC papers.
• SLAC the Spires database of SLAC record transactions, containing linis to Spires HEP, some SciDoc metadata, and OSTI transactions
• HEP the Spires database for High Energy Physics
• Inspire The Invenio database for High Energy Physics

SLAC authors are supposed to register their papers. If they do not, TechPubs (Maggie) searches out their publications via SLAC websites and pushes them into SciDoc. This process is called the Leaklist. There is a daily process run on account SLACLIB that runs a spires protocol called ..hep.leak.to.slac. This runs daily on a cron on account SLACLIB. This cron invokes a number of scripts that all run in sequence for various library and former hep functions through a spires file called spirestasks.

hep.leak.to.slac checks HEP to see if SLAC papers have been added into HEP spires and SciDoc missed them, usually submitted to arxiv. There is a further process I'm not quite clear on that scrapes LCLS and SSRL sites for papers and checks that they are in. If not, they are added. One of these scrapes works, one stopped working at sul's last recollection.

The result of this work gets presented on this page: http://www.slac.stanford.edu/~spiwork/leaklist.html

This allows Maggie to visually compare records from SLAC, SciDoc and HEP. One of the links on each record is a MERGE button. This invokes the following url: "http://spires.slac.stanford.edu:5095/MERLEAK/FIRSTSTEP" method="post"><input name="IRN" type="hidden" value="174548" This either adds or merges data into the SLAC spires subfile records.

The above url is what webspi1 listens for on sunspi4, port 5095. It invokes a spires protocol setup called webservices which is also used by the library applications. If that server is down, it can be restarted by:

• ssh webspi1@sunspi4 (secondary login, shouldn't get a password prompt) then type:
• /u/lw/webspi1/bin/spires_server -rl -p allweb -t web -d web

If the response doesn't show it 'killing' the process, then the process was indeed not running, and somehow the daily cron that is supposed to restart it, didn't!

Arsella comes in twice a week, logs on to sunspi4, calls spires and runs a spires protocol called:

• ..slac.techpubs

She is presented with a choice menu, and her task is to put the collected metadata that has come from Maggie to SLAC into the HEP file, make sure the matches are there, and if not, try to find one.

This process marks SLAC records to be ready for OSTI. OSTI harvests the records once a week on Sunday night, and sends a receipt back to SLAC. The URL OSTI uses is:

• http://www.slac.stanford.edu/spires/find/slac/xmlosti?rawcmd=find+slacr+slac+and+du+%3E%3D+[START]+and+du+%3C%3D+[STOP]+and+tpd+after+9%2F2004+not+harvestid+sgml

this url needs a start and stop date to see the output

After harvesting, OSTI sends a confirmation email with an xml file to slaclib2@sunspi4NOSPAMPLEASE.slac.stanford.edu

This email with a specific subject triggers a procmail script that:

• saves the xml receipt to this directory: /afs/slac.stanford.edu/www/spires/slac/admin/osti_receipt/
• loads the OSTI_ID into SLAC record

Once a month, metadata that has been collected by the process into Spires HEP is culled by the Inspire SLAC curators and merged or added into Inspire in this process:

• SlacInputting#Inputting_SLAC_Spires_HEP_record

Things that can go wrong

• there is a cron running on the SciDoc side, if it dies it's a Ray Horgan problem, don't know much about it.
• the webspi1 post server could be down preventing Maggie from posting changes to SLAC. Log into webspi1 and run this script: /u/lw/webspi1/bin/spires_server -rl -p allweb -t web -d web
• Leaklist isn't being generated. the spirestasks daily cron running? look at the leaklist in directory /afs/slac/u/li/spiwork/public_html and see if leaklist.html is current. If not you can invoke the protocol by this command from account slaclib: /afs/slac/g/library/bin/spirestasks.pl -k1
• Arsella not receiving records to curate: she'll check with Maggie if there's been records coming.
• OSTI url not working. try url with other dates. In SLAC, search the last week 'find du after today-7 and before today and not harvestid smgi" Show if records are there to harvest.

-- MikeSullivan - 2015-04-09