interested in using kochief

[AustinSudbury]

Howdy,

I found this project through oss4lib, looking for open source library
managing tools for a library we are starting at our school and another
organization. The demo sites I looked at were pretty slick.

Before I go too deep into this though I wanted to clarify some
things.

Kochief isn't really for generating or editing the catalog (right?) So
I would need to write some code outside of kochief. Using PyMarc for
the record managing, this backend would involve managing borrower
accounts, adding/editing records, and designating books as checked out
or checked back in. (real simple command line stuff is what I am
thinking).

Kochief can use MARC records? changing records (for check out/in for
instance) from a local computer and updating kochief is the way to go?

I know python, run linux and have done some things with web servers
before, but I generally have had poor luck with compiling programs.
Also, our website is currently hosted on godaddy. Is it likely that I
will be able to sucessfully build and run kochief? Or is it a pain to
build or requires more server control than godaddy allows?

[Gabriel Farrell]

Correct, Kochief doesn't do much in the "managing records" department.
I had some vague notions in that direction at one time -- see the
"cataloging" app in Kochief for my stabs there. Anything you'd be into
contributing would be welcomed.

You're also right in assuming that Kochief is pretty focused around
MARC records at this point. There are a couple of example alternatives
in discovery/parsers/ if you need to parse another format. At Drexel,
I index MARC dumps from our III catalog in Kochief nightly.

GoDaddy is rotten in many ways, including Django support. You'll have
some trouble there. If you have any flexibility I'd recommend
switching to WebFaction.

Great to see the interest in Kochief. Feel free to post any questions
you have to this list.

> --
> You received this message because you are subscribed to the Google
> Groups "Kochief" group.
> To post to this group, send email to koc...@googlegroups.com
> To unsubscribe from this group, send email to
> kochief+u...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/kochief

[...]

As far as the example pages go, Paul smith's site (http://library.paulsmiths.edu/catalog/) looks really good.  I especially like the break down of search criteria on the left side.

At drexel, it seems kochief is only sued for their beta Collections section, and the rest of the library is something else (proprietary).

this is not kicohief?:
http://innoserv.library.drexel.edu/search/

So, is making something as clean and clear as paul smith's page very (very) labor intensive, or is drexel trying to use only a small portion of kochief for their very small collection?  Like, where is search by subject or author? 

thanks,
Elliot

[Ross Singer]

On Sat, Sep 25, 2010 at 5:28 PM, ... <offonoff...@gmail.com> wrote:
> As far as the example pages go, Paul smith's site
> (http://library.paulsmiths.edu/catalog/) looks really good.  I especially
> like the break down of search criteria on the left side.
>

Well....

Paul Smith's is actually using VuFind: http://vufind.org/

They used to use Kochief, but migrated to VuFind a few months ago.

> At drexel, it seems kochief is only sued for their beta Collections section,
> and the rest of the library is something else (proprietary).
>
> this is not kicohief?:
> http://innoserv.library.drexel.edu/search/
>

That's just III's Innopac. I believe Drexel is about to implement
Summon (or maybe has) for the primary search.

-Ross.

> So, is making something as clean and clear as paul smith's page very (very)
> labor intensive, or is drexel trying to use only a small portion of kochief
> for their very small collection?  Like, where is search by subject or
> author?
>
> thanks,
> Elliot
>

[...]

On Sat, Sep 25, 2010 at 5:23 PM, Ross Singer <rossf...@gmail.com> wrote:

On Sat, Sep 25, 2010 at 5:28 PM, ... <offonoff...@gmail.com> wrote:
> As far as the example pages go, Paul smith's site
> (http://library.paulsmiths.edu/catalog/) looks really good.  I especially
> like the break down of search criteria on the left side.
>

Well....

Paul Smith's is actually using VuFind: http://vufind.org/

They used to use Kochief, but migrated to VuFind a few months ago.

mm, so not really a good demo site for y'all.  I guess I ill look more at VuFind. 

oh, but I do like python much more than php, though I dont have any drupal experience...

[...]

It looks like I am going to write some ILS scripts in python because existing ILS programs do not support the book lending concept out library will be using.  I plan to write these and then write a driver for VuFind:  http://vufind.org/wiki/building_an_ils_driver

My scripts will take care of creating borrowers, checking books out and in, calculating fines, probably end of day tasks (sending out "due soon" emails).  Also, towards out book lending concept, it will take care of creating book lenders and keeping track of book lease times.  It will be able to return the data necessary for a vufind driver, but university stuff like courses and instructors wont be created.  Not planning on doing holds either.

Also, I'm writting some scripts using pymarc for importing records from the library of congress, checking them for completeness and adding local information (call number, etc).

Since i'm doing it from scratch, I could easily do this in a way to make it as easy as possible to integrate into kochief at some point also.  Any thoughts on this?

thanks,
Elliot

[...]

for instance, where the heck is you're "book" model?  I would be happy to reuse it, but all I find is the Resource(object).

I was expecting something like

---
from django.db import models

class Book(models.Model):
    title = models.CharField(max_length=100)
---

[Mark A. Matienzo]

Kochief doesn't use models for the most part. It is mostly a discovery
interface over a Solr index, just like VuFind and Blacklight. Gabe did
do some work on a "cataloging" application which was previously
referenced in this thread, which is where you're getting the Resource
model from.

Mark A. Matienzo
Digital Archivist, Manuscripts and Archives
Yale University Library

[...]

On Tue, Oct 5, 2010 at 7:34 PM, Mark A. Matienzo <mark.m...@gmail.com> wrote:

Kochief doesn't use models for the most part. It is mostly a discovery
interface over a Solr index, just like VuFind and Blacklight.

ok, I assumed VuFind (and kochief) took marc records and stored them in a database.  (then, I'm assuming the way for django to access a db is through having model objects referencing the tables). I'll have to look more at what a solr index is.
 

Gabe did
do some work on a "cataloging" application which was previously
referenced in this thread, which is where you're getting the Resource
model from.

yes, and so that is a small model that is somehow mapped to the solr index or marc records.  so each marc record has a resource object that knows about it.

Maybe I'm in over my head trying to contribute to this project.  But a django ILS seems easily achievable.  It would probably be very good for me to understand how marc records are used (if not just loaded into a database) in kochief before I make the mistake you avoided, of creating book objects with titles and authors.

thanks,
Elliot

[Gabriel Farrell]

I had created the Resource model while looking in the direction of
Linked Data. At first I thought I could plug into a triplestore like
RDFLib but decided to abuse the Django ORM instead. I also considered
using a document store like CouchDB, or again twisting the ORM in a
way similar to ThingDB, since Solr is document-oriented.

Much of this was an attempt to keep Kochief more flexible, able to
ingest according to a wide variety of datamodels. For your own
purposes, however, you might simplify things by using the standard
Django models in the way they're meant to be used, with a Book class
and a Journal class and whatnot.

[...]

 For your own
purposes, however, you might simplify things by using the standard
Django models in the way they're meant to be used, with a Book class
and a Journal class and whatnot.

I''m still confused about the use of marc records.  how should the book model be tried to actual records? or, more specifically, how could an item returned from a solr search be tied to a django db item used in circulation calculations.  MARC records dont have a primary key (that I know of).  So importing all the records (title, isbn and all) into a django.db.model doesn't ensure that solr and circulation are refering to the same thing. 

[Gabriel Farrell]

Documents in Solr must have an ID. The parser assigns them on import.
Records from III, for example, get the III BibID. The default we use
for MARC record parsing is the 001 field.