[Fedora-commons-users] Fedora 2.2 and ImageCollection demo problems

[Scott Prater]

Hello,

I've run into a problem with the ImageCollection demo "SmileyStuff" in
Fedora 2.2. When I attempt to access the collection "view" disseminator:

http://myhost/fedora/get/demo:SmileyStuff/demo:Collection/view/

I get a blank page, and the following Java exception in catalina.out:

Error on line 6 column 6 of null:
Error reported by XML parser: The element type "meta" must be
terminated by the matching end-tag "</meta>".
javax.xml.transform.TransformerException: org.xml.sax.SAXParseException:
The element type "meta" must be terminated by the matching end-tag
"</meta>".
[...]

And when I try the "list" disseminator:

http://myhost/fedora/get/demo:SmileyStuff/demo:Collection/list/

I get a HTML page showing a different Java exception:

Unsupported tuple query language: itql

org.trippi.TrippiException: Unsupported tuple query language: itql
at org.trippi.impl.mpt.MPTSession.query(MPTSession.java:112)

[...]

I was able to configure and build the Resource Index, and I've been able
to run all the other demos without any problems.

I'm using Postgresql as my backend database, both for the regular Fedora
datastore (localPostgresqlPool) and for the resource index
(localPostgresMPTTriplestore). I've configured the resource index
according to the instructions in the Fedora 2.2 documentation, and I've
confirmed that the database user can connect to both the fedora database
and the triplestore database. I've been able to rebuild the triple
store a number of times, but to no avail.

I saw that MPTStore does not support itql or rdql, but only SPO. Could
my problems be arising from that fact that I don't have tools for
searching and retrieving from the MPTStore in Postgres? (As an aside,
possibly related: I can start up /fedora/risearch, but only
"unsupported" appears in the "Language" list under the "Find tuples" tab.)

Any help would be much appreciated. I can give more details about my
environment, if needed.

thanks,

-- Scott

--
Scott Prater
Library, Instructional, and Research Applications (LIRA)
Division of Information Technology (DoIT)
University of Wisconsin - Madison
prater -> wisc.edu

[Aaron Birkland]

> I saw that MPTStore does not support itql or rdql, but only SPO. Could
> my problems be arising from that fact that I don't have tools for
> searching and retrieving from the MPTStore in Postgres? (As an aside,
> possibly related: I can start up /fedora/risearch, but only
> "unsupported" appears in the "Language" list under the "Find tuples" tab.)
>

SPO is the only language that MPTStore supports that you can use through
the risearch interface, so what you're seeing is normal. If you wish
to do complex queries, you can do so by using the MPTStore client
libraries directly, if that's possible in your situation. Currently,
however, there is are no grammar parsers for a language more
sophisticated than SPO. There are no plans for an itql or rdql parser
(itql is essentially kowari/mulgara specific, and rdql is deprecated, I
believe), but a sparql parser is high on the wish list for MPTStore.

-Aaron

[Scott Prater]

Thanks, Aaron.

So is the Image Collection demo designed to work only with Kowari and
itql? To get it to work, should I write demo:Collection disseminators
for "view" and "list" that utilize MPTStore and the SPO query language?

Pardon the greenhorn questions... I'm coming up to speed on several
things at once here.

thanks,

-- Scott

--
Scott Prater
Library, Instructional, and Research Applications (LIRA)
Division of Information Technology (DoIT)
University of Wisconsin - Madison

pra...@wisc.edu

[Wallace, Iain]

*** Apologies for cross-posting ***

I hope this job will be of interest to some people on this list - the focus is very much on Fedora development.

Digital Repository Developer
Spoken Word Services - Saltire Centre
Glasgow Caledonian University, Glasgow
up to £25,070 per annum (currently under job evaluation review) Fixed term until end of July 2008 in the first instance Closing date: Friday 25th May 2007 Post No: T.17199

Please note: this job is based in Glasgow, UK

Spoken Word Services is based in the Library within the multi-award winning Saltire Centre. The team maintain and develop a repository of several hundred hours of digital audio and video resources for the educational community. These resources are sourced primarily, though not exclusively, from the BBC Television and Radio Archives.

The Digital Repository Developer will build on existing technical infrastructure, to develop a Fedora-based digital object repository and associated tools, and migrate content from the existing MySQL/PHP repository. The central task of the Developer will be to manage and develop the repository, containing digital images, text, audio, video, metadata and other content. The Developer will also be expected to adapt and develop a set of tools and services to support ingestion, search and discovery, digital object delivery and presentation, annotation and preservation.

Candidates should have an undergraduate degree or equivalent experience, and experience in the development of complex software systems, including work in Java. The successful candidate will be self-motivated, able to demonstrate a logical approach to problem solving and a keen, analytical mind. As the post operates as part of a team candidates must demonstrate good communication skills and the potential to work within a cross-disciplinary group.

Please note that CVs will not be accepted for this post. 'Glasgow Caledonian University is committed to being an equal opportunities employer'. Please note applicants are strongly encouraged to apply online - to do so visit www.caledonian.ac.uk/jobs/. However, if you do not have access to our website please call 0141 331 8864. Closing date: Friday 25th May 2007.

To discuss this job informally please contact Iain Wallace on +44(0)141 273 1901.

For more information on Spoken Word please visit: http://www.spokenword.ac.uk

------------------

Iain Wallace
Digital Services Development Librarian
Spoken Word Services
The Saltire Centre
Glasgow Caledonian University
Scotland, UK
web: www.spokenword.ac.uk
email: iain.w...@gcal.ac.uk
tel: +44 (0)141 273 1230

[Aaron Birkland]

Scott,

> So is the Image Collection demo designed to work only with Kowari and
> itql? To get it to work, should I write demo:Collection disseminators
> for "view" and "list" that utilize MPTStore and the SPO query language?
>
> Pardon the greenhorn questions... I'm coming up to speed on several
> things at once here.
>

Ahh, I see. I've actually never seen or used that demo. As far as I
understand it, that demo pre-dates MPTStore (which is a really new,
alternative triple store that was just introduced in the last
release). You would probably have to do a fair bit of modifications to
get it to work with MPTStore. I'm really not familiar with the demo, so
I don't know what basic functionality the demo requires to function
(what types of queries, etc), but I imagine you would probably have to
write your own adaptation using the MPTStore libraries directly if it
really needed graph queries of any complexity. I can help you out if
you'd like to know how to formulate complex queries in MPTStore, but as
far as the specifics of the image collection demo goes, I really don't
know anything about it. That probably wasn't very helpful. Sorry!

-Aaron

[Scott Prater]

Actually, that was quite helpful. At least now I know the scope of the
problem, and I'm starting to feel relatively confident that the errors
running the demo haven't arisen as a result of misconfiguration.

Rewriting the demo to use MPTStore might be a good starting exercise for
me. But in the meantime, I have a lot more ground to cover.

thanks!

-- Scott

Aaron Birkland wrote:

[Alicia López Medina]

Hi all:
We want to include our audio/video collection in our fedora repository
and we would appreciate some advice about the metadata schema that we
could use, besides Dublin Core. Is anyone using MPEG tools?. How does
this standards fit with Fedora datastreams? It will help very much if
you could provide us with an example of a video/fedora object in xml.

Thanks a lot!

alicia.

[Susan Schreibman]

Dear Alicia,

We've been using video (and now working on audio) in Fedora. Jennifer
Roper will be getting in touch with you to tell you how we've handled
the metadata (not till next week though). In the meantime, you may want
to look at our documentation & DTDs

http://www.lib.umd.edu/dcr/publication/

susan

> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> _______________________________________________
> Fedora-commons-users mailing list
> Fedora-com...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/fedora-commons-users
>

--
Susan Schreibman, PhD
Assistant Dean
Head of Digital Collections and Research
McKeldin Library
University of Maryland
College Park, MD 20742

Phone: 301 314 0358
Fax: 301 314 9408
Email: ssch...@umd.edu

http://lib.umd.edu/dcr
http://irith.org
http://macgreevy.org

[fe...@fedora-commons.org]

Alicia,

Rutgers developed a METS implementation, with descriptive (MODS), source
(for analog materials), technical and rights metadata. Our METS
administrative metadata incorporates PREMIS, MIX (images), video metadata
and the recently developed AES sound metadata. We received a contract
last year to develop the bibliographic utility for the Library of Congress
to use with Moving Image Collections portal with moving image archives.
Our video metadata incorporates data elements from PBcore, MPEG-7, and the
Library of Congress technical metadata for video. We've had lots of
assistance from members of the Association of Moving Image Archivists
(co-sponsors of the MIC portal). As one of the earliest adopters of
MPEG-7, we incorporated a lot of that metadata into our bibliographic
utility. I'm also a metadata advisor for PBcore, so I drew heavily upon
that excellent schema, particularly since we want strong interoperability
with PBcore, a very important video metadata schema, IMO.

Our bibliographic utility will go open source this spring and our WMS
(workflow management system), which incorporates file ingest and handling
and interfaces with Fedora this summer. Our WMS has become really
functional over time, since we use it to support participants across the
state for our statewide cultural heritage portal, which is a Fedora
repository. The WMS allows us to support a rich metadata infrastructure,
to abstract users across the state from direct contact with the Fedora
repository when they create metadata and upload objects, but they are now
able to easily access metadata records that are ingested into the fedora
repository, pull them into the WMS to make corrections or additions, and
reload to Fedora, very quickly, over the web. Since we use a lot of event
metadata, particularly in descriptive and source where you may want to
capture a condition assessment or physical intervention or use of a
resource in an exhibit, people can now quickly access collection or item
level records, add the event to the metadata and re-ingest into Fedora, in
a very short period of time.

We are currently working on a more nuanced administrative module and
collection owners portal (to run reports on types of files, condition,
use, etc.) but hope to be complete shortly for delivery to the Library of
Congress.

You can learn more about our WMS at: http://rucore.libraries.rutgers.edu
in the developers area, or feel free to contact me off list for more
information. I think we definitely have the most comprehensive video and
audio metadata possible, thanks to the experts at the Library of Congress,
PBcore, and within AMIA, and to our own knowledge of MPEG-7, but since we
also support very small archives and libraries, the template features of
the WMS can enable you to create very simple templates that ignore and
hide any high end metadata that you don't need--until the point where you
need it. You can also read about our product in The Code4Lib Journal
online: http://journal.code4lib.org/articles/25

Grace Agnew
Rutgers University Libraries