Yes, I think I should have said CSVPlugin. CSVPlugin is used when each
row of the csv file is its own separate "document" or record.
MetadataCSVPlugin is for csv files which contain metadata to be attached
to some other file, eg a PDF, or word file etc. There needs to be a
document field which contains the name of the doc to attach the metadata
to. So sorry about that.
For an "untrained non-techie", Greenstone 2 is definitely the way to go
at the moment. Greenstone 3 is still not as stable or as complete as
Greenstone 2, and is more for people who want to get in and heavily
customise their Greenstone.
Regarding metadata, the greenstone runtime doesn't care what metadata
you use. GLI offers metadata sets, which can be easily extended (via
manage metadata sets, select the set you want, and click edit).
For a csv file though, just put any names you want in the field names
line, and greenstone will use them. As they don't show up in GLI enrich
panel anyway, it doesn't matter that they are not officially part of the
You won't get them showing up in drop down lists but you can use them
If you want to specify that some are dublin core and some aren;t, use
the dc namespace on the dublin core ones, then no namespace (or your own
one) for the other fields.
does this make sense? the only place this might fall down in GLI is when
defining indexes. I think you get to select from a list, you can't just
free type. If the new metadata names have no namespace, then they should
show up as extracted metadata I think. (but haven't tried).
Anyway, have a go at just defining new field names in your csv file, and
see how you get on.
(Note, I have sent this on greenstone-users mailing list instead of
greenstone3 mailing list.)
Joy Payton wrote:
> Hi Katherine,
> Funny you should mention it, I've been back-and-forthing the CSV idea
> on the spanish language mailing list, mostly because my fellow library
> types here (we're all completely untrained non-techies) don't have the
> command of English that I do (I live in Spain although I'm from the
> USA). So MarcEdit wasn't a good long term solution and I needed a
> Spanish-friendly option.
> I've just yesterday done a trial (on GS 2.82 now, I've chickened out
> of 3 for the moment) of CSV files with Dublin Core metadata labels as
> the first row. (incidentally... take this for what you will...
> MetadataCSVPlugin caused me no end of problems, I removed it, and IT
> STILL WORKS... I still get my fields associated with the metadata in
> the first row using only CSVPlugin, and the labels in Spanish.)
> I only wish there were more metadata fields in DC... I have read
> somewhere that one can define her own metadata, but I'm Greenstoned
> out right now and contemplating adding my own metadata layer plus the
> CSV conversion makes me a little nervous. So I'm killing time by
> making cute icons to distinguish my collections when cross-referencing
> searches are done.
> The display I got hammered out by html tables in the GLI.
> Thanks for the reply!
> Joy, aci
> 2009/9/12 Katherine Don <firstname.lastname@example.org
> Hi Joy
> 3 hours seems like a long time. Do you particularly need MARC
> format? You can just use csv files directly in greenstone - they
> need the field names to be the first row in the file. And you need
> to add MetadataCSVPlugin to the plugin list. Maybe that will be
> faster to process.
> To get a better display, you need to edit the format statements.
> Which version of greenstone 3 are you using? 3.03 or 3.04 or other?
> Use <gsf:metadata name="Title"/> type elements to output different
> metadata values.
> For search results, modify the search documentNode template. For
> doc display, I don't think you can currently do it in GLI - you
> need to add a documentContent template inside a format element
> inside the display element.
> Take a look at the manual in greenstone3/docs/manual - this is
> incomplete and needs updating but there is quite a lot about
> format statements in here.
> I hope this helps a bit,
> Joy Payton wrote:
> Update: I tried my build again and waited 3 hours... it
> finally built and my primitive search shows that my
> bibliographic records are intact and searchable. Now I'm off
> to find or build an interface that will display the output in
> a user-friendly manner... anyone point me in the direction of
> a results interface for MARC records? The standard output in
> greenstone 2 was a document icon plus the book title, but
> greenstone 3 only gives me the document icon without anything
> else. When I click, I get the MARC record, which I will want
> to put in a more attractive format than a raw record with codes.
> Sr. Joy, aci
> 2009/9/2 Stephen De Gabrielle <email@example.com
> Did you try splitting the marc file into smaller parts.
> It could be the MARCPlugin is choking on the file for
> some reason.
> On Wed, Sep 2, 2009 at 7:20 AM, Joy Payton
> <firstname.lastname@example.org <mailto:email@example.com>
> <mailto:firstname.lastname@example.org>>> wrote:
> Hi friends,
> I'm trying to convert my convent library to
> Greenstone 3. We
> have about 3400 bibliographic records that were
> previously in
> an Access DB. From that, I eventually whittled the
> bibliographic data to a CSV file and now I have a .marc
> (generated from said CSV text file by MarcEdit) which
> then be builtable by the librarian interface. But,
> when I try
> to build, my .marc file hangs at 40 % complete... the
> GLI is
> using MARCPlugin as expected, but it just doesn't get
> 40%. I did check the validity of the MARC file using the
> validator in MarcEdit and it comes thru without errors.
> Sister Joy, aci
> Greenstone3 mailing list
> Stephen De Gabrielle
> Telephone +44 (0)20 85670911
> Mobile +44 (0)79 85189045
> Greenstone3 mailing list