Re: [greenstone-users] trouble with mgpp

From Katherine Don
DateMon, 19 Apr 2004 09:08:19 +1200
Subject Re: [greenstone-users] trouble with mgpp
In-Reply-To (40802BB8-TOP1S46)
Hi

MGPP's allfields element is perhaps badly named, as it doesn't mean all
the metadata available, but rather a combined index of all the
*specified* metadata.

So, 'indexes Title allfields'
actually only indexes the Title.

'indexes allfields Title Keywords Comment'
for example, would index all three metadata elements separately, then
provide a combined index over all three together.

Do you think that having an index over all the *available* metadata
without separate indexes for each one is a useful feature? If so, I'll
add it to the Greenstone TODO list.

Regards
Katherine Don

jraynor wrote:
> I am trying to build a collection of archival photographs, and am having
> trouble with mgpp.
>
> The documents being collected are all simply JPEG images.
>
> I put all of the metadata elements in a separate "metadata.xml" file, to be
> added to the individual documents with RecPlug's "-use_metadata_files" option.
>
> Each of the photographs already has a caption, so I created a metadata element
> called "Caption" to put it in.
>
> These captions, however, leave much to be desired (they're not descriptive
> enough to be useful for keyword-searching) so I added a second metadata
> element called "Keywords".
>
> Each photograph has a negative number, which I'm putting in the "Title"
> metadata element. This fact, however, is to be concealed from the users (note
> the "collectionmeta" statement which renames "Title" as "negative number" in
> the drop-down menu of searchable indexes).
>
> Here is one entry from the "metadata.xml" file:
>
> - - - - - - - - - - - - -
>
> <FileSet>
> <FileName>02632a-SmJ.jpg</FileName>
> <Description>
> <Metadata name="Title" mode="override">02632a</Metadata>
> <Metadata name="Caption" mode="override">R. I. Red Festival.</Metadata>
> <Metadata name="Keywords" mode="override">exterior cooking
> barbeque</Metadata>
> <Metadata name="Date" mode="override">20010625</Metadata>
> </Description>
> </FileSet>
>
> - - - - - - - - - - - - -
>
> I would like users to be able to search for keywords *both* the *Caption*
> field and the *Keywords* field at once, and have thus opted to use mgpp's
> "allfields" option to create a searchable master index.
>
> Here's the problem:
>
> Right now, I *can* search the "allfields" index for titles (e.g. searching for
> "02632a" yields the document whose metadata is shown above).
>
> However, searching the "allfields" index for anything that I know to be in a
> Keywords or Caption metadata element (e.g. "Red" or "cooking" from the example
> provided above) yields absolutely *nothing*.
>
> I'm very puzzled.
>
> The only explanation that I can think of is that mgpp is confused by
> non-standard metadata elements (neither "Comment" nor "Keywords" correspond to
> anything in the Dublin Core). On the other hand, Greenstone seems to be
> perfectly capable of sorting the documents alphabetically by Caption
> ("classify AZList -metadata Caption" is working nicely).
>
> Any help would be most gratefully appreciated! Thanks!
>
> Here is the collection's configuration file:
>
> - - - - - - - - - - - - -
>
>
> creator jraynor@mail.uri.edu
> maintainer jraynor@mail.uri.edu
> public true
>
>
> searchtype plain form
> indexes Title allfields
> collectionmeta .Title "negative number"
> collectionmeta .allfields "keyword"
>
> plugin ImagePlug
> plugin GAPlug
> plugin ArcPlug
> plugin RecPlug -use_metadata_files
>
> classify AZList -metadata Title
> classify AZList -metadata Caption
> classify DateList
>
> collectionmeta collectionname "ArcPhot April 13th Number 1"
> collectionmeta iconcollection ""
> collectionmeta collectionextra "This is just the April 7th version, with
> mgpp instead of mg."
>
> format VList "<td>[link][icon][/link]</td>
> <td valign=top><b>Negative Number:</b>
> [Title]<br><b>Caption:</b> [Caption]<br><b>Added Keywords:</b>
> [Keywords]<br><b>Date:</b> [Date]</td>"
>
> format DocumentText '<center><table>
> <tr><td><img src="/gsdl/collect/gsarch/index/assoc/[assocfilepath]/[Image]">
> <br>Negative Number: [Title]<br>Caption:[Caption]<br>Keywords:
> [Keywords]<br>Date: [Date]
> </td></tr></table></center>'
> format DocumentButtons ""
>
>
> _______________________________________________
> greenstone-users mailing list
> greenstone-users@list.scms.waikato.ac.nz
> https://list.scms.waikato.ac.nz/mailman/listinfo/greenstone-users
>