[Geodata] Re: Online conversion and validation services on OSGEO?
Jo Walsh
jo at frot.org
Fri Mar 16 11:16:27 EDT 2007
On Thu, Mar 15, 2007 at 06:29:05PM -0400, Frank Warmerdam wrote:
> Yes, I do think that some of the proposed metadata can be extracted
> automatically with OGR using ogrinfo, or perhaps better the python
> bindings. My understanding was that Jo was pursuing some of this,
> though I haven't been paying close attention.
Ja, your understanding is quite correct FrankW :) There's some
metadata autoextraction code using the python bindings to OGR in
amongst my python widgets, and similar using OWSLib to grab metadata
from different GetCapabilities and stick it into the same model.
I started running it over the many directories of undocumented geodata
on the freemap.in system and ran into a fun situation where the
daemon would find a set of tiled shapefiles before finding the
tileindex, and start creating records for 10,000 tiny datasets which
are all 'really' subcomponents of the same data set.
I don't see any specific provision for relating tiles to broader
datasets in the metadata standards i have been reading. I see concepts
of 'dataset series' which looks more like a generic term for a
collection of data sets which share some metadata properties (contact
info, possibly spatial extents) but not others (possibly spatial
extents, likely temporal extents)
Yet subset-of-tiles of a larger scope body of data is probably a
common thing to want to collect into an application and redistribute.
So i got this far before those couple of months mostly-offline.
I promised John and Norm something that could be run over some of the
telascience file storage and i would like to try this out soon,
a lot of the freemap.in data is handmade or deliberately
'experimental' and probably not typically representative of real world
environments.
What this *didn't* do was connect data just in local filestorage to
data available at, say, a HTTP URL. That sort of thing might be so
specific to an installation, or dare i say it to a 'workflow process',
that providing more than libraries and a pile of example scripts might
not really be worth it?
cheers,
jo
More information about the Geodata
mailing list