Re: xml representation of TDS dataset contents

• To: Ben@xxxxxxxxxxxxxxxx
• Subject: Re: xml representation of TDS dataset contents
• From: John Caron <caron@xxxxxxxxxxxxxxxx>
• Date: Tue, 24 Oct 2006 10:28:19 -0600

Hi Ben:

If you append "&wantXML" to the showForm URL, you get back the XML before its 
been XSL transformed to HTML, eg:

http://motherlode.ucar.edu:9080/thredds/ncServer/fmrc/NCEP/GFS/CONUS_80km/files/GFS_CONUS_80km_20061023_0600.grib1?showForm&wantXML

unfortunately, the returned XML is an unpublished format, and may change in the 
future. ignore the vertCoord amd missing stuff, and you see its a list of the 
available variables and what forecast hours they are available at. The forecast 
hours do change for some models depending on the run time. This is a 
complication for operational models like NCEP.

the XML returned by WCS is standardized but more complicated, eg:

 
http://motherlode.ucar.edu:9080/thredds/wcs/fmrc/NCEP/GFS/CONUS_80km/files/GFS_CONUS_80km_20061023_0600.grib1?request=GetCapabilities&version=1.0.0&service=WCS

we have a new "latest" dataset that should be useful. What problem are you 
trying to solve with knowing the available forecast hours?

Ben Domenico wrote:

Hi,

As many of you know, I've been working on scripts to automate the delivery and processing of subsets of datasets from TDS sites. I've managed to make this work for a few of the model output datasets (NAM, RUC on motherloade, WRF on lead4). However, it is clear that the approach I'm using -- basically scraping the HTML of the netCDFserver form for selecting variables, bounding boxes and forecast times -- presents problems because the HTML is very different for different models. In some cases, it actually seems to change from time to time even for the same model. Robb has mentioned XML documents that contain much of the same information in a more systematic and/or consistent format. I think that would be a better approach. Maybe the WCS interactions might be a better approach.

An example of the difficulties that arise is the fact that different variables have different forecast times and the lists appear in a variety of different ways in the netCDFserver forms for different models.

Please send me whatever suggestions you have for the best way to proceed.

Thanks.

-- Ben

PS I realize this is probably not the best time to be distracting you from the training workshop but I thought I'd mention it while it is fresh in my mind.