search

NOAA-PMEL / LAS

Live Access Server
https://ferret.pmel.noaa.gov/LAS/
The Unlicense
13 stars 5 forks source link

addXML needs to create configuration for curvilinear data #388

Closed karlmsmith closed 6 years ago

karlmsmith commented 6 years ago

Reported by @kevin-obrien on 7 Feb 2008 22:14 UTC None

Migrated-From: http://dunkel.pmel.noaa.gov/trac/las/ticket/382

karlmsmith commented 6 years ago

Modified by @noaaroland on 29 Feb 2008 21:13 UTC

karlmsmith commented 6 years ago

Comment by @AnsleyManke on 16 Oct 2008 22:03 UTC The properties that the Ferret scripts use to determine that a dataset (or variable) has curvilinear grid are Ferret properties curvi_coord_lon and curvi_coord_lat

The scripts also either need to compute or be given the range of the curvilinear longitudes, so that they can figure out what to do with the x range of data requested. This range can be passed with properties curv_lon_min and curv_lon_max. If these aren't present, Ferret computes them by just reading the curvilinear coordinate variable.

The other thing needed in the xml configuration is ranges of X and Y to present on the user interface. My documentation for configuring these datasets by hand has the configurer just use the ranges from the curvilinear coordinate variables and define some axes with these ranges.

http://ferret.pmel.noaa.gov/LAS/documentation/installer-documentation/adding-curvilinear-data

The Ferret scripts COULD figure all this out themselves; if they have a property that says this is a curvilinear dataset, and if the dataset contains a coordinates attribute, then Ferret can read that attribute, perhaps do some checking on the dimensions and ranges of the coordinate variables, and then call the right scripts to use the curvi data.

karlmsmith commented 6 years ago

Comment by steven.c.hankin on 16 Oct 2008 22:23 UTC Very helpful. Thanks.

To make the curvilinear scripts robust and high performance, the best strategy would seem to be that addXML should pre-compute whatever it can -- curv_lon_min and curv_lon_max. And also that the Ferret script be able to compute whatever it finds to be missing. For high resolution datasets, pre-computing the limits X limits is a performance issue. Similarly the full configuration of a curvilinear dataset will also require the creation of a remap grid, won't it? (Used to determine the IJ range needed to encompass a particular XY range.)

Whether addXML should include the attributes curvi_coord_lon and curvi_coord_lat is debatable. My impulse is to say yes, but I could be talked out of it.

karlmsmith commented 6 years ago

Comment by @noaaroland on 19 Dec 2008 14:16 UTC Checked in with r3525.