29.05.2012 WP9 Geoprocessing

WP9 Skype 29 may 2012

Participants: G. Coro, F. Brito, F. Barchetta

Topics

• Summary on Thredds activities;
• Summary of the work about WPS.

On Update on Thredds activities

GC illustrates the procedure implemented so far about the invocation of Thredds by a Java library (Environment Explorer Library or EnvEx)

GC demonstrates how the layers contained in the NetCDF files, lying on Thredds, can be indexed on the GeoNetwork by means of another automatic procedure in EnvEx

A discussion on netCDF-CF files follows:

FBrito says that the files the community is interested in, belong to the NetCDF-CF category

GC says that formally there can be files with a NetCDF-CF structure which are not associated to layers but that can be, for example, images or points-trajectories. The files we may be interested into are the NetCDF-CF ones presenting a GRID structure

FBarchetta says that the NetCDF-CF files from MyOcean are generally of GRID type

FBrito proposes to GC to test the system with some files from MyOcean which Terradue will provide

FBrito says that Terradue is interested in WCSWeb Coverage Service access to the files, because this would allow parallelization on the WPS-Hadoop framework

FBrito advices the usage of WPS-Hadoop for all the geo spatial processing

GC will introduce a method, in EnvEx, for retrieving the WCSWeb Coverage Service address for a certain netCDF layer from Geo Network

GC explains the details of the designed architecture for future usage by the D4ScienceAn e-Infrastructure operated by the D4Science.org initiative. services. More information can be found here

FBrito says it is important that even the suggestions by IRD, about the evaluation of the geospatial processing quality, will have to be taken into account by WPS plugins

Summary of the work about WPS

The discussion moves on WPS:

FBrito explains that the activity on the WPS Intersection process between two WFSWeb Feature Service layers was asked by FAO and indicated as an activity in WP3.

GC remarks that geospatial activity on WPS should be part of WP9 or at least mentioned in WP9

FBrito explains that the indication came from FAO, as demonstrated by some WP3 meeting reports

GC says he missed such information as he is not fully involved in WP3

GC reports that, currently, the activity on WPS at CNR is investigating more on understanding WPS and Hadoop separately, before going to understand the details of the framework by Terradue

FBrito says that all the activity on WPS should rely on the Terradue framework

GC agrees but explains that the early tests with the package distributed by Terradue returned some errors. It was our opinion to be more expert about the arguments before asking to Terradue for help

FBrito advices that the right way should be to make the Terradue package work on CNR machines and then to investigate about the details. He offers their support for rapidly solve the possible issues

GC agrees

Both the parts agree on going forward in the installation and usage of the WPS-Hadoop framework.

Short terms plan:

• Test of the Thredds Library on NetCDF files from MyOcean
• WPS-Hadoop installation troubleshooting at CNR
• Add of a method for retrieving the WCSWeb Coverage Service address from a layer on the Geo Network