PPUG_pp_config

Configuring the post processing case files

The PP_CASE_PATH created by the create_postprocess command contains all the necessary XML files, batch submission scripts and some helper scripts to run the post processing tools.

The env_*.xml files in the PP_CASE_PATH, with the exception of env_timeseries.xml, adhere to the following entry tag schema:

<entry id="ID" value="VALUE"/>

where each ID and VALUE pairs are used to define environment variables used by the post process wrapper scripts and their calls to the underlying NCL, python and perl scripts.

The env_*.xml files include:

• env_postprocess.xml
• env_timeseries.xml
• env_diags_atm.xml
• env_diags_ice.xml
• env_diags_lnd.xml
• env_diags_ocn.xml

Every ID in the env_diags_*.xml file is proceeded by either ATMDIAG_, ICEDIAG_, LNDDIAG_ or OCNDIAG_. This constraint ensures unique ID and VALUE pairs associated with each component.

You can edit the env_*.xml files directly in a standard text editor (vi, emacs, etc...) and then validate the XML using the following command:

>xmllint --schema env_files.xsd env_diags_[atm, ice, lnd, ocn, postprocess].xml

** OR **

You can use the pp_config script to "get" and "set" XML values. (TODO - add synopsis).

The pp_config script can be called from other scripts as well in order to create custom work flows.

The env_timeseries.xml file

The env_timeseries.xml file adheres to a different XML schema. The env_timeseries.xml file defines the rules and settings for the history time slice to variable time series parallel python conversion tool which is called by the **timeseries** batch submission script.

The basic env_timeseries.xml XML schema is defined as follows: (TODO - add a single stanza)

You must edit the env_timeseries.xml file manually using a text editor (vi, emacs, etc...) and then validate the XML using xmllint as follows:

>xmllint --schema $POSTPROCESS_PATH/Config/config_timeseries.xsd env_timeseries.xml

The submission script wrappers

Running the post processing scripts follows a logical step-by-step approach to generating the desired output data whether they are variable time series files, diagnostic plots, or CMOR compliant output files. The parallel python tools called by the wrapper scripts are modular in design.

The submission script wrappers contain batch submission stanza settings associated with machine that you are using. These wrapper scripts include:

• timeseries
• atm_averages
• atm_diagnostics
• ice_averages
• ice_diagnostics
• lnd_averages
• lnd_regrid
• lnd_diagnostics
• ocn_averages
• ocn_diagnostics

Depending on the queueing system of the machine, you will need to submit each individual script. For example, on yellowstone which uses LSF, you would submit that atm_averages wrapper script as follows:

>bsub < atm_averages