Databases

ITM data entries are defined by the following information : (user, machine, shot, run). "user" is either "public" (public ITM database) or any Gateway username, allowing the creation of private ITM databases.

In order to run a KEPLER workflow, you must first create your private database(s). A private database is created for a given machine (tokamak name) and data structure version, by the following command:

$ITMSCRIPTDIR/create_user_itm_dir TokamakName \
 DataVersion

Example (creates a tree for tokamak name test, allowed for testing purposes):

$ITMSCRIPTDIR/create_user_itm_dir test 4.08a

The database is created under ~my_username/public/itmdb/itm_trees. Since it is located in your public directory, all Gateway users can read from it.

Standard tokamak names are : asdex_upgrade, jet, mast, tore_supra, ... Nonetheless, any arbitrary machine name is allowed by the system for testing purposes (e.g. test). It is however strongly recommended to use the standard machine names when using real experimental data.

Before using KEPLER or the UAL , you must specify on which database you wish to work, using the following command :

source $ITMSCRIPTDIR/set_itm_data_env public|user TokamakName DataVersion

Example (set the environment variables to work to the previously created directory):

source $ITMSCRIPTDIR/set_itm_data_env my_username test 4.08a

Note that with this command, you can read any ITM database on the Gateway, including the public database (replace my_username by "public").

The public database is located at /pfs/itmdb/itm_trees/public/. Its content is summarised in ITM Public Database, which contains also the location of the various files, the list of standard machine names ...

For the moment, the ITM software (KEPLER, UAL) are able to work only on one database at a time. Reading is allowed from all databases (by using the set_itm_data_env command above). However, if you wish to write also some results during the workflow, you can write only to your private database. Therefore, if you wish to use data entries from other users in your workflow, you need first to copy them in your own directory.

For example, to copy the test example from the public database (user=public, machine=test, shot=1, run=1) to your directory, type:

cp /pfs/itmdb/itm_trees/public/test/4.08a/mdsplus/0/euitm_10001.* \
   ~my_username/public/itmdb/itm_trees/test/4.08a/mdsplus/0/.

This data entry can then be used as (user=my_username, machine=test, shot=1, run=1). It is not recommended to change the shot or run number when copying data entries like this. More flexible tools for working on multiple databases simultaneously will be provided in the near future.

Machine descriptions are stored in shot 0 of each tokamak. Different versions (valid e.g. for different shot ranges) can be stored, using different run numbers. To copy the Tore Supra machine description (run 1) to your local folder, type (NB you must first create your private tore_supra 4.08a database):

cp /pfs/itmdb/itm_trees/public/tore_supra/4.08a/mdsplus/0/euitm_001.* \
   ~my_username/public/itmdb/itm_trees/tore_supra/4.08a/mdsplus/0/.

To copy the Tore Supra experimental data set for shot 39736, run 1, type:

cp /pfs/itmdb/itm_trees/public/tore_supra/4.08a/mdsplus/0/euitm_397360001.* \
   ~my_username/public/itmdb/itm_trees/tore_supra/4.08a/mdsplus/0/.

Note for advanced users : an ITM data entry consists in three file, euitm_SSSSSSRRRR.* where SSSSSS is the shot number (truncated, e.g. shot=1 is SSSSSS=1) and RRRR the run number (exception : if shot = 0, SSSSSSRRRR = (R)RRR, e.g. shot 0 run 1 is euitm_001.*). In the example above, it is assumed that one is using the default storage method MDS+ (otherwise, for HDF5, replace in the path the "mdsplus" folder name by "hdf5"), and that the run number is below 9999 (otherwise, replace the "0" folder name by "1" for run numbers between 10000 and 19999, ...).

---

last update: 2013-09-12 by dpc