Supporting Change Sets

Adding support for change_sets (groups) to SEACAS applications

• exodus
  • Need to figure out caching in the exodus file_item struct. The values are not valid when switching to a new group which can be done without closing/opening a file...
    • Do we need a ex_open_group and ex_close_group which update the cached values...
  • Need to decide how to handle the group file_id vs root_id in file_item and check_valid_file type calls...
• io_info
  • change groups to change_sets
  • use new API in IOSS instead of going down to exodus level
  • show which change_set the info is being show for and number of change_sets in file
  • show info for multiple change sets
• io_shell
  • change groups to change_sets
  • If copying a file, output should have same change_sets as input
  • add option to copy only selected change_sets from input file to output file.
  • add option to extract all change_sets or list of change_sets to output file (Use %G in filename to substitute change_set name)
  • --compare option should compare all change_sets in the file by default
  • --compare option should have option to compare specified change_sets in the file
  • parallel io_shell can decompose a file with change sets (all change sets in decomposed file)
  • parallel io_shell can compose a file with change sets (all change sets in composed file)
  • parallel io_shell can decompose a file with change sets (specified change set in output file)
  • parallel io_shell can compose a file with change sets (specified change set in composed file)
• exodiff
  • if input and output have change_sets, do diff on each change_set
  • allow diffing a specific change_set in a file with a specific change_set in the other file
  • allow diffing a specific change_set in a file with a file that has no change_sets.
  • allow specifying one or more change sets (index or name) to compare with the same number of specified change sets in the other file.
  • summary mode works on files with change sets
  • how does the command file option work...
• epu
  • a set of file-per-rank files with change_sets should produce a single file with the same set of change_sets.
  • specify a change_set (1-based index) from the input file(s) to write to output file.
  • specify a change_set by name from the input file(s) to write to output file
• conjoin
  • "union" all change_sets in a file to a single file with status variables (treat the change_sets similar to how individual files are currently handled)
  • Handle multiple input files each containing 0 or more change sets...
• aprepro
  • create a variable containing the names of the change_sets in the file
  • create a variable containing the count of change_sets in the file
  • permit opening a specific change_set in an exodus file.