Decide on naming of simulation files

[gilrrei]

This issue was originally created by @gilrrei on gitlab.lrz.de on 2024-10-25.

Motivation and Context

During the merge request 754 the question regarding naming of simulation data came up. As discussed in developer_meetings74 here are the viable options:

Current approach:

/home/rei/Desktop/queens-simulation-data/experiments/smc_battery_manage_paths_current
  ├── 0
  │   ├── metadata.yaml
  │   ├── smc_battery_geometry_file_0.exo
  │   ├── jobscript.sh
  │   ├── smc_battery_input_file_0.dat
  │   └── output
  │       ├── smc_battery_0.vtu
  │       ├── smc_battery_0.log
  │       └── smc_battery_0.err
  └── 1
      ├── smc_battery_input_file_1.dat
      ├── metadata.yaml
      ├── smc_battery_geometry_file_1.exo
      ├── jobscript.sh
      └── output
          ├── smc_battery_1.vtu
          ├── smc_battery_1.log
          └── smc_battery_1.err

Advantages:

• redundant information: the job can be identified by the filenames

Disadvantage:

• redundant information: job_id and experiment_name needs to be provided for every script you create. Also for the data processor regex expression (i.e. smc_battery_*.vtu) and checks if the file is unique are required

Without the job_id

/home/rei/Desktop/queens-simulation-data/experiments/smc_battery_manage_paths_no_job_id
  ├── 0
  │   ├── metadata.yaml
  │   ├── smc_battery_input_file.dat
  │   ├── jobscript.sh
  │   ├── smc_battery_geometry_file.exo
  │   └── output
  │       ├── smc_battery.log
  │       ├── smc_battery.vtu
  │       └── smc_battery.err
  └── 1
      ├── metadata.yaml
      ├── smc_battery_input_file.dat
      ├── jobscript.sh
      ├── smc_battery_geometry_file.exo
      └── output
          ├── smc_battery.log
          ├── smc_battery.vtu
          └── smc_battery.err

Advantages:

• no redundant information: no need to add the job_id when looping or looking into files. For the data processor simply use: smc_battery.vtu

Disadvantage:

• redundant information: experiment_name needs to be provided for every script you create. Identifying the job from a file on its own is hard

Nothing:

/home/rei/Desktop/queens-simulation-data/experiments/smc_battery_manage_paths_nothing
  ├── 0
  │   ├── input_file.dat
  │   ├── metadata.yaml
  │   ├── jobscript.sh
  │   ├── geometry_file.exo
  │   └── output
  │       ├── output.err
  │       ├── output.log
  │       └── output.vtu
  └── 1
      ├── input_file.dat
      ├── metadata.yaml
      ├── jobscript.sh
      ├── geometry_file.exo
      └── output
          ├── output.err
          ├── output.log
          └── output.vtu

Advantages:

• no redundant information at all: no need to add the job_id or experiment_name when loop or looking into files. For the data processor simply use: output.vtu

Disadvantage:

• no redundant information: Identifying the job and experiment from a file on its own is hard

Related Issues

• Blocks
• Is blocked by
• Follows 754
• Precedes
• Related to
• Part of
• Composed of

Additional Information

Interested Parties

@queens_community/QUEENS_maintainers