pycon22-jsonargparse

Examples used in the talk "jsonargparse - Say goodbye to configuration hassles" at PyConDE / PyData Berlin 2022.

Links

• Slides: https://speakerdeck.com/stecklin/jsonargparse-say-goodbye-to-configuration-hassles
• Recording: to be added soon
• jsonargparse: https://github.com/omni-us/jsonargparse

1st Live Demo

• same usage as argparse, additional information in the help (types, default values, docstring of function)
  • Compare python 0_argparse.py --help and python 1_cli.py --help
  • Run python 1_cli.py PyCON --year 2022
• possible to parse config files, pattern "Print config, edit config, run config"
  • Print config using python 1_cli.py --print_config > config.yaml
  • Edit config.yaml
  • Run python 1_cli.py --config config.yaml
• order of command line arguments decides if year is picked from config or from command line argument
  • Run python 1_cli.py --config config.yaml --year 2024
  • Run python 1_cli.py --year 2024 --config config.yaml
• possible to parse from environment variables
  • Run PYCON_CONFERENCE_NAME=PyCON PYCON_YEAR=2024 python 2_cli_with_env_variables.py

2nd Live Demo

• help shows only the top level configuration options, we can then look deeper into single groups
  • Run python 3_image_classification.py --help
  • Run python 3_image_classification.py --dataset.help Cifar10Dataset
• type checks: useful error messages when invalid values are provided, access permissions are checked for paths
  • Run python 3_image_classification.py --config configs/3_config.yaml --dataset.augment maybe
  • Run python 3_image_classification.py --config configs/3_config.yaml --dataset.data_dir /home/ghosts/

3rd Live Demo

• select classes for --print_config, list of classes is displayed in help
  • Run python 4_subclasses.py --help
  • Run python 4_subclasses.py --dataset ImageNetDataset --model ConvNet --print_config
• possible to split config into subconfigs
  • Run python 4_subclasses.py --config configs/4_config.yaml
  • Run python 4_subclasses.py --dataset configs/4_cifar10.yaml --model configs/4_conv_net.yaml
  • Run python 4_subclasses.py --config configs/4_config_with_subconfigs.yaml
• arbitrary levels of nesting supported
  • Run python 5_model_composition.py --config configs/5_config.yaml