pythonic-config 0.5.3

No project description provided

Config

This package allows you to configure functions explicitly and safely. You will be able to create an intuitive type-checked configuration file that directly sets function arguments, globally.

This is a lightweight package with only two widely used dependencies and only a couple hundred line of code.

Contents

• Config - Contents
  • Install
  • Usage 🤗
    • Configuring a function
    • Writing a configuration file
    • Configuring via the command line
    • Logging the configuration
    • Advanced: Sharing configurations between processes
    • Advanced: Ensuring the configuration is used
  • How does this work?

Install

Make sure you have Python 3, then you can install pythonic-config using pip:

pip install pythonic-config

Install the latest code via:

pip install git+https://github.com/PetrochukM/Config.git

Usage 🤗

Configuring a function

Any function can be configured, and then used anywhere, see below:

import config

# Define function
def do_something_cool(how_many_times: int):
    pass

# Configure function
config.add({do_something_cool: config.Args(how_many_times=5)})

# Use the configured function anywhere! 🎉
do_something_cool(how_many_times=config.get())

This approach is simple but powerful. Now, each configuration can be directly attributed to a documented function argument.

Furthermore, config incorporates typeguard 💂‍♀️ so every configuration is type checked at runtime.

Writing a configuration file

The simple example above can be extended to create a configuration file, for example:

import config
import data
import train

config.add({
  data.get_data: config.Args(
      train_data_path="url_lists/all_train.txt",
      val_data_path="url_lists/all_val.txt"
  ),
  data.dataset_reader: config.Args(
      type_="cnn_dm",
      source_max_tokens=1022,
      target_max_tokens=54,
  ),
  train.make_model: config.Args(type_="bart"),
  train.Trainer.make_optimizer: config.Args(
      type_="huggingface_adamw",
      lr=3e-5,
      correct_bias=True
  )
  train.Trainer.__init__: config.Args(
      num_epochs=3,
      learning_rate_scheduler="polynomial_decay",
      grad_norm=1.0,
  )
})

With this approach, this configuration file will make it clear which (hyper)parameters are set and where. This improves overall readability of the configuration file.

🐍 Last but not least, the configuration file is written in Python, you can use variables, lambdas, etc to further modularize.

Configuring via the command line

In case you want to change one variable at a time, this package supports configuration from the command line, for example:

python example.py --sorted='Args(reverse=True)'

import sys
import config

config.add(config.parse_cli_args(sys.argv[1:]))

Logging the configuration

Lastly, it's useful to track the configuration file by logging it. This package supports that via config.log. In the example below, we log the configuration to Comet.

from comet_ml import Experiment
import config

experiment = Experiment()
experiment.log_parameters(config.log())

Advanced: Sharing configurations between processes

In multiprocessing, it may be useful to share the configuration file between processes. In this case, the configuration can be exported to another process and then subsequently imported, see below:

from multiprocessing import Process
import config

def handler(configs: config.Config):
    config.add(configs)

if __name__ == "__main__":
    process = Process(target=handler, args=(config.export(),))
    process.start()
    process.join()

Advanced: Ensuring the configuration is used

In a large code base, it might be hard to tell if the configuration has been set for every function call. In this case, we've exposed config.trace which can double check every function call against the configuration, see below:

import sys
import config

def configured(a=111):
    pass

sys.settrace(config.trace)
config.add({configured: config.Args(a=1)})

configured()  # `config.trace` issues a WARNING!
configured(a=config.get())

We also have another option for faster tracing with enable_fast_trace. Instead of a system wide trace, this will inject new code into the __code__ of every function in your configuration for tracing. This has much lower overhead; however, it is still in beta due to the number of edge cases.

How does this work?

Our approach is simple, config maintains a global configuration mapping each function to its associated arguments. When a user calls config.get, it'll attempt to parse the code to determine the calling function and associated argument. In most cases, this will be successful and it will fetch the appropriate value. If it's not successful, it'll raise an error that'll help you fix the issue.