6. Format Conversion

Frequently, specific configurations are saves as files in various formats such as JSON, YAML, TOML, or Python scripts. The pydantic_sweep.convert module provides utilities to facilitate conversion between these formats while maintaining the integrity of the Pydantic models. Concertely, the module supports

• Loading models from JSON, YAML, TOML, and Python files

• Writing models to JSON, YAML, TOML, and Python files

• Code generation to convert models into executable Python code

• Command-line interface for batch conversions

6.1. Conversion between formats

The convert.load function and convert.write functions allow reading and writing Pydantic models from and to various file formats. For example, the following saves a model as yaml and then reads it back. The model argument to convert.load is the dot-separated import path to the Pydantic model class (in this case __main__.Config).

import tempfile
from pathlib import Path

from pydantic_sweep import BaseModel, convert


class Config(BaseModel):
    name: str
    value: int
    path: Path


config = Config(name="test", value=42, path=Path("/tmp/data"))

with tempfile.TemporaryDirectory() as temp_dir:
    config_file = Path(temp_dir) / "config.yaml"
    convert.write(config_file, model=config)
    loaded_config = convert.load(config_file, model="__main__.Config")

assert loaded_config == config
print(loaded_config)

name='test' value=42 path=PosixPath('/tmp/data')

6.1.1. Python Code

Storing configuration directly in Python files can be advantageous for complex setups, as they allow for dynamic generation of configurations using code. They are supported the same was as other formats, with the difference that the model parameter to convert.load is the variable name inside the Python file that holds the model instance.

with tempfile.TemporaryDirectory() as temp_dir:
    config_file = Path(temp_dir) / "config.py"
    convert.write(config_file, model=config)

    print("Written Python file:")
    with open(config_file) as f:
        print(f.read())

    loaded_config = convert.load(config_file, model="model")

assert loaded_config == config

Written Python file:
from __main__ import Config


model = Config(
    name='test',
    value=42,
    path='/tmp/data'
)

6.2. Command-Line Interface

The conversion module also provides a command-line interface (CLI) for converting configuration files between different formats. This is particularly useful for batch processing or integrating into existing workflows. To use the CLI, you can run the following command in your terminal:

python -m pydantic_sweep.convert input_file output_file --model model_path/name

Here, input_file is the path to the source configuration file, output_file is the path where the converted file should be saved, and model_path is the dot-separated import path to the Pydantic model class (or variable name for Python files). For example, to convert a JSON configuration file to YAML, you would run:

python -m pydantic_sweep.convert config.json config.yaml --model my_library.MyModel