Skip to content

README.md

Latest commit

 

History

History
71 lines (49 loc) · 1.61 KB

README.md

File metadata and controls

71 lines (49 loc) · 1.61 KB

argdump

Serialize and deserialize Python argparse parsers.

Installation

pip install argdump

Usage

import argparse
import argdump

# Create a parser
parser = argparse.ArgumentParser(prog="mytool")
parser.add_argument("input")
parser.add_argument("-v", "--verbose", action="count", default=0)
parser.add_argument("--format", choices=["json", "csv"])

# Serialize
data = argdump.dump(parser)       # dict
json_str = argdump.dumps(parser)  # JSON string

# Deserialize
restored = argdump.load(data)
restored = argdump.loads(json_str)

# Use normally
args = restored.parse_args(["input.txt", "-vvv", "--format", "json"])

JSON Schema

A JSON Schema for validating serialized output is available at docs/schema-v1.json.

Features

  • All standard actions (store, append, count, etc.)
  • Subparsers with aliases
  • Mutual exclusion and argument groups
  • Type converters (builtins, FileType, importable functions)
  • Choices, defaults, metavar, help text
  • Environment metadata ($env) for reproducibility

Options

# Exclude environment metadata
argdump.dump(parser, include_env=False)

# Non-strict mode: skip unresolvable types instead of raising
argdump.load(data, strict=False)

Limitations

Lambdas and closures cannot be serialized. Use strict=False to skip them:

parser.add_argument("--value", type=lambda x: int(x) * 2)

argdump.load(argdump.dump(parser), strict=False)  # type becomes None
argdump.load(argdump.dump(parser), strict=True)   # raises UnresolvableTypeError

License

MIT