Features

• Map fields using simple dot-separated string paths for nested access (literal keys, no quoting needed)

• Transform field values in-place with 25+ built-in conversion actions (string ops, math, URL cleanup, phone parsing, etc.)

• Compute output values with command processors (GeoJSON points, joins, splits, list transforms)

• Convert between date formats (Unix timestamps, DD-MM-YYYY, YYYY_MM_DD, etc.)

• Run records through Pydantic models or dataclasses with optional method invocation

• Conditionally execute rules or skip entire records

• Extend with custom field convertors, date formatters, and command classes via protocols

Installation

pip install record-convertor

Quick Start

from record_convertor import RecordConvertorWithRulesDict

rules = {
    "name": "item.name",
    "brand": "item.brand.name",
    "price": "item.price",
}

convertor = RecordConvertorWithRulesDict(rule_dict=rules)

result = convertor.convert({
    "item": {
        "name": "Widget",
        "brand": {"name": "Acme"},
        "price": 9.99,
    }
})
# {"name": "Widget", "brand": "Acme", "price": 9.99}

Documentation

Full documentation is available at https://record-convertor.readthedocs.io/

License

LGPL-3.0-or-later