data-mapper

A declarative data mapper

Description

Most projects work with different representations of the same data. The code that is written every time to morph data between its representations is mostly very repetitive. More over, it is always a deposit of bugs and issues which requires a developer or tester to unit-test it.

This package is an attempt to solve these problems... well, at least the most common ones.

And to make a developer job easier it is primarily designed to be used in declarative fashion: describe what you want and get it right after.

Use Cases and Features

Here are examples of the most common use-cases and features:

Different Naming Schemes

This mapper looks for properties first_name and last_name in the data. For property first_name it tries to resolve it by the first key 'first_name', if not found it tries the second key 'name'. The similar process goes for property last_name.

from data_mapper.mappers import Mapper
from data_mapper.properties import Property

class PersonMapper(Mapper):
    first_name = Property('first_name', 'name')
    last_name = Property('last_name', 'surname')

mapper = PersonMapper()

assert mapper.get({
    'first_name': 'Ivan', 
    'surname': 'Bogush',
}) == {
    'first_name': 'Ivan', 
    'last_name': 'Bogush',
}

assert mapper.get({
    'name': 'Ivan', 
    'surname': 'Bogush',
}) == {
    'first_name': 'Ivan', 
    'last_name': 'Bogush',
}

This use-case has a story :)

It was the first issue I wanted to solve in my other project. I had different naming schemes in different data sources, and in my databases. All of them used different names for product categories: 'categories', 'category', 'categoryId'. I found it very boring to write repeatable code to convert the same data.

Arbitrary functions on resolved values

Full Name String Construction

This one resolves properties first_name, middle_name [optionally] and last_name and combines them into a single string — full_name.

from data_mapper.shortcuts import F, Str, L

full_name = F(
    ' '.join,
    L(
        Str('first_name'),
        Str('middle_name', required=False),
        Str('last_name'),
        skip_none=True,
    ),
)

assert 'Anton Pavlovich Chekhov' == full_name.get(dict(
    first_name='Anton',
    middle_name='Pavlovich',
    last_name='Chekhov',
))

assert 'Anton Chekhov' == full_name.get(dict(
    first_name='Anton',
    last_name='Chekhov',
))

Object mapping

Dict to Object

Let's assume we have a class Person:

class Person:
    def __init__(
            self,
            id_: int,
            first_name: str,
            last_name: str,
            middle_name: str = None,
    ):
        self.id = id_
        self.first_name = first_name
        self.last_name = last_name
        self.middle_name = middle_name

A mapper from dict with corresponding keys to an instance of class Person could be defined by subclassing ObjectMapper:

from data_mapper.mappers.object import ObjectMapper
from data_mapper.properties import (
    CompoundProperty, CompoundListProperty, IntegerProperty, StringProperty,
)


class PersonMapper(ObjectMapper):
    init = Person
    args = CompoundListProperty(
        IntegerProperty('id'),
        StringProperty('first_name'),
        StringProperty('last_name'),
    )
    kwargs = CompoundProperty(
        middle_name=StringProperty(required=False),
    )

first, middle, last = 'Iosif Aleksandrovich Brodsky'.split()
person = PersonMapper().get(dict(
    id=1940,
    first_name=first,
    middle_name=middle,
    last_name=last,
))

assert isinstance(person, Person)
assert person.id == 1940
assert person.first_name == first
assert person.middle_name == middle
assert person.last_name == last

Exactly the same can be done by instantiating the ObjectMapper:

from data_mapper.mappers.object import ObjectMapper
from data_mapper.properties import (
    CompoundProperty, CompoundListProperty, IntegerProperty, StringProperty,
)


mapper = ObjectMapper(
    init=Person,
    args=CompoundListProperty(
        IntegerProperty('id'),
        StringProperty('first_name'),
        StringProperty('last_name'),
    ),
    kwargs=CompoundProperty(
        middle_name=StringProperty(required=False),
    ),
)

first, middle, last = 'Iosif Aleksandrovich Brodsky'.split()
person = mapper.get(dict(
    id=1940,
    first_name=first,
    middle_name=middle,
    last_name=last,
))

assert isinstance(person, Person)
assert person.id == 1940
assert person.first_name == first
assert person.middle_name == middle
assert person.last_name == last