morphit 1.2.0

makes shitty data more fit (a smarter serializer)

MorPhit

Usage

    >>> from morphit import Parser
    >>> Parser([], '[1,2,3]')
    [1, 2, 3]

pip install morphit

sudo -H pip install morphit

morphit is a no nonsense parser that takes a template for data and converts whatever you feed it into whatever you want it to be.

Do you have "My data is fucked"? If you find yourself saying any of the following, you may to be entitled to financial compensation.

• WTF half of these are strings instead of insert type here
• How do I even convert from insert type A to insert type B
• I have so much bad data I'm going to have to manually fix
• I have hundreds of if conditions to fix weird data
• How the hell am I supposed to fix the type of all these nested objects
• You're telling me this float is supposed to be a date?...
• You're telling me that I cant put a datetime object anywhere?
• I don't know how any of these strings became numbers but...

Features

• Processor chaining

>>> Processor.flow(["<parser|lambda|func|obj>",...])

>>> p = Processor(dict)
>>> p.then(lambda d: d['date'])
     .then(Processor.types['Date'])
>>> res = p('{"date": "2018-01-31T06:17:45.547"}')

datetime(2018, 1, 31, 6, 17, 45, 547000)

• datetime serialization

>>> from datetime import date
>>> Parser(str, datetime(2018, 1, 31, 6, 17, 45, 547000))
"2018-01-31T06:17:45.547"

• deep serialization of dicts
• supports custom serializers using methods/lambda functions
• templated parsers
• nested templated parsers

Supported Conversions

• any -> list
• any -> object
• any -> tuple
• any -> type
• any -> str
• dict -> str[json]
• dict -> dict
• dict -> list
• dict -> tuple
• datetime -> str
• datetime -> str[json]
• datetime -> int
• datetime -> float
• datetime.time -> str
• datetime.date -> str
• float -> str
• float -> float
• float -> list
• float -> tuple
• float -> type
• float -> datetime
• int -> str
• int -> int
• int -> list
• int -> tuple
• int -> datetime
• int -> type
• list -> str[json]
• list -> list
• list -> tuple
• tuple -> str[json[list]]
• tuple -> list
• tuple -> tuple
• none -> str
• none -> bool
• none -> list
• none -> tuple
• str -> bool
• str -> datetime
• str -> dict
• str -> list
• str -> tuple
• str[int] -> float
• str[int] -> int
• str[float] -> float
• str[float] -> int
• str -> type
• str[iso8601] -> datetime
• str[iso8601] -> float
• str[unicode] -> any string conversions
• str[python] -> any string conversions

iso8601 a date string formatted to the iso8601 spec json a json encoded equivilent of the data unicode strings that get messed up with u' ex: "[u'this',u'that']" python strings that are weirdly single quoted ex: "['photo', 2, 'pic', 'pics']"

Installation

pip3 install morphit

sudo -H pip3 install morphit

Deploy

python3 setup.py build       #Build:
python setup.py coverage         #Test
python3 setup.py install     #Install:

Deploy

• Run tests as outlined above
• Make sure you bump the version
• Check venv is active

echo "__version__=1.x.y" > ./morphit/version.py # where x and y are the next version
python3 setup.py coverage && python3 setup.py sdist bdist_wheel
twine upload dist/*

Development

virtualenv --python=python3 venv
source ./venv/bin/activate
python setup.py easy_install
pip install -r requirements-dev.txt

Roadmap

• Custom exceptions for bad casting (int/list/tuple -> dict)
• Add strict parameter that enables enforcement of keys and list/tuple length
• Add default parameter that remplaces None or non existing values with a specified default
• Add fallback parameter that accepts a value or a lambda function to replace the value

CHANGELOG

1.2.0

Added custom aggregators for processors to allow result merging without needing to update the chained partial object.

• FIX: Coverage was broken due to the import of version in setup.py
• FEATURE: Custom aggregators for Processor class

1.1.0

Breaking changes: removed Template class, types are now attached to Processor

• FIX: 100% test coverage, all dead code eliminated
• FIX: primitive to iterable casting
• FEATURE: Removed & Replaced Template with Processor callable class
• FEATURE: Added processor chaining then flow
• FEATURE: Added lambda Processor original data & result parameters

BUGS

Contributing

Read CONTRIBUTING_.

Make a PR and I'll give it a look

Odds are good / the goods are odd