BnpDataclass

BioNumPy extends the concept of Python’s dataclass decorator, by providing a bnpdataclass-decorator. This decorator enables dataclass-like objects with numpy-array-like fields (EncodedArray, EncodedRaggedArray, RaggedArray or pure numpy-arrays). See examples of usage in the documentation below.

API documentation

bionumpy.bnpdataclass.bnpdataclass(base_class: type) → Type[BNPDataClass][source]

Create a bnpdataclass from a class with fields specified

A wrapper around @npdataclass that includes implicit format conversion for strings and other encoded data. @npdataclass is again a wrapper around dataclasses.dataclass but where all the fields are assumed to be objects that supports advanced numpy indexing so that any dataclass objects are also indexible like a numpy array.

bnpdataclass classes are meant to be dataclasses and so should not have any methods other than those implicitly given by npdataclass.

Parameters

base_classtype: Base class that defines the fields of the dataclass.

Returns

bnpdataclass: bnpdataclass object that supports numpy like indexing

Examples

>>> from bionumpy.bnpdataclass import bnpdataclass
>>> @bnpdataclass
... class Person:
...       name: str
...       age: int
... 
>>> data = Person(["Knut", "Ivar", "Geir"], [35, 30, 40])
>>> print(data)
Person with 3 entries
                     name                      age
                     Knut                       35
                     Ivar                       30
                     Geir                       40

>>> print(data[[0,2]])
Person with 2 entries
                     name                      age
                     Knut                       35
                     Geir                       40

>>> print(data[[False,True, False]])
Person with 1 entries
                     name                      age
                     Ivar                       30

bionumpy.bnpdataclass.replace(obj: BNPDataClass, **kwargs) → BNPDataClass[source]

Replace the values of a dataclass with new values

Parameters

objBNPDataClass: The dataclass to be replaced
kwargsdict: The new values to be replaced

Returns

BNPDataClass: The new dataclass with the replaced values

Examples

>>> import bionumpy as bnp
>>> entry = bnp.SequenceEntry(['seq1'], ['acgt'])
>>> entry
SequenceEntry with 1 entries
                     name                 sequence
                     seq1                     acgt
>>> bnp.replace(entry, name=['seq2'])
SequenceEntry with 1 entries
                     name                 sequence
                     seq2                     acgt

bionumpy.bnpdataclass.make_dataclass(fields: List[Tuple], name: str = 'DynamicDC', bases=()) → Type[BNPDataClass][source]

Constructs a dynamic dataclass from a list of attributes

Parameters

fields: list: a list of tuples in format (field_name, field_type) to be used to construct the dynamic bnp dataclass
name: str: optional name of new class

Returns

new BNPDataClass