Data Types

These are the data types mimicking the NumPy’s scalar hierarchy of types and allowing to describe NumPy’s dtypes in the Nani format.

They can later on be converted into NumPy’s dtypes by calling the resolve() function.

Bool	Type corresponding to numpy.bool_.
Object	Type corresponding to numpy.object_.
Number	Type corresponding to numpy.number.
String	Type corresponding to numpy.string_.
Unicode	Type corresponding to numpy.unicode_.
Array	Type corresponding to a NumPy (sub)array.
Structure	Type corresponding to a NumPy structured array.
Bytes	alias of String
Str	alias of Unicode

---

class nani.Bool(default=False, view=None)

Type corresponding to numpy.bool_.

default

bool – Default value.

view

type or None – If None, the owning array returns a direct reference to this boolean value, otherwise it is expected to be a class object wrapping it and accepting 2 parameters: data, the NumPy array owning the boolean value, and index, its position in the array.

---

class nani.Object(default=None, view=None)

Type corresponding to numpy.object_.

default

object – Default value.

view

type or None – If None, the owning array returns a direct reference to this Python object, otherwise it is expected to be a class object wrapping it and accepting 2 parameters: data, the NumPy array owning the Python object, and index, its position in the array.

---

class nani.Number(type=numpy.float_, default=0, view=None)

Type corresponding to numpy.number.

type

type – Type of the number. Either one inheriting from numbers.Number or numpy.number.

default

numbers.Number or numpy.number – Default value.

view

type or None – If None, the owning array returns a direct reference to this numeric value, otherwise it is expected to be a class object wrapping it and accepting 2 parameters: data, the NumPy array owning the numeric value, and index, its position in the array.

---

class nani.String(length, default='', view=None)

Type corresponding to numpy.string_.

length

int – Number of characters.

default

str on PY2 or bytes on PY3 – Default value.

view

type or None – If None, the owning array returns a direct reference to this string value, otherwise it is expected to be a class object wrapping it and accepting 2 parameters: data, the NumPy array owning the string value, and index, its position in the array.

---

class nani.Unicode(length, default='', view=None)

Type corresponding to numpy.unicode_.

length

int – Number of characters.

default

unicode on PY2 or str on PY3 – Default value.

view

type or None – If None, the owning array returns a direct reference to this unicode value, otherwise it is expected to be a class object wrapping it and accepting 2 parameters: data, the NumPy array owning the unicode value, and index, its position in the array.

---

class nani.Array(element_type, shape, name=None, view=None)

Type corresponding to a NumPy (sub)array.

element_type

nani type – Type of each element.

shape

int or tuple of int – Shape of the array. Passing an int defines a 1D array.

name

str or None – Name for the view type if view is None.

view

type or None – If None, a view for this array is dynamically generated by Nani, otherwise it is expected to be a class object wrapping it and accepting 1 parameter: data, the corresponding NumPy array.

---

class nani.Structure(fields, name=None, view=None)

Type corresponding to a NumPy structured array.

fields

tuple of nani.Field or compatible tuple – Fields defining the structure.

name

str or None – Name for the view type if view is None.

view

type or None – If None, a view for this structured array is dynamically generated by Nani, otherwise it is expected to be a class object wrapping it and accepting 1 parameter: data, the corresponding NumPy structured array.

---

class nani.Bytes(length, default='', view=None)

Alias for String.

---

class nani.Str(length, default='', view=None)

Alias for String on PY2 or Unicode on PY3.