mangadap.util.datatable module

Light-weight class for handling data tables using numpy record arrays.

class mangadap.util.datatable.DataTable(keys, types, element_shapes=None, descr=None, shape=None)[source]

Bases: object

A light-weight class that wraps a numpy record array.

Parameters:

keys (str, list) – Table column keys. Must be a single string or a list of strings.
types (type, list) – Column data type. Must be a single type or a list of types. Supported types are limited to integers and floating point values. If a single type and more than one key is provided, all columns will have the same type.
element_shapes (tuple, list, optional) – Shape of each column element. If None, all elements are assumed to be single values. If a list is provided, the elements of the list must be either None (for a single value) or a tuple with the shape of each element. If a single shape and more than one key is provided, all columns will have the same element shape.
descr (str, list, optional) – Human-readable description of each column. If None, no descriptions are available. If a list, each element must be a string or None. If not None, must provide a description or a None element for each key.
shape (int, tuple) – Shape of the data table to instantiate. If None, data will be None and instantiation of the array will require a call to init().

Raises:

TypeError – Raised if any of the provided arguments or list item have an incorrect type.
ValueError – Raised if any of the provided arguments do not provide the same number of columns as determined by the column keys.

keys

See argument list.

Type:: list

types

See argument list.

Type:: list

element_shapes

See argument list.

Type:: list

descr

See argument list.

Type:: list

data

Array with the table data.

Type:: numpy.recarray

_dtype

List of tuples with the data type of the record array.

Type:: list

append(rhs)[source]

Append the data in the provided datatable to this one.

Both objects (self and rhs) must have the same derived class. This instance is directly modified.

Parameters:

rhs (DataTable) –

The instance with the data to append. The derived class of this object must be the same as the object to which you want to append the data. I.e.:

assert isinstance(rhs, self.__class__)

property dtype

Return the data type of the record array.

If data has been initialized, this is identically:

self.data.dtype

Returns:: Table data type.
Return type:: numpy.dtype

init(shape)[source]

(Re)Initialize data.

Note that this always re-initializes the array, regardless of whether or not there was a pre-existing data array.

Parameters:: shape (tuple) – The shape of the array to construct.

property shape

Return the shape of the data array.

Returns:: Returns the shape of the data array. If the data has not been instantiated, returns None.
Return type:: tuple

property size

Return the size of the data array.

Returns:: Returns the number of elements in the array. If the data has not been instantiated, returns 0.
Return type:: int

to_hdu(name=None, add_dim=False)[source]

Use the data table to build an astropy.io.fits.BinTableHDU.

Todo

add_dim shouldn’t be optional; the code should just do this always or know when it needs to.

Parameters:

name (str, optional) – HDU extension name.
add_dim (bool, optional) – Include the dimensionality of the column in the construction of the astropy.io.fits.Column.

Returns:

FITS binary table with the data.

Return type:

astropy.io.fits.BinTableHDU

to_rst_table(header=True, class_link=True)[source]

Construct a reStructuredText table describing the data table.

Parameters:

header (bool, optional) – Include a section header
class_link (bool, optional) – Include an rst-style link to the class instantiation documentation.

Returns:

Returns a list of lines that can be written to an *.rst file.

Return type:

list