landlab.utils.jaggedarray_ma¶
Store arrays of variable-length arrays implemented with masked arrays.
Implements a MaskedJaggedArray class using numpy masked arrays.
Examples
Create a MaskedJaggedArray that stores link IDs for the links attached to the nodes of a 3x3 grid.
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray
>>> links_at_node = MaskedJaggedArray(
... [
... [0, 6],
... [1, 7, 0],
... [8, 1],
... [2, 9, 6],
... [3, 10, 2, 7],
... [11, 3, 8],
... [4, 7],
... [5, 10, 4],
... [5, 11],
... ]
... )
Make up some data that provides values at each of the links.
>>> value_at_link = np.arange(12, dtype=float)
Create another MaskedJaggedArray. Here we store the values at each of the links attached to nodes of the grid.
>>> values_at_node = MaskedJaggedArray.empty_like(links_at_node, dtype=float)
>>> values_at_node.array = value_at_link[links_at_node.array]
Now operate on the link values for each node.
>>> values_at_node.foreach_row(np.sum)
array([ 6., 8., 9., 17., 22., 22., 11., 19., 16.])
>>> values_at_node.foreach_row(np.min)
array([0., 0., 1., 2., 2., 3., 4., 4., 5.])
>>> values_at_node.foreach_row(np.max)
array([ 6., 7., 8., 9., 10., 11., 7., 10., 11.])
Note: np.ptp doesn’t work on masked arrays but since max and min seem to, you can create a new ptp that returns the expected values.
>>> ptp = lambda x, axis=None: np.max(x, axis=axis) - np.min(x, axis=axis)
>>> values_at_node.foreach_row(ptp)
array([6., 7., 7., 7., 8., 8., 3., 6., 6.])
Or access the underlying masked array directly,
>>> values_at_node.masked_array.ptp(axis=1).data
array([6., 7., 7., 7., 8., 8., 3., 6., 6.])
- class MaskedJaggedArray[source]¶
Bases:
objectA container for an array of variable-length arrays.
MaskedJaggedArray([row0, row1, …]) MaskedJaggedArray(values, values_per_row)
Examples
Create a MaskedJaggedArray with an array of arrays.
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.array array([0, 1, 2, 3, 4])
Create a MaskedJaggedArray as a 1D array and a list or row lengths.
>>> x = MaskedJaggedArray([0, 1, 2, 3, 4], (3, 2)) >>> x.array array([0, 1, 2, 3, 4])
MaskedJaggedArray([row0, row1, …]) MaskedJaggedArray(values, values_per_row)
Examples
Create a MaskedJaggedArray with an array of arrays.
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.array array([0, 1, 2, 3, 4])
Create a MaskedJaggedArray as a 1D array and a list or row lengths.
>>> x = MaskedJaggedArray([0, 1, 2, 3, 4], (3, 2)) >>> x.array array([0, 1, 2, 3, 4])
- __init__(*args)[source]¶
MaskedJaggedArray([row0, row1, …]) MaskedJaggedArray(values, values_per_row)
Examples
Create a MaskedJaggedArray with an array of arrays.
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.array array([0, 1, 2, 3, 4])
Create a MaskedJaggedArray as a 1D array and a list or row lengths.
>>> x = MaskedJaggedArray([0, 1, 2, 3, 4], (3, 2)) >>> x.array array([0, 1, 2, 3, 4])
- __iter__()[source]¶
Iterate over the rows of the array.
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> for row in x: ... row ... array([0, 1, 2]) array([3, 4])
- __new__(**kwargs)¶
- property array¶
The jagged array as a 1D array.
- Returns:
A view of the underlying 1D array.
- Return type:
array
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.array array([0, 1, 2, 3, 4])
>>> x.array = np.array([1, 1, 2, 3, 4]) >>> x.array array([1, 1, 2, 3, 4])
- static empty_like(jagged, dtype=None)[source]¶
Create a new MaskedJaggedArray that is like another one.
- Parameters:
jagged (MaskedJaggedArray) – A MaskedJaggedArray to copy.
dtype (np.dtype) – The data type of the new MaskedJaggedArray.
- Returns:
A new MaskedJaggedArray.
- Return type:
- foreach_row(func, out=None)[source]¶
Apply an operator row-by-row.
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.foreach_row(np.sum) array([3, 7])
>>> out = np.empty(2, dtype=int) >>> x.foreach_row(np.sum, out=out) is out True >>> out array([3, 7])
- length_of_row(row)[source]¶
Number of values in a given row.
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.length_of_row(0) 3 >>> x.length_of_row(1) 2
- static ma_from_flat_array(array, values_per_row)[source]¶
Create a masked array from a flat array.
- Parameters:
array (array_like) – Values of the jagged array.
values_per_row (array_like of int) – Number of values in each row of the jagged array.
- Returns:
A new masked array.
- Return type:
np.masked_array
- static ma_from_list_of_lists(rows, dtype=None)[source]¶
Create a masked array from a list of lists.
- Parameters:
rows (array_like or array_like) – Rows of the jagged array.
dtype (np.dtype, optional) – The data type of the new masked array.
- Returns:
A new masked array.
- Return type:
np.masked_array
- property masked_array¶
The jagged array as a masked array.
- Returns:
The underlying masked array.
- Return type:
np.masked_array
- property number_of_rows¶
Number of array rows.
- Returns:
Number of rows in the array.
- Return type:
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.number_of_rows == 2 True
- row(row)[source]¶
Get the values of a row as an array.
- Parameters:
row (int) – Index to a row.
- Returns:
Values in the row as a slice of the underlying array.
- Return type:
array
Examples
>>> from landlab.utils.jaggedarray_ma import MaskedJaggedArray >>> x = MaskedJaggedArray([[0, 1, 2], [3, 4]]) >>> x.row(0) array([0, 1, 2]) >>> x.row(1) array([3, 4])