Skip to content

Indexing Module#

The indexing module provides efficient functions for working with numpy arrays, particularly for indexing and extracting values based on various conditions.

Functions#

get_indices(arr, mask_val) #

Get the array indices for the non-zero cells.

This function returns the row and column indices of cells in a 2D array that match a specific value or all non-zero values.

Parameters:

Name Type Description Default
arr ndarray

2D array with values you need to get the indexes of the cells that are filled with these values.

 required
mask_val Union[int, float]

If you need to locate only a certain value, and not all values in the array. If None or falsy, will return indices of all non-zero values.

 required

Returns:

Type Description
Tuple[ndarray, ndarray]

A tuple of two numpy arrays: - first array is the row indices - second array is the column indices

Raises:

Type Description
ValueError

If the input array is not 2D.

Examples:

  • Import numpy 
    >>> import numpy as np
  • Create a sample array 
    >>> arr = np.array([[0, 1, 0], [0, 0, 2], [3, 0, 0]])
  • Get indices of all non-zero values 
    >>> i, j = get_indices(arr, None)
>>> print(i)
[0 1 2]
>>> print(j)
[1 2 0]
  • Get indices of a specific value 
    >>> i, j = get_indices(arr, 2)
>>> print(i)
[1]
>>> print(j)
[2]
Source code in src/hpc/indexing.py 
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
def get_indices(
    arr: np.ndarray, mask_val: Union[int, float]
) -> Tuple[np.ndarray, np.ndarray]:
    """Get the array indices for the non-zero cells.

    This function returns the row and column indices of cells in a 2D array that match
    a specific value or all non-zero values.

    Args:
        arr: 2D array with values you need to get the indexes of the cells that are filled with these values.
        mask_val: If you need to locate only a certain value, and not all values in the array.
            If None or falsy, will return indices of all non-zero values.

    Returns:
        A tuple of two numpy arrays:
            - first array is the row indices
            - second array is the column indices

    Raises:
        ValueError: If the input array is not 2D.

    Examples:
        - Import numpy
        ```python
        >>> import numpy as np

        ```
        - Create a sample array
        ```python
        >>> arr = np.array([[0, 1, 0], [0, 0, 2], [3, 0, 0]])

        ```
        - Get indices of all non-zero values
        ```python
        >>> i, j = get_indices(arr, None)
        >>> print(i)
        [0 1 2]
        >>> print(j)
        [1 2 0]

        ```
        - Get indices of a specific value
        ```python
        >>> i, j = get_indices(arr, 2)
        >>> print(i)
        [1]
        >>> print(j)
        [2]

        ```
    """
    # Use the arr to get the indices of the non-zero pixels.
    if mask_val is not None:
        (i, j) = (arr == mask_val).nonzero()
    else:
        (i, j) = arr.nonzero()

    return i, j

get_indices2(arr, mask=None) #

Get indices of array cells after filtering values based on mask values.

This function returns the indices of array cells that don't match the values in the mask. If mask is None, returns indices of all cells in the array. This function is particularly useful for filtering out specific values (like NoData values) from an array.

Parameters:

Name Type Description Default
arr ndarray

2D numpy array to get indices from.

 required
mask Optional[List[Union[int, float, number]]]

List of values to exclude from the result. - If None, returns indices of all cells. - If list with one value, returns indices of cells not equal to that value. - If list with two values, returns indices of cells not equal to either value.

 None

Returns:

Type Description
List[Tuple[int, int]]

List of tuples (row, col) representing the indices of cells that pass the filter.

Raises:

Type Description
ValueError

If mask contains more than two values.
ValueError

If the input array is not 2D.

Examples:

  • Import numpy 
    >>> import numpy as np
  • Create a sample array 
    >>> arr = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
  • Get all indices 
    >>> indices = get_indices2(arr, None)
>>> len(indices)  # 3x3 array = 9 indices
9
  • Filter out cells with value 5 
    >>> indices = get_indices2(arr, [5])
>>> sorted(indices) #doctest: +SKIP
[(0, 0), (0, 1), (0, 2), (1, 0), (1, 2), (2, 0), (2, 1), (2, 2)]
  • Filter out cells with values 1 and 9 
    >>> indices = get_indices2(arr, [1, 9])
>>> sorted(indices)     #doctest: +SKIP
[(0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1)]
  • Works with NaN values 
    >>> arr_with_nan = np.array([[1, 2, np.nan], [4, 5, 6]])
>>> indices = get_indices2(arr_with_nan, [np.nan])
>>> sorted(indices) #doctest: +SKIP
[(0, 0), (0, 1), (1, 0), (1, 1), (1, 2)]
Source code in src/hpc/indexing.py 
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
def get_indices2(
    arr: np.ndarray, mask: Optional[List[Union[int, float, np.number]]] = None
) -> List[Tuple[int, int]]:
    """Get indices of array cells after filtering values based on mask values.

    This function returns the indices of array cells that don't match the values in the mask.
    If mask is None, returns indices of all cells in the array. This function is particularly
    useful for filtering out specific values (like NoData values) from an array.

    Args:
        arr: 2D numpy array to get indices from.
        mask: List of values to exclude from the result.
            - If None, returns indices of all cells.
            - If list with one value, returns indices of cells not equal to that value.
            - If list with two values, returns indices of cells not equal to either value.

    Returns:
        List of tuples (row, col) representing the indices of cells that pass the filter.

    Raises:
        ValueError: If mask contains more than two values.
        ValueError: If the input array is not 2D.

    Examples:
        - Import numpy
        ```python
        >>> import numpy as np

        ```
        - Create a sample array
        ```python
        >>> arr = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

        ```
        - Get all indices
        ```python
        >>> indices = get_indices2(arr, None)
        >>> len(indices)  # 3x3 array = 9 indices
        9

        ```
        - Filter out cells with value 5
        ```python
        >>> indices = get_indices2(arr, [5])
        >>> sorted(indices) #doctest: +SKIP
        [(0, 0), (0, 1), (0, 2), (1, 0), (1, 2), (2, 0), (2, 1), (2, 2)]

        ```
        - Filter out cells with values 1 and 9
        ```python
        >>> indices = get_indices2(arr, [1, 9])
        >>> sorted(indices)     #doctest: +SKIP
        [(0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1)]

        ```
        - Works with NaN values
        ```python
        >>> arr_with_nan = np.array([[1, 2, np.nan], [4, 5, 6]])
        >>> indices = get_indices2(arr_with_nan, [np.nan])
        >>> sorted(indices) #doctest: +SKIP
        [(0, 0), (0, 1), (1, 0), (1, 1), (1, 2)]

        ```
    """
    # get the position of cells that is not zeros
    if mask is not None:
        if len(mask) > 1:
            mask = np.logical_and(
                ~np.isclose(arr, mask[0], rtol=0.001),
                ~np.isclose(arr, mask[1], rtol=0.001),
            )
        else:
            if np.isnan(mask[0]):
                mask = ~np.isnan(arr)
            else:
                mask = ~np.isclose(arr, mask[0], rtol=0.001)

        rows = np.nonzero(mask)[0]
        cols = np.nonzero(mask)[1]

        ind = list(zip(rows, cols))
    else:
        rows = arr.shape[0]
        cols = arr.shape[1]
        ind = [(i, j) for i in range(rows) for j in range(cols)]

    return ind

get_pixels(arr, mask=None, mask_val=None) #

Get pixels from a raster (with optional mask).

This function extracts pixel values from an array based on a mask. It can work with both 2D and 3D arrays and can extract pixels based on specific mask values.

Parameters:

Name Type Description Default
arr ndarray

Array of raster data in the form [bands][y][x] or [y][x].

 required
mask ndarray

Array (2D) of values to mask data (from rasterizing a vector). If None, returns the original array.

 None
mask_val Union[int, float, None]

Value of the data pixels in the mask to extract. If None or falsy, will extract all non-zero values.

 None

Returns:

Type Description
ndarray

np.ndarray: Array of non-masked data. - For 2D input: 1D array of values - For 3D input: 2D array with bands as rows and filtered pixels as columns

Raises:

Type Description
ValueError

If the mask dimensions don't match the array dimensions.

Examples:

  • Import numpy 
    >>> import numpy as np
  • 2D array example 
    >>> arr_2d = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
>>> mask = np.array([[0, 1, 0], [0, 0, 1], [0, 0, 0]])
>>> get_pixels(arr_2d, mask)
array([2, 6])
  • 3D array example 
    >>> arr_3d = np.array([[[1, 2, 3], [4, 5, 6], [7, 8, 9]],
...                    [[10, 20, 30], [40, 50, 60], [70, 80, 90]]])
>>> get_pixels(arr_3d, mask)
array([[ 2,  6],
       [20, 60]])
  • With specific mask value 
    >>> mask2 = np.array([[0, 2, 0], [0, 0, 2], [0, 0, 0]])
>>> get_pixels(arr_2d, mask2, 2)
array([2, 6])
Source code in src/hpc/indexing.py 
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
def get_pixels(
    arr: np.ndarray, mask: np.ndarray = None, mask_val: Union[int, float, None] = None
) -> np.ndarray:
    """Get pixels from a raster (with optional mask).

    This function extracts pixel values from an array based on a mask. It can work with
    both 2D and 3D arrays and can extract pixels based on specific mask values.

    Args:
        arr: Array of raster data in the form [bands][y][x] or [y][x].
        mask: Array (2D) of values to mask data (from rasterizing a vector).
            If None, returns the original array.
        mask_val: Value of the data pixels in the mask to extract.
            If None or falsy, will extract all non-zero values.

    Returns:
        np.ndarray: Array of non-masked data.
            - For 2D input: 1D array of values
            - For 3D input: 2D array with bands as rows and filtered pixels as columns

    Raises:
        ValueError: If the mask dimensions don't match the array dimensions.

    Examples:
        - Import numpy
        ```python
        >>> import numpy as np

        ```
        - 2D array example
        ```python
        >>> arr_2d = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
        >>> mask = np.array([[0, 1, 0], [0, 0, 1], [0, 0, 0]])
        >>> get_pixels(arr_2d, mask)
        array([2, 6])

        ```
        - 3D array example
        ```python
        >>> arr_3d = np.array([[[1, 2, 3], [4, 5, 6], [7, 8, 9]],
        ...                    [[10, 20, 30], [40, 50, 60], [70, 80, 90]]])
        >>> get_pixels(arr_3d, mask)
        array([[ 2,  6],
               [20, 60]])

        ```
        - With specific mask value
        ```python
        >>> mask2 = np.array([[0, 2, 0], [0, 0, 2], [0, 0, 0]])
        >>> get_pixels(arr_2d, mask2, 2)
        array([2, 6])

        ```
    """
    if mask is None:
        return arr

    # Validate that mask dimensions match array dimensions
    if arr.ndim == 2:
        if mask.shape != arr.shape:
            raise ValueError(
                f"Mask shape {mask.shape} does not match array shape {arr.shape}"
            )
    elif arr.ndim == 3:
        if mask.shape != arr.shape[1:]:
            raise ValueError(
                f"Mask shape {mask.shape} does not match array shape {arr.shape[1:]}"
            )
    else:
        raise ValueError(f"Unsupported array dimensions: {arr.ndim}")

    i, j = get_indices(mask, mask_val)
    # get the corresponding values to the indices from the array
    vals = arr[i, j] if arr.ndim == 2 else arr[:, i, j]
    return vals

get_pixels2(arr, mask=None) #

Get pixels from a raster using the get_indices2 function for filtering.

This function extracts pixel values from an array based on indices that don't match the mask values. It works with both 2D and 3D arrays and is particularly useful for filtering out specific values (like NoData values) from an array.

Parameters:

Name Type Description Default
arr ndarray

Array of raster data in the form [y][x] for 2D arrays or [bands][y][x] for 3D arrays.

 required
mask Optional[List[Union[int, float, number]]]

List of values to exclude from the result. - If None, returns all pixels. - If list with values, returns pixels not matching those values. See get_indices2 for more details on mask behavior.

 None

Returns:

Type Description
ndarray

np.ndarray: Array of filtered pixel values. - For 2D input: 1D array of values - For 3D input: 2D array with bands as rows and filtered pixels as columns

Raises:

Type Description
ValueError

If mask contains more than two values.
ValueError

If the input array dimensions are not supported (must be 2D or 3D).

Examples:

  • Import numpy 
    >>> import numpy as np
  • 2D array example 
    >>> arr_2d = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
  • Get all pixels 
    >>> pixels = get_pixels2(arr_2d, None)
>>> len(pixels)
9
  • Filter out pixels with value 5 
    >>> pixels = get_pixels2(arr_2d, [5])
>>> sorted(pixels) #doctest: +SKIP
[1, 2, 3, 4, 6, 7, 8, 9]
  • 3D array example 
    >>> arr_3d = np.array(
...     [
...         [[1, 2, 3], [4, 5, 6]],
...         [[10, 20, 30], [40, 50, 60]]
...     ]
... )
  • Filter out pixels with value 5 and 50 
    >>> pixels = get_pixels2(arr_3d, [5, 50])
>>> pixels.shape
(2, 5)
>>> pixels[0]  # First band values #doctest: +SKIP
array([1, 2, 3, 4, 6])
>>> pixels[1]  # Second band values #doctest: +SKIP
array([10, 20, 30, 40, 60])
Source code in src/hpc/indexing.py 
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
def get_pixels2(
    arr: np.ndarray, mask: Optional[List[Union[int, float, np.number]]] = None
) -> np.ndarray:
    """Get pixels from a raster using the get_indices2 function for filtering.

    This function extracts pixel values from an array based on indices that don't match the mask values.
    It works with both 2D and 3D arrays and is particularly useful for filtering out specific values
    (like NoData values) from an array.

    Args:
        arr: Array of raster data in the form [y][x] for 2D arrays or [bands][y][x] for 3D arrays.
        mask: List of values to exclude from the result.
            - If None, returns all pixels.
            - If list with values, returns pixels not matching those values.
            See get_indices2 for more details on mask behavior.

    Returns:
        np.ndarray: Array of filtered pixel values.
            - For 2D input: 1D array of values
            - For 3D input: 2D array with bands as rows and filtered pixels as columns

    Raises:
        ValueError: If mask contains more than two values.
        ValueError: If the input array dimensions are not supported (must be 2D or 3D).

    Examples:
        - Import numpy
        ```python
        >>> import numpy as np

        ```
        - 2D array example
        ```python
        >>> arr_2d = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

        ```
        - Get all pixels
        ```python
        >>> pixels = get_pixels2(arr_2d, None)
        >>> len(pixels)
        9

        ```
        - Filter out pixels with value 5
        ```python
        >>> pixels = get_pixels2(arr_2d, [5])
        >>> sorted(pixels) #doctest: +SKIP
        [1, 2, 3, 4, 6, 7, 8, 9]

        ```
        - 3D array example
        ```python
        >>> arr_3d = np.array(
        ...     [
        ...         [[1, 2, 3], [4, 5, 6]],
        ...         [[10, 20, 30], [40, 50, 60]]
        ...     ]
        ... )

        ```
        - Filter out pixels with value 5 and 50
        ```python
        >>> pixels = get_pixels2(arr_3d, [5, 50])
        >>> pixels.shape
        (2, 5)
        >>> pixels[0]  # First band values #doctest: +SKIP
        array([1, 2, 3, 4, 6])
        >>> pixels[1]  # Second band values #doctest: +SKIP
        array([10, 20, 30, 40, 60])

        ```
    """
    if arr.ndim == 2:
        ind = get_indices2(arr, mask)
        fn = lambda x: arr[x[0], x[1]]
        values = np.fromiter(map(fn, ind), dtype=arr.dtype)
    else:
        ind = get_indices2(arr[0, :, :], mask)
        fn = lambda x: arr[:, x[0], x[1]]
        values = list(map(fn, ind))
        values = np.array(values, dtype=arr.dtype)
        values = values.transpose()

    return values

locate_values(values, grid_x, grid_y) #

Locate coordinates in a grid by finding the closest grid points.

This function takes a set of (x,y) coordinates and finds the closest matching indices in the provided grid_x and grid_y arrays. It's particularly useful for mapping point data to grid cells in spatial analysis and interpolation tasks.

Parameters:

Name Type Description Default
values ndarray

Array with shape (n, 2), where each row contains [x, y] coordinates to locate. Example: 

array([[454795, 503143],
       [443847, 481850],
       [454044, 481189]])

required
grid_x ndarray

Array of x-coordinates (west to east) to search within. These are typically the x-coordinates of a grid.

 required
grid_y ndarray

Array of y-coordinates (north to south) to search within. These are typically the y-coordinates of a grid.

 required

Returns:

Type Description
ndarray

np.ndarray: Array with shape (n, 2) containing the [row, col] indices of the closest grid points to each input coordinate. Example: 

array([[ 5,  4],
       [ 2,  9],
       [ 5,  9]])

Raises:

Type Description
ValueError

If values array doesn't have shape (n, 2).
ValueError

If grid_x or grid_y are empty arrays.

Examples:

  • Import numpy 
    >>> import numpy as np
  • Create sample coordinates to locate 
    >>> coords = np.array([[10, 20], [30, 40], [50, 60]])
  • Create grid coordinates 
    >>> grid_x = np.array([0, 10, 20, 30, 40, 50])
>>> grid_y = np.array([0, 20, 40, 60, 80])
  • Find the closest grid indices 
    >>> indices = locate_values(coords, grid_x, grid_y)
>>> print(indices)
[[1 1]
 [3 2]
 [5 3]]
  • Verify the first coordinate [10, 20] maps to grid_x[1]=10, grid_y[1]=20 
    >>> grid_x[indices[0, 0]], grid_y[indices[0, 1]]
(np.int64(10), np.int64(20))
  • Verify the second coordinate [30, 40] maps to grid_x[3]=30, grid_y[2]=40 
    >>> grid_x[indices[1, 0]], grid_y[indices[1, 1]]
(np.int64(30), np.int64(40))
  • Verify the third coordinate [50, 60] maps to grid_x[5]=50, grid_y[3]=60 
    >>> grid_x[indices[2, 0]], grid_y[indices[2, 1]]
(np.int64(50), np.int64(60))
Source code in src/hpc/indexing.py 
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
def locate_values(
    values: np.ndarray, grid_x: np.ndarray, grid_y: np.ndarray
) -> np.ndarray:
    """Locate coordinates in a grid by finding the closest grid points.

    This function takes a set of (x,y) coordinates and finds the closest matching indices
    in the provided grid_x and grid_y arrays. It's particularly useful for mapping point
    data to grid cells in spatial analysis and interpolation tasks.

    Args:
        values: Array with shape (n, 2), where each row contains [x, y] coordinates to locate.
            Example:
            ```
            array([[454795, 503143],
                   [443847, 481850],
                   [454044, 481189]])
            ```
        grid_x: Array of x-coordinates (west to east) to search within.
            These are typically the x-coordinates of a grid.
        grid_y: Array of y-coordinates (north to south) to search within.
            These are typically the y-coordinates of a grid.

    Returns:
        np.ndarray: Array with shape (n, 2) containing the [row, col] indices of the
            closest grid points to each input coordinate.
            Example:
            ```
            array([[ 5,  4],
                   [ 2,  9],
                   [ 5,  9]])
            ```

    Raises:
        ValueError: If values array doesn't have shape (n, 2).
        ValueError: If grid_x or grid_y are empty arrays.

    Examples:
        - Import numpy
        ```python
        >>> import numpy as np

        ```
        - Create sample coordinates to locate
        ```python
        >>> coords = np.array([[10, 20], [30, 40], [50, 60]])

        ```
        - Create grid coordinates
        ```python
        >>> grid_x = np.array([0, 10, 20, 30, 40, 50])
        >>> grid_y = np.array([0, 20, 40, 60, 80])

        ```
        - Find the closest grid indices
        ```python
        >>> indices = locate_values(coords, grid_x, grid_y)
        >>> print(indices)
        [[1 1]
         [3 2]
         [5 3]]

        ```
        - Verify the first coordinate [10, 20] maps to grid_x[1]=10, grid_y[1]=20
        ```python
        >>> grid_x[indices[0, 0]], grid_y[indices[0, 1]]
        (np.int64(10), np.int64(20))

        ```
        - Verify the second coordinate [30, 40] maps to grid_x[3]=30, grid_y[2]=40
        ```python
        >>> grid_x[indices[1, 0]], grid_y[indices[1, 1]]
        (np.int64(30), np.int64(40))

        ```
        - Verify the third coordinate [50, 60] maps to grid_x[5]=50, grid_y[3]=60
        ```python
        >>> grid_x[indices[2, 0]], grid_y[indices[2, 1]]
        (np.int64(50), np.int64(60))

        ```
    """
    # Validate inputs
    if len(grid_x) == 0:
        raise ValueError("grid_x cannot be empty")
    if len(grid_y) == 0:
        raise ValueError("grid_y cannot be empty")

    # Handle empty values array
    if len(values) == 0:
        return np.zeros((0, 2), dtype=int)

    # Validate values shape
    if values.ndim != 2 or values.shape[1] != 2:
        raise ValueError(f"values must have shape (n, 2), got {values.shape}")

    def find(point_i):
        x_ind = np.abs(point_i[0] - grid_x).argmin()
        y_ind = np.abs(point_i[1] - grid_y).argmin()
        return x_ind, y_ind

    indices = np.array(list(map(find, values)))

    return indices