You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* feat: `meshgrid` (#740)
• Default implementation equivalent to NumPy's with `copy=True` and `sparse=False`
• Add 2D, 3D, 4D test cases
• Update docs
* docs: update version to 6.12.0
Copy file name to clipboardExpand all lines: docs/ulab-ndarray.ipynb
+55-1Lines changed: 55 additions & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -507,7 +507,7 @@
507
507
"source": [
508
508
"# Array initialisation functions\n",
509
509
"\n",
510
-
"There are nine functions that can be used for initialising an array. Starred functions accept `complex` as the value of the `dtype`, if the firmware was compiled with complex support.\n",
510
+
"There are several functions that can be used for initialising an array. Starred functions accept `complex` as the value of the `dtype`, if the firmware was compiled with complex support.\n",
"The `meshgrid` function returns coordinate matrices from coordinate vectors. Given *N* one-dimensional arrays, it returns a tuple of *N* N-dimensional arrays.\n",
1165
+
"\n",
1166
+
"Make sure *N* is not larger than the maximum dimension of your firmware.\n",
1167
+
"\n",
1168
+
"The `indexing` keyword argument determines the arrangement of the output. The default is `'xy'` (Cartesian indexing), where the first two dimensions are swapped. This is commonly used for generating grids for 2D plotting. If `indexing` is set to `'ij'` (matrix indexing), the dimensions of the output arrays correspond exactly to the order of the input arrays.\n",
1169
+
"\n",
1170
+
"Note that unlike `numpy`, `ulab`'s implementation always behaves as if `sparse=False` and `copy=True` i.e. it always returns fully allocated dense arrays."
0 commit comments