diff --git a/core/algorithms/nevaluate.cnb b/core/algorithms/nevaluate.cnb
new file mode 100644
index 0000000000..964bc6bf20
--- /dev/null
+++ b/core/algorithms/nevaluate.cnb
@@ -0,0 +1,127 @@
+{
+ "cell_id": 7056241270934378170,
+ "cells": [
+ {
+ "cell_id": 6442002833303548429,
+ "cell_origin": "client",
+ "cell_type": "latex",
+ "cells": [
+ {
+ "cell_id": 14156260080777624593,
+ "cell_origin": "client",
+ "cell_type": "latex_view",
+ "source": "\\algorithm{nevaluate}{Numerically evaluate an expression.}\n\nGiven a scalar expression of one or more variables, evaluate it for a range of values \nof those variables. This algorithm accepts a Cadabra expression and one or more numpy arrays\ncontaining the values of the variables, in the form of a dictionary. A simple example with \nan expression of one variable:"
+ }
+ ],
+ "hidden": true,
+ "source": "\\algorithm{nevaluate}{Numerically evaluate an expression.}\n\nGiven a scalar expression of one or more variables, evaluate it for a range of values \nof those variables. This algorithm accepts a Cadabra expression and one or more numpy arrays\ncontaining the values of the variables, in the form of a dictionary. A simple example with \nan expression of one variable:"
+ },
+ {
+ "cell_id": 7717361411841115229,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "cells": [
+ {
+ "cell_id": 16317621934812508465,
+ "cell_origin": "server",
+ "cell_type": "latex_view",
+ "cells": [
+ {
+ "cell_id": 3643042606220920068,
+ "cell_origin": "server",
+ "cell_type": "input_form",
+ "source": "\\cos(x) \\exp( - 1/4 (x)**2)"
+ }
+ ],
+ "source": "\\begin{dmath*}{}\\cos{x} \\exp\\left( - \\frac{1}{4}{x}^{2}\\right)\\end{dmath*}"
+ }
+ ],
+ "source": "ex:= \\cos(x) \\exp(-x**2/4);\n\nimport numpy as np\n\nxv = np.linspace(0, 3, 100)\nz = np.array( nevaluate(ex, {$x$: xv} ) )"
+ },
+ {
+ "cell_id": 9285492536561123852,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "cells": [
+ {
+ "cell_id": 9668872934801381964,
+ "cell_origin": "server",
+ "cell_type": "verbatim",
+ "source": "\\begin{verbatim}[1. 0.99931146 0.99724785 0.99381515 0.98902334 0.98288631\n 0.97542182 0.96665142 0.95660038 0.94529757]\\end{verbatim}"
+ }
+ ],
+ "source": "z[0:10];"
+ },
+ {
+ "cell_id": 5511376990109163867,
+ "cell_origin": "client",
+ "cell_type": "latex",
+ "cells": [
+ {
+ "cell_id": 3830134294778153283,
+ "cell_origin": "client",
+ "cell_type": "latex_view",
+ "source": "The \\algo{nevaluate} function thus takes as its second argument a Python\ndictionary which maps each variable in the expression to a list of values.\n\nFor expressions of multiple variables, the logic is the same: just list all \nthe variables in the dictionary, as in the example below."
+ }
+ ],
+ "hidden": true,
+ "source": "The \\algo{nevaluate} function thus takes as its second argument a Python\ndictionary which maps each variable in the expression to a list of values.\n\nFor expressions of multiple variables, the logic is the same: just list all \nthe variables in the dictionary, as in the example below."
+ },
+ {
+ "cell_id": 2992782121045095320,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "cells": [
+ {
+ "cell_id": 3251060037967502714,
+ "cell_origin": "server",
+ "cell_type": "latex_view",
+ "cells": [
+ {
+ "cell_id": 6149534026651365282,
+ "cell_origin": "server",
+ "cell_type": "input_form",
+ "source": "\\cos(x) \\sin(y)"
+ }
+ ],
+ "source": "\\begin{dmath*}{}\\cos{x} \\sin{y}\\end{dmath*}"
+ }
+ ],
+ "source": "ex:= \\cos(x) \\sin(y);"
+ },
+ {
+ "cell_id": 13003303353493457346,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "source": "xv = np.linspace(0, np.pi, 100)\nyv = np.linspace(0, np.pi, 100)"
+ },
+ {
+ "cell_id": 14177849469379369091,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "source": "z = np.array( nevaluate(ex, {$x$: xv, $y$: yv}) )"
+ },
+ {
+ "cell_id": 10078872525967387623,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "cells": [
+ {
+ "cell_id": 26823108033460839,
+ "cell_origin": "server",
+ "cell_type": "verbatim",
+ "source": "\\begin{verbatim}0.3106205340965772\\end{verbatim}"
+ }
+ ],
+ "source": "z[3,10];"
+ },
+ {
+ "cell_id": 10368637782920619472,
+ "cell_origin": "client",
+ "cell_type": "input",
+ "source": ""
+ }
+ ],
+ "description": "Cadabra JSON notebook format",
+ "version": 1.0
+}
diff --git a/examples/ref_numerical.cnb b/examples/ref_numerical.cnb
index 6e05f4d2cc..a982330f94 100644
--- a/examples/ref_numerical.cnb
+++ b/examples/ref_numerical.cnb
@@ -56,14 +56,14 @@
"cell_type": "latex",
"cells": [
{
- "cell_id": 17518277435515688271,
+ "cell_id": 18054204918676316375,
"cell_origin": "client",
"cell_type": "latex_view",
- "source": "The \\verb|nevaluate| function returns an \\verb|NTensor|, which is Cadabra's\nobject to store numerical values of tensors. It can be converted to a numpy\narray, after which you can plot it:"
+ "source": "The \\algo{nevaluate} function returns an \\verb|NTensor|, which is Cadabra's\nobject to store numerical values of tensors. It can be converted to a numpy\narray by wrapping it in `np.array`, after which you can plot it:"
}
],
"hidden": true,
- "source": "The \\verb|nevaluate| function returns an \\verb|NTensor|, which is Cadabra's\nobject to store numerical values of tensors. It can be converted to a numpy\narray, after which you can plot it:"
+ "source": "The \\algo{nevaluate} function returns an \\verb|NTensor|, which is Cadabra's\nobject to store numerical values of tensors. It can be converted to a numpy\narray by wrapping it in `np.array`, after which you can plot it:"
},
{
"cell_id": 5634273073315875314,
diff --git a/web2/CMakeLists.txt b/web2/CMakeLists.txt
index 34e3a69b13..f04ea07c07 100644
--- a/web2/CMakeLists.txt
+++ b/web2/CMakeLists.txt
@@ -107,6 +107,7 @@ set(ALGOS
lr_tensor
map_sympy
meld
+ nevaluate
product_rule
raise_free_indices
reduce_delta
diff --git a/web2/cadabra2/source/index.html b/web2/cadabra2/source/index.html
index be2ab94c43..7dc0388b96 100644
--- a/web2/cadabra2/source/index.html
+++ b/web2/cadabra2/source/index.html
@@ -26,6 +26,13 @@ News
+ - 20 Sep 2023
+ - Release of 2.4.4 (changes).
+ Many new features for numerical evaluation and plotting of expressions
+ using nevaluate (see
+ the numerical
+ evaluation chapter in the book) and many bug fixes.
+
- 25 Aug 2022
- Release of 2.4.0 (changes).
Update to accompany the release of