benruijl / symbolica Goto Github PK

Hello,

I tried running the program below,

from symbolica import Expression
e = Expression.num(0)
print(e)
print(e.factor())

and it produced

0
1

I'm wondering if that's expected. Thanks!

Sorry i couldn't find any "comment" feature on your blog so i decided to write here.
Your 'loop' pattern can be written like this:

    let mut items = vec![3, 5, 7, 11, 13, 17];
    let item = 19;
    let location = items.iter().position(|x| *x == item).or_else(|| {
        items.push(item);
        Some(items.len() - 1)
    });

To reproduce:

>>> x, y = Expression.vars('x', 'y')
>>> e = 5*x + x * y + x**2 + 5
>>> print(e.coefficient(x*y))
0

But of course it should be 1.

With the current version (0.8.0) of symbolica that includes the float representations I see that rational factors in python are sometimes converted to floats. This example demonstrates where depending on how I write the expression it either is represented as a float or a rational number:

e = Expression.symbol("e")
e1 = 2/5 * e
e2 = 2 * e / 5
print(e1)  # gives:  4.00000000000000e-1*e
print(e2)  # gives:  2/5*e

This page just isn't very useful, removing it should be as simple as removing the pub keyword in front of the module name.

Since Symbolica is advertised as a blazing fast language, are there any benchmark comparisons to Form?

Imagined use: Suppose we have an ODE and we wish to solve it with the Frobenius method. The we would like to convert the ansatz

$$y(t) = \sum_{i=0}^{\infty} c_{i} t^{i+\alpha}$$

to some recurrence involving only the $c_i$. Can this be done in an an automated way in symbolica? It seems this would require support for an infinite array of symbols, maybe?

c = Expression.var('c[0:infinity]')

or

c = Expression.vector_var('c[0:infinity]')

I submit this issue with some trepidation as I know absolutely nothing about this field and have no clue if I've made a feature request for something that's NP complete or undecidable.