Christopher Smith

tests were somehow added in #25579

@AdvaitPote @moorepants , please give this another look.

Further modifications can be made in a new PR.

yes - if there was no test for this sort of thing added, then one should be added. Otherwise, if there was a test added, then this can just be...

In the `test_simplify.py` where nsimplify is being tested, add a line that says `assert nsimplify(pi/7.0) == pi/7`. But we already have `assert nsimplify(pi/1e2) == pi/100` so I think this can...

> I am unable to reproduce this issue. I am running Perhaps the OP intended to show `.doit()` on the Integral?

I am not aware of anything more that is needed for this. The test suite does not give much feedback with this change -- only one test was changed for...

> This breaks a key invariant I am guessing that I will have to dig into `from_sympy` to fix this.

> if there isn't any reason to do so. The motivation to do so is to make Rational-based powers behave the same as symbolic bases: ```python >>> [len(Poly((x+x**2).simplify()).gens) for x...

Yes, the factorization bothers me, too. Perhaps the better thing to do is attempt a unification of generators that are powers in Poly. Finding gcd is easy compared to factorization.