Fix divergence in cylindrical coordinates

[lindsayad]

This fix is more general than my particular use case, but I wanted to
compute:

from sympy.vector import divergence, gradient, Vector, CoordSys3D
R = CoordSys3D('R', transformation='cylindrical')
divergence(R.i)

The correct solution to this is 1/r, however, sympy returned zero
because it was detecting that no base scalars were in the provided
expression, and then using the special case code, returning zero. A
straight-forward solution to this is removing the special case code, and
simply returning a Derivative instantiation every time.

• vector
  • Always return Derivative instance from _diff_conditional in operators.py since the coeffs may be functions of the base scalars even if expr is not

[sympy-bot]

✅

Hi, I am the SymPy bot (v162). I'm here to help you write a release notes entry. Please read the guide on how to write release notes.

Your release notes are in good order.

Here is what the release notes will look like:

• vector
  • Always return Derivative instance from _diff_conditional in operators.py since the coeffs may be functions of the base scalars even if expr is not (#20047 by @lindsayad and @smichr)

This will be added to https://github.com/sympy/sympy/wiki/Release-Notes-for-1.10.

Click here to see the pull request description that was parsed.

This fix is more general than my particular use case, but I wanted to
compute:

```python
from sympy.vector import divergence, gradient, Vector, CoordSys3D
R = CoordSys3D('R', transformation='cylindrical')
divergence(R.i)
```

The correct solution to this is 1/r, however, sympy returned zero
because it was detecting that no base scalars were in the provided
expression, and then using the special case code, returning zero. A
straight-forward solution to this is removing the special case code, and
simply returning a `Derivative` instantiation every time.

<!-- BEGIN RELEASE NOTES -->
- vector
    - Always return `Derivative` instance from `_diff_conditional` in `operators.py` since the `coeffs` may be functions of the base scalars even if `expr` is not

Update

The release notes on the wiki have been updated.

[friyaz]

Looks good. Please add test cases.

[friyaz]

@lindsayad please fix the flake test. Importing BaseScalar is not needed. I think then we can get this merged.

[czgdp1807]

@lindsayad Would you like to conclude this PR? Please let us know and do NOT close this PR. Thanks for your contributions.

[lindsayad]

I would like to see it concluded. It looks like my change broke this test:

Failed example:
    delop.dot(C.x*C.i)
Expected:
    Derivative(C.x, C.x)
Got:
    Derivative(0, C.y) + Derivative(0, C.z) + Derivative(C.x, C.x)

I'm not surprised by this. I guess from my perspective, I would rather have more verbose output than sometimes have wrong output. If you guys agree, then I will change that test. I'm also open to other suggestions.

[czgdp1807]

Derivative(0, C.y) + Derivative(0, C.z) + Derivative(C.x, C.x)

This notation seems better to me. At least it makes clear that y and z components are 0. Though what happens if we apply simplify to the above expression i.e., (Derivative(0, C.y) + Derivative(0, C.z) + Derivative(C.x, C.x)).simplify() or simplify(Derivative(0, C.y) + Derivative(0, C.z) + Derivative(C.x, C.x)).

@moorepants Any thoughts on this?

changes have been made

[github-actions]

Benchmark results from GitHub Actions

Lower numbers are good, higher numbers are bad. A ratio less than 1
means a speed up and greater than 1 means a slowdown. Green lines
beginning with + are slowdowns (the PR is slower then master or
master is slower than the previous release). Red lines beginning
with - are speedups.

Significantly changed benchmark results (PR vs master)

Significantly changed benchmark results (master vs previous release)

       before           after         ratio
     [907895ac]       [87cbdd9a]
+      7.47±0.3ms       12.1±0.8ms     1.62  matrices.TimeMatrixPower.time_Case1

Full benchmark results can be found as artifacts in GitHub Actions
(click on checks at the top of the PR).

[smichr]

I will open an issue for adding a test.

[lindsayad]

Just yesterday I was on a new machine and manually updated this file when doing some MMS testing and I thought to myself that I really need to get this in. Thanks for getting this over the finish line!