typo in and suggestions for SHGO docs

[podkop]

Page https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.shgo.html,
the description of "constraints" argument ("constraints : dict or sequence of dict, optional"),
2-nd line:
g(x) <= 0 ...

Correct text: g(x) >= 0 ...

Explanation: the inequality constraints ("type": "ineq") are actually treated in the function as >= constraints. It is also confirmed in an example below on the same page (after the line "To demonstrate solving problems with non-linear constraints consider ...").

As a side note from a simple user: some phrases in the scipy docs are formally incorrect.

For example, in the same description of the "constraints" argument we read: "Function(s) R**n in the form:", while below we see the form of constraints, not just functions.
Besides that, below we see sets of constraints, and corresponding vector-valued functions (R**n -> R**p), but then we read "Each constraint is defined..." (the singular form suggesting a single constraint defined by a scalar function). In the example below, only single constraints are defined.

I would write something like following (assuming one can define a set of constraints at once):

Constraints definitions. A sequence where each element defines a set of constraints in either of the following forms:
g(x) >= 0 defined by a function g : R^n -> R^m
h(x) == 0 defined by a function h : R^n -> R^p
Each system of constraints is defined in a dictionary with fields:
...

[rgommers]

There's a couple of suggestions here that @Stefan-Endres may want to incorporate in SHGO, so pinging him.

Thanks for the report @podkop. Note that the correct repo for SciPy documentation issues is https://github.com/scipy/scipy (this one is just for the top-level website), so please use that for future issues.

[Stefan-Endres]

Thank you very much @podkop for the detailed issue and @rgommers for pinging me. I'm almost ready to release a major performance update as well as a number of bug fixes in the stand alone repository, so I will include this doc fix issue and make a pull request to scipy with the latest version.

[rgommers]

Thanks @Stefan-Endres!

I'll close this issue then, since you have one on your own repo.