* The rewrite no longer bails out when dimshuffle affects both unique param dimensions and repeated param dimensions from the size argument. This requires:
  1) Adding broadcastable dimensions to the parameters, which should be "cost-free" and would need to be done in the `perform` method anyway.
  2) Extend size to incorporate implicit batch dimensions coming from the parameters. This requires computing the shape resulting from broadcasting the parameters. It's unclear whether this is less performant, because the `perform` method can now simply broadcast each parameter to the size, instead of having to broadcast the parameters together.
* The rewrite now works with Multivariate RVs
* The rewrite bails out when dimensions are dropped by the Dimshuffle. This case was not correctly handled by the previous rewrite

This allows PyTensor to infer more broadcastable patterns, by placing the casting inside the MakeVector Op

Co-authored-by: Maxim Kochurov <max.kochurov@pymc-devs.org>

* Redirect installation issues to Discourse

* Remove required reference to documentation on template issue

* Remove required reference to discussion on feature request template

* Fix automatic label on feature request template

* Replace plain issue template with free form developer template

* Fix line endings of empty files

* Use `not in` for membership test

* Remove unused imports

* Fix implicit string concatenation

* Remove python 2 coding statement

* Change python2's Text into str

* Remove useless linebreaks

* Remove useless parentheses

* Use new-style dictionary and set constructors

* There is no more IOError (alias to OSError now)

* Use .format instead of % string interpolation

* Use super() without redundant arguments

* Remove redundant read flag from open

* Avoid building lists unnecessarily

* Use f-strings with python 3.6

* Black reformat
Co-authored-by: Michael Osthege <michael.osthege@outlook.com>

Co-authored-by: Maxim Kochurov <maxim.v.kochurov@gmail.com>

- Lazily create and cache `FunctionGraph`s, the `Composite.perform`
  implementation, C code, and name values
- Use `fgraph_to_python` for `Composite.perform`
- Use the `HasInnerGraph` interface