Add type annotations to pyreverse

[mbyrnepr2]

Closes #1548

Steps

• Add yourself to CONTRIBUTORS if you are a new contributor.
• Add a ChangeLog entry describing what your PR does.
• If it's a new feature or an important bug fix, add a What's New entry in
  doc/whatsnew/<current release.rst>.
• Write a good description on what the PR does.

Description

demo.py:

from typing import Dict, List, NamedTuple


class A:
    classvar1: Dict = None
    classvar2 = 'classvar2'

    def __init__(self, a: str = None, b: NamedTuple = None, c = 1):
        self.a = a
        self.b = b
        self.c = c

    def m1(self, i: str, j: int) -> List:
        pass

development branch:

master branch:

Type of Changes

	Type
✓	✨ New feature

Related Issue

[coveralls]

Coverage increased (+0.02%) to 92.036% when pulling 194e6d3 on mbyrnepr2:1548_pyreverse_type_annotations into 99589b0 on PyCQA:master.

[Pierre-Sassoulas]

Thank you for this changes :) I made some suggestions that could maybe simplify the code let me know if there is something that I did not see :)

[Pierre-Sassoulas]

Suggested change

	ann = getattr(node.parent, "annotation", None)
	values = {ann} if ann else set(node.infer())
	values = {getattr(node.parent, "annotation", node.infer())}

Would that work ?

[Pierre-Sassoulas]

Suggested change

	ann = None
	ann = node.infer()

I think doing that would simplify that following code.

[mbyrnepr2]

@Pierre-Sassoulas thank you! I'll try that later today along with the further changes required to display variables a optional in the diagram.

[alexchandel]

This would be fantastic! There is no Python UML generator that understands type annotations right now. This also lays the groundwork for the future generation of dependency arrows for class attributes.

@mbyrnepr2 Did the requested changes work?

[mbyrnepr2]

@alexchandel the suggested changes work perfectly. Aiming to update the pr with those + Optional output this week.

[Pierre-Sassoulas]

@DudeNr33 I don't know if you saw that MR but someone else if working on pyreverse in parallel :) Maybe you can rebase on it as soon as it's merged to avoid handling conflicts late.

[DudeNr33]

Thanks for the heads up!
Great work, I actually also had the corresponding issue bookmarked for some time now because I think this is a great improvement! Will rebase my own branch as soon as I got time.

[Pierre-Sassoulas]

Thank you for going back to this. I think there's should be a handling for the inference fail that can happen with infer() call. I don't have an example of code where there inference would fail, maybe we can patch NodeNg so it raises one in order to be able to test ?

[Pierre-Sassoulas]

Here there can be an inference fail, we should catch it (like line 238). Btw it look like the code is similar, maybe we can create a function for it ?

[Pierre-Sassoulas]

There can be an InferenceError raised here.

[mbyrnepr2]

Thanks again @Pierre-Sassoulas I agree with all the above. I've aimed to address the points in 194e6d3. I could easily have missed something though so happy to look again if something doesn't look right.

[Pierre-Sassoulas]

Thank you for all the work that went into this !