Add type hints for plot

[janga1997]

#808
@marcharper @drvinceknight Obviously there's a lot left. But I need to discuss before attempting any further.

• The parameter data I think, can be list of ints, floats, etc. I think it would be better to just do data: list.

• The parameter title too can be an int, float, and object, list or anything. I can't think of way to assign a broad enough type for it.

• I think the type of the parameter ax is matplotlib.axes._subplots.AxesSubplot, although I am not sure.

• I don't understand what is names.

[marcharper]

I think you've got ax right; title should be a strings and names is probably a list of strings.

[janga1997]

And data?

[marcharper]

Looking through the code it looks like it's a list of lists of floats (look at self._boxplot_dataset which creates it in one case). However I didn't check every instance so it might be different for some of the other plots.

[janga1997]

Yes, but it could very well be a list of list of ints in another case, right?
Wouldn't it better to just check if it is a list?

[marcharper]

Maybe List[List[Any]] is best unless int can check as a subtype of float (I'm not if that's how the numeric type hierarchy in Python works but I doubt it), or Union[int, float] which should accept either.

[marcharper]

Looks like you have some syntax errors -- check the build errors or try to import the library locally in a python interpreter.

[janga1997]

@marcharper The error has been showing up since the first commit on this PR, which was quite simple.
The only thing I changed was an
import axelrod

and set result_set: axelrod.result_set.ResultSetFromFile

I am assuming it has something to do with the way I import?

[janga1997]

@marcharper Never mind. It is an import error.
It says, ImportError: Failed to import test module: axelrod
, and then AttributeError: module 'axelrod' has no attribute 'result_set'

I saw one of your commits on type hinting, you did something to the effect of from .result_set import ResultSet .
Why does that work ?

[marcharper]

The reason that axelrod.results_set.ResultSet doesn't work is because we directly import ResultSet in the top level __init__.py for the library. Instead you would need axelrod.ResultSet.

See PEP328 for info on absolute versus relative imports (or Google around). That's what's going on with from .result_set import ResultSet.

[janga1997]

@drvinceknight @marcharper I've exhausted all the options. I've tried different ways to import ResultSet, and none of them passed, even though I am doing it the same way as in other files.
from axelrod import Actions
from .game import Game is how we do it in interaction_utils. I've tried both of them in plot, and none worked. Am I missing something silly?

And is there a way to run tests locally? The commit history on this PR is ugly.

[drvinceknight]

And is there a way to run tests locally? The commit history on this PR is ugly.

Yup, it's easy enough to run the tests locally, there's information about it here http://axelrod.readthedocs.io/en/latest/tutorials/contributing/strategy/running_tests.html but here's how you would run the tests on the plot.py file:

$ python -m unittest axelrod.tests.unit.test_plot

I've just done that locally and gotten the following error:

Axelrod(janga1997-typeHint-plot) ✗: python -m unittest axelrod.tests.unit.test_plot
Traceback (most recent call last):
  File "/home/vince/anaconda3/lib/python3.5/runpy.py", line 184, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/vince/anaconda3/lib/python3.5/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/vince/anaconda3/lib/python3.5/unittest/__main__.py", line 18, in <module>
    main(module=None)
  File "/home/vince/anaconda3/lib/python3.5/unittest/main.py", line 93, in __init__
    self.parseArgs(argv)
  File "/home/vince/anaconda3/lib/python3.5/unittest/main.py", line 140, in parseArgs
    self.createTests()
  File "/home/vince/anaconda3/lib/python3.5/unittest/main.py", line 147, in createTests
    self.module)
  File "/home/vince/anaconda3/lib/python3.5/unittest/loader.py", line 219, in loadTestsFromNames
    suites = [self.loadTestsFromName(name, module) for name in names]
  File "/home/vince/anaconda3/lib/python3.5/unittest/loader.py", line 219, in <listcomp>
    suites = [self.loadTestsFromName(name, module) for name in names]
  File "/home/vince/anaconda3/lib/python3.5/unittest/loader.py", line 153, in loadTestsFromName
    module = __import__(module_name)
  File "/home/vince/src/Axelrod/axelrod/__init__.py", line 11, in <module>
    from .plot import Plot
  File "/home/vince/src/Axelrod/axelrod/plot.py", line 23, in <module>
    axType = matplotlib.axes._subplots.AxesSubplot
AttributeError: module 'matplotlib.axes._subplots' has no attribute 'AxesSubplot'

Which indicates that axType = matplotlib.axes._subplots.AxesSubplot isn't the correct type.

[janga1997]

@drvinceknight I removed the AxesSubplot, and it passed the tests, but that doesn't seem right.
Because we do _, ax = plt.subplots(), which means that ax is the second element in the tuple plt.subplots().
I checked type(plt.subplots()[1]) and it returned matplotlib.axes._subplots.AxesSubplot, which is what I did. Where am I going wrong?

[drvinceknight]

@drvinceknight I removed the AxesSubplot, and it passed the tests, but that doesn't seem right.
Because we do _, ax = plt.subplots(), which means that ax is the second element in the tuple plt.subplots().
I checked type(plt.subplots()[1]) and it returned matplotlib.axes._subplots.AxesSubplot, which is what I did. Where am I going wrong?

It looks like that's a dynamically created class. Here's a SO question with relevant info: http://stackoverflow.com/questions/11690597/there-is-a-class-matplotlib-axes-axessubplot-but-the-module-matplotlib-axes-has

Try changing it to:

axType = matplotlib.axes.SubplotBase

[drvinceknight]

Try changing it to:

axType = matplotlib.axes.SubplotBase

when I do that on this branch the tests pass fine but I get a few other errors when running mypy.

[janga1997]

@drvinceknight Can we take this to gitter?

[drvinceknight]

This looks good to me, one final thing could you add axelrod/plots.py to type_tests.sh please @janga1997?

[drvinceknight]

Thanks! Looks good to me.