884 gobymajority, gradualkiller, grudger, grumpy #989

eric-s-s · 2017-04-25T18:45:27Z

There is a likely off-by-one error in ForgetfulGrudger. The docstring was the following.

    def strategy(self, opponent: Player) -> Action:
        """Begins by playing C, then plays D for mem_length rounds if the
        opponent ever plays D."""

mem_length was 10 but played D for 11 rounds. I changed the code to follow the docstring.

eric-s-s · 2017-04-25T18:49:51Z

axelrod/strategies/grudger.py

-    only punishes after its opponent defects a specified amount of times consecutively.
-    The punishment is in the form of a series of defections followed by a 'penance' of
-    a series of consecutive cooperations.
+    A generalization of the SoftGrudger strategy. SoftGrudger punishes by


this is simply fitting docstring to 80 chars per line.

eric-s-s · 2017-04-25T18:54:47Z

axelrod/strategies/grudger.py

@@ -67,12 +67,12 @@ def strategy(self, opponent: Player) -> Action:
            self.grudge_memory = 0
            self.grudged = False

+        if not self.grudged and D in opponent.history[-1:]:


The code and docstring didn't match. the docstring said to play D for 10 rounds, but the code played D for 11 rounds. I wasn't sure which to follow, so for now, I changed the code to fit the docstring.

for a small issue, is this a correct way to raise it?

The code and docstring didn't match. the docstring said to play D for 10 rounds, but the code played D for 11 rounds. I wasn't sure which to follow, so for now, I changed the code to fit the docstring.

for a small issue, is this a correct way to raise it?

Yup: this is fine and I think you've changed it correctly. The behaviour was not matching the description. 👍

drvinceknight

I feel these changes have made the tests a bit more obscure. If possible I'd prefer if we stuck to the very explicit:

opponent = ...
actions = ...
self.versus_test

If some of the setup for opponent and actions need to be repeated I would prefer that in favour of very explicit tests. Of course this is not a hard fast rule and in some cases there might just not be a way around things but looking through this I don't think the helper functions are helping. (Although without studying them closely I don't actually understand some of them which is my point)

drvinceknight · 2017-04-25T19:58:23Z

axelrod/tests/strategies/test_gobymajority.py

-                                [D] * (L // 2 + 1) + [C] * (L // 2 - 1))
+            opponent_actions = [C] * memory_depth + [D] * memory_depth
+            """
+            with memory_depth=3 always switches after


Could these docstrings be made in to inline comments # (or actual docstrings).

drvinceknight · 2017-04-25T20:00:43Z

axelrod/tests/strategies/test_grudger.py

-        self.responses_test([D], [C, C, D, D, D, D, C, C, D],
-                            [C, D, C, C, C, C, C, D, C])
+        vs_cooperator = [(C, C)] * 30
+        self.versus_test(axl.Cooperator(), expected_actions=vs_cooperator)


To be in line with tests already written could we stick with actions=[(C, C)] * 30 instead of vs_cooperator. (Similarly elsewhere).

drvinceknight · 2017-04-25T20:02:26Z

axelrod/tests/strategies/test_grudger.py

-                         init_kwargs={"n": 2})
-
-        # Testing n=1, d=1, c=1
-        actions = [(C, D), (D, D), (C, C), (C, C), (C, D), (D, D), (C, C), (C, C)]


Why are you re writing these tests? They are already in the correct form.

drvinceknight · 2017-04-25T20:04:55Z

axelrod/tests/strategies/test_gradualkiller.py

+        self.versus_test(axl.MockPlayer(actions=opponent),
+                         expected_actions=expected)
+
+    def get_actions_for_test(self, opponent_start, action_six_and_seven,


Unless they're truly necessary it's nicer to not make use of these helper functions and write out tests explicitly: otherwise technically we also want a test for this helper function.

If we do actually need this helper function then it should certainly be better documented.

drvinceknight · 2017-04-25T20:08:32Z

axelrod/tests/strategies/test_gobymajority.py

+
+    def test_set_soft_to_false(self):
+        self.eq_play = D
+        expected, opponent_actions = self.get_infinite_memory_depth_actions()


Not a fan of these helper functions. I have no idea what this does and subsequently am not sure what this test is doing. Can you just write out the test explicitly?

I'd prefer it if we kept the tests very clear (technically these helper functions need tests). Of course there might be a valid argument for using a helper function here but as is it just obfuscates the test here (no docstring etc). In this case for example it looks like you're just concatenating 3 lists together: so I'd just go with that.

drvinceknight · 2017-04-25T20:09:31Z

axelrod/tests/strategies/test_gobymajority.py

+        expected, opponent_actions = self.get_infinite_memory_depth_actions()
+        self.versus_test(MockPlayer(actions=opponent_actions),
+                         expected_actions=expected, init_kwargs={'soft': False})
+        self.eq_play = C


Not sure why this needs to be a class attribute. (I'm not actually sure what this is doing.)

drvinceknight · 2017-04-25T20:18:13Z

If it's helpful: please do feel free to open multiple PRs for each module (it helps with reviewing especially when/if you're going to make more than just the straightforward changes converting test_responses to versus_test). For example I did that recently with #981 and #982, which I opened one after the other, even though they were standard conversions. :)

eric-s-s · 2017-04-25T20:37:32Z

If it's helpful: please do feel free to open multiple PRs for each module (it helps with reviewing especially when/if you're going to make more than just the straightforward changes converting test_responses to versus_test).

yes. that seems like a good idea. i'll open PR's for each one. close this one? (i'll, of course be working off the comments you made and change those things before pushing up the new PR's. this will be in a couple of days)

eric-s-s · 2017-04-25T20:38:18Z

bah! wrong button. i meant to ask if i should close, not close. >(

drvinceknight · 2017-04-25T20:41:31Z

yes. that seems like a good idea. i'll open PR's for each one. close this one? (i'll, of course be working off the comments you made and change those things before pushing up the new PR's. this will be in a couple of days)

Sounds good to me :) 👍

…n#989

eric-s-s added 6 commits April 14, 2017 14:30

cosmetic changes and typehint change to gobymajority

53f8fb8

branch switch

9272327

branch switch. no big changes

ddeee72

finished gobymajority

1711b6b

finished grumpy and gradualkiller

027a365

grudger tests. fixed possible off-by-one error in ForgetfulGrudger

912cb67

eric-s-s commented Apr 25, 2017

View reviewed changes

drvinceknight requested changes Apr 25, 2017

View reviewed changes

eric-s-s closed this Apr 25, 2017

eric-s-s reopened this Apr 25, 2017

drvinceknight closed this Apr 25, 2017

eric-s-s added a commit to eric-s-s/Axelrod that referenced this pull request Apr 26, 2017

original PR state. will fix according to suggestions on Axelrod-Pytho…

96d9880

…n#989

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

884 gobymajority, gradualkiller, grudger, grumpy #989

884 gobymajority, gradualkiller, grudger, grumpy #989

eric-s-s commented Apr 25, 2017

eric-s-s Apr 25, 2017

eric-s-s Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight left a comment

drvinceknight Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight Apr 25, 2017

drvinceknight commented Apr 25, 2017 •

edited

Loading

eric-s-s commented Apr 25, 2017

eric-s-s commented Apr 25, 2017

drvinceknight commented Apr 25, 2017

884 gobymajority, gradualkiller, grudger, grumpy #989

884 gobymajority, gradualkiller, grudger, grumpy #989

Conversation

eric-s-s commented Apr 25, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

drvinceknight left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

drvinceknight commented Apr 25, 2017 • edited Loading

eric-s-s commented Apr 25, 2017

eric-s-s commented Apr 25, 2017

drvinceknight commented Apr 25, 2017

drvinceknight commented Apr 25, 2017 •

edited

Loading