Fix AgentSet inplace shuffle (and thus RandomActivation), add tests #2007
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Adds a test that checks if the RandomActivation doesn't trigger agents in a sequential order. In theory this could give false positives (a test passing when it shouldn't, but that chance is around ~0.1^18).
EwoutH
force-pushed
the
test-randomactivation
branch
from
612caa2 to
a7a2c87
Compare
|
Performance benchmarks:
|
|
If you change def do_each(self, method, shuffle=False):
if shuffle:
self._agents.shuffle(inplace=True)
self._agents.do(method)it works correctly. I am going to check RandomActivationByType as well. Do I have write access to this? |
|
I also think this bug fix will actually produce a speedup because we avoid a needless copy operation of the AgentSet. RandomActivationByType should not have this bug in the first place because it does not rely on |
5 tasks
quaquel
force-pushed
the
test-randomactivation
branch
from
b7603d9 to
90f0408
Compare
|
Thanks! For clarity, we should document the difference between |
|
@quaquel could you add unittests for |
|
do you mean for AgentSet.shuffle? |
for more information, see https://pre-commit.ci
|
Can you trigger the benchmarks? I am curious to see what this fix has done with those given that it removes an unnecessary copy operation. |
|
/rerun-benchmarks |
…into test-randomactivation
for more information, see https://pre-commit.ci
|
Looks good! Benchmarks depend on #2002. Can’t approve this PR, but looks good to me. Recommend squash merge. |
EwoutH
commented
Jan 26, 2024
EwoutH
changed the title
fix for the second issue in projectmesa#2006. If super is not present, we create the data structure but forget to add the agent to it. This is just a backward compatibility fix.
…into test-randomactivation
No the chance on a false negative is one in 12! instead of 4! (40 million instead of 24)
rht
reviewed
Jan 26, 2024
| @@ -54,6 +54,7 @@ def __init__(self, unique_id: int, model: Model) -> None: | |||
| except AttributeError: | |||
| # model super has not been called | |||
| self.model.agents_ = defaultdict(dict) | |||
| self.model.agents_[type(self)][self] = None | |||
There was a problem hiding this comment.
This fix is out of the scope of this PR, but it's fine to add it here.
There was a problem hiding this comment.
Yeah preferably this would have been a seperate PR.
rht
approved these changes
Jan 26, 2024
EwoutH
added a commit
that referenced
this pull request
Jan 26, 2024
…2007) * tests: Add test to check if RandomActivation is not sequential Adds a test that checks if the RandomActivation doesn't trigger agents in a sequential order. In theory this could give false positives (a test passing when it shouldn't, but that chance is around ~0.1^18). * fix for RandomActivation bug fixes #2006 * add agentset.shuffle unittest * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * ruff fix * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add agent to model.agent correctly even if super was not called fix for the second issue in #2006. If super is not present, we create the data structure but forget to add the agent to it. This is just a backward compatibility fix. * test: Shuffle more agents to prevent false negatives No the chance on a false negative is one in 12! instead of 4! (40 million instead of 24) --------- Co-authored-by: Jan Kwakkel <j.h.kwakkel@tudelft.nl> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
EwoutH
added a commit
that referenced
this pull request
Jan 26, 2024
…2007) * tests: Add test to check if RandomActivation is not sequential Adds a test that checks if the RandomActivation doesn't trigger agents in a sequential order. In theory this could give false positives (a test passing when it shouldn't, but that chance is around ~0.1^18). * fix for RandomActivation bug fixes #2006 * add agentset.shuffle unittest * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * ruff fix * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add agent to model.agent correctly even if super was not called fix for the second issue in #2006. If super is not present, we create the data structure but forget to add the agent to it. This is just a backward compatibility fix. * test: Shuffle more agents to prevent false negatives No the chance on a false negative is one in 12! instead of 4! (40 million instead of 24) --------- Co-authored-by: Jan Kwakkel <j.h.kwakkel@tudelft.nl> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
|
I was torn. It was raised in the same issue, and it was just 1 line. |
This was referenced Jan 26, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Adds a test that checks if the
RandomActivationdoesn't trigger agents in a sequential order. This test now fails, because as noted in #2006, it currently does do it in sequential order.In theory this could give false positives (a test passing when it shouldn't, but that chance is around ~0.1^18).