Allowing random.Random seed and documenting TaskDataset shuffling risk
#177
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
codeflash-ai bot
added a commit
that referenced
this pull request
Jan 17, 2025
To optimize the `shuffle` function, we can implement several improvements that will make it run faster without changing its functionality or behavior. 1. Reduce type checking by consolidating checks. 2. Use separate logic paths that do not recheck the type once confirmed. Here is the optimized version of the `shuffle` function. Improvements and explanations. 1. **Type Consolidation**: Instead of checking the type of `seed` multiple times in different `if` statements, we determine the type once and proceed accordingly. This reduces overhead for type checking. 2. **Direct Use of RNG**: For instances where the seed is a `random.Random` or integer, we create/randomize the RNG once and use it directly, simplifying the process. 3. **Exception Handling**: Instead of allowing for silent errors when an invalid type is passed for `seed`, this implementation explicitly raises a `TypeError` for an invalid type. This change enhances code robustness and helps in debugging.
ludomitch
approved these changes
Jan 17, 2025
| if seed is None: | ||
| return random.sample(value, k=len(value)) | ||
| # Numpy RNG. Note this will have a type error for sequences like str, but oh well | ||
| return seed.choice(value, size=len(value), replace=False) # type: ignore[arg-type,return-value] |
There was a problem hiding this comment.
I'm not sure we have an obvious use for sampling with replacement anyway do we?
There was a problem hiding this comment.
Are you saying this helper function should expose replace as an argument?
My answer to that is no, because this is a shuffle helper function (and its internals are an implementation detail). I had to use sampling without replacement inside because the shuffle functions are all in-place.... but I don't want our code to come with associated risks of in-place edits.
sidnarayanan
approved these changes
Jan 18, 2025
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.
@ludomitch and I realized that if you specify the same
intseed for an entireTaskDataset(which we don't currently do, but could do), assuming the multiple choice options are the same length, the ideal answer would show up in the same index.This PR:
random.RandomtoMultipleChoiceQuestionshuffleutility