⚡️ Speed up is_pydantic_dataclass() by 41% in pydantic/dataclasses.py

[misrasaurabh1]

Change Summary

📄 is_pydantic_dataclass() in pydantic/dataclasses.py

📈 Performance improved by 41% (0.41x faster)

⏱️ Runtime went down from 4.80 microseconds to 3.40 microseconds

Explanation and details

Here's the optimized version of the provided program.

Explanation.

1. Short-circuit Evaluation: I reordered the conditions in the return statement to utilize short-circuit evaluation. The check for '__pydantic_validator__' in class_.__dict__ is less expensive and will fail fast if class_ is not a valid type with this attribute, reducing unnecessary calls to the relatively more expensive dataclasses.is_dataclass function.

2. Error Handling: Added a try-except block to quickly handle the scenario where class_ may not have the __dict__ attribute (__dict__ is not present for some objects that do not have attribute dictionaries). This prevents the function from throwing an attribute error and failing incompletely.

This should make the function slightly faster, especially when dealing with objects that early out on the __dict__ check.

Correctness verification

The new optimized code was tested for correctness. The results are listed below.

✅ 211 Passed − ⚙️ Existing Unit Tests

(click to show existing tests)

- test_dataclasses.py

Checklist

• The pull request title is a good summary of the changes - it will be used in the changelog
• Unit tests for the changes exist
• Tests pass on CI
• Documentation reflects the changes where applicable
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

[codspeed-hq]

CodSpeed Performance Report

Merging #9652 will not alter performance

_{Comparing misrasaurabh1:codeflash/optimize-is_pydantic_dataclass-2024-06-04T21.41.42 (0bf4c9f) with main (6b0e7de)}

Summary

✅ 13 untouched benchmarks

[sydney-runkle]

@misrasaurabh1,

How about:

if hasattr(class_, '__dict__'):
    return '__pydantic_validator__' in class_.__dict__ and dataclasses.is_dataclass(class_)
return False

What are the performance changes like with that approach?

[misrasaurabh1]

Great suggestion, when I tested your recommendation with codeflash on my laptop, I got a 23% speedup. But when I try the try/except approach (the one that codeflash originally recommended), I got a 36% speedup. So the try/except approach seems to be faster.

[sydney-runkle]

Great suggestion, when I tested your recommendation with codeflash on my laptop, I got a 23% speedup. But when I try the try/except approach (the one that codeflash originally recommended), I got a 36% speedup. So the try/except approach seems to be faster.

Awesome, thanks for looking into this. We're happy with the existing change, then!

I'll note, this isn't a super critical function to speed up, but we'll take the performance benefit where we can get it :). We're excited to collaborate more on more critical functions!

[sydney-runkle]

Looks great! Thanks folks!