Documentation: rename benchmark django-restful-framework -> django-rest-framework

[frankie567]

Change Summary

I came across this while reading the benchmark page in the docs: benchmark for django-rest-framework was named django-restful-framework. This confused me a bit.

I changed the package name in the benchmark script and run the script instead of modifying directly the generated MD file, as specified.

Problem is that, on my machine, the benchmark is not in favor of pydantic 😟So I don't know what you prefer for this (I would be okay to just edit the previous results).

Related issue number

Nope.

Checklist

• Unit tests for the changes exist
• Tests pass on CI and coverage remains at 100%
• Documentation reflects the changes where applicable
• changes/<pull request or issue id>-<github username>.md file added describing change
  (see changes/README.md for details)

[codecov]

Codecov Report

Merging #1119 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1119   +/-   ##
=======================================
  Coverage   99.97%   99.97%           
=======================================
  Files          20       20           
  Lines        3395     3395           
  Branches      664      664           
=======================================
  Hits         3394     3394           
  Misses          1        1

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a7ef2d5...288b784. Read the comment docs.

[dmontagu]

Did you cythonize before running the benchmark locally? That will make pydantic about 2x faster.

[samuelcolvin]

Also, I always use the same cases.json for consistency, you can download it from here.

---

I know that .benchmarks_table.md says "do not edit", but I think here for simplicity you can just do a find a replace on "django-restful-framework" and thereby edit both files without regenerating benchmarks.

[frankie567]

@dmontagu I didn't, and TBH I never did that. Shouldn't it be mentioned in the benchmarks page? It just says "run with Python 3.7.4". That doesn't seem very fair regarding the other libs.

@samuelcolvin Ok, I'll do that 👍

[samuelcolvin]

@dmontagu I didn't, and TBH I never did that. Shouldn't it be mentioned in the benchmarks page? It just says "run with Python 3.7.4". That doesn't seem very fair regarding the other libs.

The benchmarks say:

Benchmarks were run with Python 3.7.4 and the package versions listed above installed via pypi on Ubuntu 18.04.

I think that's the fair way of running benchmarks - with the library versions distributed via pypi.

Pydantic provides binaries compiled with cython from python which is modified in numerous ways to be compilable and as fast as possible when compiled - that's one of the reasons it's fast. If other libraries provided similar binaries for linux, they would be used too; some may already be, I haven't checked.

[frankie567]

@samuelcolvin Thank you, I understand better now. So, since I ran the benchmark with a pydantic version pulled from PyPi (not the one on my local machine), I guess then it actually ran with the Cython optimizations?

[samuelcolvin]

Only if you're on linux, you can checked via print(pydantic.compiled).

[frankie567]

I didn't then, i was on macOS. Thank you for providing those details 👍

[samuelcolvin]

thanks.