Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Don't recompile the regexes every Note::new/Note::write_to_db #15

Open
wants to merge 2 commits into
base: master
Choose a base branch
from

Conversation

jacksonriley
Copy link

@jacksonriley jacksonriley commented Oct 29, 2023

Hi there - first of all thank you for making this crate, it's super useful!

Motivation

I noticed when profiling my application that quite a chunk of time was being spent in Note::new (I end up creating 193 notes in this example):

jacksonriley@192 ~/zw_to_anki (main)> hyperfine -- "./target/release/zw_to_anki -f E08_417.txt -o test.apkg"
Benchmark #1: ./target/release/zw_to_anki -f E08_417.txt -o test.apkg
  Time (mean ± σ):      4.450 s ±  0.176 s    [User: 4.002 s, System: 0.216 s]
  Range (min … max):    4.254 s …  4.867 s    10 runs

image

This looks to be mainly due to compiling the "sentinel regex" for every field every time:

Regex::new(&format!(

Since the fields never change, I think we can pre-compile the regexes to save this time.

Benchmark post-change

With this change:

jacksonriley@192 ~/zw_to_anki (main)> hyperfine -- "./target/release/zw_to_anki -f E08_417.txt -o test.apkg"
Benchmark #1: ./target/release/zw_to_anki -f E08_417.txt -o test.apkg
  Time (mean ± σ):      1.978 s ±  0.223 s    [User: 1.628 s, System: 0.190 s]
  Range (min … max):    1.820 s …  2.573 s    10 runs

image

Which comes out as making Note::new go from taking 12.6ms per-call to 0.1ms.

Adding some once-cells

I then noticed that there were some more regexes getting compiled every Note::write_to_db, so have made these only get compiled once using once_cell.

This then cuts the time down further:

jacksonriley@192 ~/zw_to_anki (main)> hyperfine -- "./target/release/zw_to_anki -f E08_417.txt -o test.apkg"
Benchmark #1: ./target/release/zw_to_anki -f E08_417.txt -o test.apkg
  Time (mean ± σ):      1.613 s ±  0.051 s    [User: 1.374 s, System: 0.186 s]
  Range (min … max):    1.548 s …  1.737 s    10 runs

and now genanki-rs is barely visible anymore on the flamegraph!
image

Implementation

I think that what I've done is reasonable - fine to add a new private field to Model and the implementation seems to work. Happy to change up the implementation however you see fit though!

@jacksonriley jacksonriley changed the title Don't recompile the sentinel regexes every Note::new Don't recompile the sentinel regexes every Note::new Oct 29, 2023
@jacksonriley jacksonriley changed the title Don't recompile the sentinel regexes every Note::new Don't recompile the regexes every Note::new/Note::write_to_db Oct 29, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant