Memoize manual section presenter #3516
Merged
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
We're currently calling the main method, and then looping over the
result. In each loop, we call the method again so that we can get the
length. See manual_section.html.erb, which does roughly:
@content_item.main.each.with_index do |item, index|
# ...
{
data_attributes: {
ga4_event: {
# ...
index: index,
index_total: @content_item.main.length,
... the effect of this is that we run the main method n times, where n
is the number of <h2> elements in the body. This results in us spending
a lot of time parsing the same document and running the same CSS / XPath
selectors on it, which adds up to a lot of time.
Memoizing the result of the main method shaves several seconds off the
time to render www.gov.uk/guidance/complete-the-school-census/data-items-2024-to-2025
on my machine.
I did look at some other optimizations (e.g. traversing the nokogiri
document manually, instead of using css / xpaths), but once it's
memoized those only save microseconds so it's not worth it.
This saves about 70ms re-parsing the body, as the intro method is called twice.
returning early from a begin block won't assign the variable, but it sort of looks like it will. Clearer / more explicit to have the early returns outside the block.
richardTowers
force-pushed
the
faster-manuals
branch
from
6a27e17 to
5e06f8b
Compare
hannako
approved these changes
Jan 15, 2025
There was a problem hiding this comment.
How do you do that rack-mini-profiler flame diagram thing? I smell a tech fortnightly presentation!!
Maybe we could also help by adding
length = @content_item.main.length
here in the template. Also maybe some of that template logic should be in the presenter? One for another time!
|
Definitely a tech fortnightly talk coming yeah. It's pretty easy though actually:
Previously I've used rbspy for these things, which is way more faff to set up. |
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.
We're currently calling the main method, and then looping over the result. In each loop, we call the method again so that we can get the length. See manual_section.html.erb, which does roughly:
... the effect of this is that we run the main method n times, where n is the number of
<h2>elements in the body. This results in us spending a lot of time parsing the same document and running the same CSS / XPath selectors on it, which adds up to a lot of time.Memoizing the result of the main method shaves several seconds off the time to render www.gov.uk/guidance/complete-the-school-census/data-items-2024-to-2025 on my machine.
The flamegraph before these changes (using the wonderful rack-mini-profiler) shows manual_section.html.erb taking over 2 seconds to render:
Note that
ManualSectionPresenter#mainis called many times.After the changes in this PR, this reduces to 186ms:
These flamegraphs also identified a further ~70ms we could save by memoizing the intro method, so I've done that too.
I did look at some other optimizations (e.g. traversing the nokogiri document manually, instead of using css / xpaths), but once it's memoized those only save microseconds so it's not worth it.