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

Skip folded deeper levels when rendering page tree #1324

Merged
merged 1 commit into from
Nov 24, 2017

Conversation

pascalj
Copy link
Contributor

@pascalj pascalj commented Nov 16, 2017

This change is a performance improvement and should not affect the behavior of the CMS.

Previously the page tree was rendered in full even if pages are folded.
Consider the following tree: even if level 1 was folded, level 2 and 3
would get rendered and sent to the client.

- root
 |+ level 1
   |- level 2
     |- level 3

This change does two things:

  1. load all user-folded pages for the current user in advance, reducing
    the number of queries to 1
  2. skips all pages that are under a folded page

The result is:

- root
 |+ level 1

Together with 5a4c0d3 this speeds up the page tree rendering from 20
seconds to ~100ms when only 10 folded pages are shown (for Alchemy::Page.count == 4000). The number of queries per page tree is also constant now.

Feedback is welcome! 👂

Copy link
Member

@tvdeyen tvdeyen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Makes sense! Thank you, Pascal.

I tried this locally with my "Seed 1000 Alchemy Pages" gist and did not saw any noticeable improvements, but this might be because of running this locally with sqlite and in dev mode.

As I did not found any regressions, 👍

Could you add a simple model spec for the newly introduced scope? Thanks

Previously the page tree was rendered in full even if pages are folded.
Consider the following tree: even if level 1 was folded, level 2 and 3
would get rendered and sent to the client.

- root
 |+ level 1
   |- level 2
     |- level 3

This change does two things:

1. load all user-folded pages for the current user in advance, reducing
the number of queries to 1
2. skips all pages that are under a folded page

The result is:

- root
 |+ level 1

Together with 5a4c0d3 this speeds up the page tree rendering from 20
seconds to ~100ms when only 10 folded pages are shown
(Alchemy::Page.count == 4000).
@pascalj pascalj force-pushed the pagetreeserializer-speedup branch from f7eae6e to ac564d7 Compare November 19, 2017 20:26
@pascalj
Copy link
Contributor Author

pascalj commented Nov 19, 2017

Thanks for the gist! Yes, this certainly has an influence. I have no time to dig deeper into this, but it seems Postgres has certain problems with this part of the code or maybe it's my 3.5 branch (although including 5a4c0d3). I tried it with roughly 3k pages (5 levels with 5 sub pages each) and it outperforms my Postgres testsetup. However: both are faster with this patch:

spec/dummy with 10 * 100 pages, first level folded:
master: Completed 200 OK in 1383ms (Views: 1355.2ms | ActiveRecord: 19.0ms) (~350KB)
this branch: Completed 200 OK in 244ms (Views: 221.6ms | ActiveRecord: 13.5ms) (~60KB)

Also notice the difference in size. In the 3.5k pages example this goes down from 600kb to 4KB.

I added a model spec. The test for inheritance doesn't seem to be necessary any more. I copied it from the old implementation, but I'm not brave enough to remove it and added a test to verify it doesn't fail for non-AR user classes.

Copy link
Member

@tvdeyen tvdeyen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the specs and additional benchmarks.

@tvdeyen tvdeyen merged commit 01f3497 into AlchemyCMS:master Nov 24, 2017
@tvdeyen
Copy link
Member

tvdeyen commented Nov 24, 2017

Danke Pascal und lieben Gruß

@creativej
Copy link

Great work! What's the timeline for the 4.1 release?
Any chance for this patch to go into 4.0?
We have a lot of pages in our cms it takes over 5s to load the page tree so getting this improvement would be amazing!

@tvdeyen tvdeyen added this to the 4.1 milestone Sep 10, 2018
@tvdeyen tvdeyen mentioned this pull request Sep 10, 2018
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.

3 participants