Skip to content

WEB: JupyterLite terminal too slow #47530

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

Closed
datapythonista opened this issue Jun 28, 2022 · 9 comments · Fixed by #49807
Closed

WEB: JupyterLite terminal too slow #47530

datapythonista opened this issue Jun 28, 2022 · 9 comments · Fixed by #49807
Labels
Web pandas website

Comments

@datapythonista
Copy link
Member

The Try it in your browser terminal in the https://pandas.pydata.org/getting_started.html page, seems to be too slow to me. I've got 10Mb download, and an i7 processor. Nothing too fancy, but quite reasonable to expect things to load not too slow. For me the terminal loads after 20 seconds, and cells are executed and the terminal is ready to run commands 4 minutes after the terminal is loaded.

Can someone else have a look and see if the times are similar? @jtpio is this what it needs to take? If this is what everybody is experiencing, and there is nothing we can do about it, I think we should remove the terminal from the website, until things can be faster.

@datapythonista datapythonista added the Web pandas website label Jun 28, 2022
@jtpio
Copy link
Contributor

jtpio commented Jun 28, 2022

Thanks @datapythonista for reporting 👍

4 minutes sounds like a long time. Normally it should be within 30-60s, but can depend on the internet connection since it's a couple of MBs to download indeed.

After the first download most of the assets should be cached by the browser and make it faster to load.

Will look into it more. In the meantime I opened #47532 to update to the latest jupyterlite release, which should help speed things up a little bit on subsequent reloads.

@datapythonista
Copy link
Member Author

I assumed that the download finishes after 20 seconds, when the terminal is visible, and the problem is running things, which is what takes 4 minutes. Is it still downloading things when the terminal with the code is visible?

@datapythonista
Copy link
Member Author

I tried today with a faster connection, 80Mbps, and displaying the terminal took shorter, a bit less than ten seconds. But running the code took more than one minute.

I did a test to run print(1) instead of the pandas code, and the loading time is the same as expected, but running the print takes like 15 or 20 seconds.

I tried the numpy terminal. It loads much faster, like in 3 seconds. I'm using a new incognito session every time, I think caches shouldn't be used. But running the import numpy takes like 1m30s.

I'm using Brave on Linux. Would be good if someone else can give it a try, to see if those are the times everybody is experiencing.

@rhshadrach
Copy link
Member

I got 23 seconds from incognito, 10 seconds when just refreshing the page. I'm using Chrome on Ubuntu.

If the experience isn't too bad for most users, I'm +1 on it remaining. In fact, I just used it while working on a PR to check 1.4.2 behavior and it was quite convenient.

@datapythonista
Copy link
Member Author

That sounds more reasonable. Was it 23 seconds until the terminal finished the execution of the cells? What Browser are you using?

@rhshadrach
Copy link
Member

I'm using Chrome on Ubuntu. That's correct; the terminal finished execution of the cells that are there when the page loads as well as one I provided (1 + 1)

@rhshadrach
Copy link
Member

I recently tried this again, it seems to be hanging now. Even after 30 minutes, I don't get output from any commands in the terminal.

@datapythonista
Copy link
Member Author

I tried two days ago from mobile and I couldn't get the default commands executed after couple of minutes. I thought it was a mobile only problem.

@jtpio maybe a problem with the last upgrade?

@jtpio
Copy link
Contributor

jtpio commented Sep 26, 2022

Ah that looks odd and should have been fixed by #48329.

I'll have another look to see what's happening.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Web pandas website
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants