Long install time from windows.

[T4rk1n]

When pip install plotly from a windows machine, it takes anywhere from a minute to 5 minutes. From linux it took no more than 10 seconds.

[jonmmease]

Thanks for the report. I've only tried the windows install on a VM so I wasn't sure what the real difference was. I'm wondering if it's related to compiling the *.pyc files, similar to the conda slowness reported in #1175. Though I don't know why this would be so much slower on windows.

In that issue @fcollonval reported that the conda install from our plotly anaconda channel (which includes precompiled *.pyc files) took less than 20 seconds.

[DarkGhost7]

taking me forever to install on windows 10 with pip install dash, seems like plotly is taking up most of the time?

[Shaban0818]

Same in my case. Windows 10 and plotly taking forever to install.

[Limits0x]

I've been trying to install for about 30 mins. It still hasn't finished. Should I cancel ?
I am on Windows 10 Python 3.8.

[nicolaskruchten]

That does seem like a long time... did it eventually succeed?

[Limits0x]

No I had to cancel it after 1 and 1/2 hours.

[nitrocalcite]

I also experienced this on Windows 10, Python 3.6.10, and after ~5 minutes I canceled it. However, I didn't get any error messages & plotly at least appears to be working fine; perhaps it's getting hung somewhere?

[parmanchaddha]

Same thing here. Using poetry on my mac, running poetry add plotly on an empty environment took about 30 seconds. Executing the same command on my windows 10 pro machine took about 5 minutes...

[Limits0x]

It was probably an issue with my python installation. I reinstalled and now it takes about 5 mins.

[Prajval-1608]

Same here, windows 10 and python 3.7.9. Waiting for plotly to install since more than 5 minutes.

[beingandbrian]

I am also having this issue. Is there a solve for this?

[florian6973]

Same issue on Windows 11

[adridem]

Same issue, stuck on Windows 11

[eroller]

plotly install/update is pretty slow on WSL too. Slower than all the other packages I use. Is there a good justification for it?

[ahmedwaqas92]

Installation was getting stuck on Windows 11 for me as well, I aborted the installation and tried the following:

pip install plotly -vvv

installed pretty much within 30 seconds.

[nick-st]

pip install plotly took ages, but
pip install plotly==5.14.1 is fine

[okoye-dev]

pip install dash is taking almost 30minutes here and it's still now done.
Is this just because it's a windows machine? Windows 11 btw.

[nezza]

We had a lot of Windows user reports for our library reporting installation times as very slow. After some investigation we found:

pip3 install --no-compile plotly 

Takes around 7 seconds

pip3 install plotly

Takes multiple minutes

pip3 install plotly 

With having Real-time protection disabled - takes ~20 seconds

So it looks like it's an antivirus issue

[bmaranville]

I love the plotly package - it does things that no other package does (offers a compact, stable JSON serialization of a plot that can be generated using the public API and passed around between program components and plotted with javascript).

It has nearly 30000 files in the installed python package though, which is why the antivirus slows down installation (it is looking at each one). By comparison matplotlib has between 500 and 1000, depending on how it is installed. >26000 of the files in plotly are in the /validators subdirectory.

If there is a way to reduce the number of files, I'd be happy to help. This would cut down the installation time directly, particularly on Windows (though it affects the installation time on all platforms). Refactoring so that validators are grouped into larger files instead of a lot of small files could reduce the total file count by >80%

[gvwilson]

@bmaranville help consolidating the validators would be very welcome. our next plotly.py release (6.1) reduces the bundle size using some other tricks, but there wasn't time to do what you're suggesting - if you have time to chat online this week or next I'd be happy to walk through options. cheers - @gvwilson