Will try thank you! 
Congrats on the release guys! 
Q: do we need to go through these same steps now that “custom components” have made it to the official release?
1 Like
Just doing a pip install streamlit-pandas-profiling and using it in your app should work directly now 
Phoawwwww! Guess what I’m about to try as soon as I get home?! 
2 Likes
Still facing the pyarrow error and can’t seem to avoid it
Failed to build pyarrow
ERROR: Could not build wheels for pyarrow which use PEP 517 and cannot be installed directly
I may try on another machine later this week to see if the issue persists.
Thanks,
Charly
Hmm weird. Could you try to perform those steps one by one in a new python environment?
-
pip install --upgrade pip: pip is most often not up to date when creating a new env. -
pip install --upgrade setuptools wheel: it’ll install/upgrade wheel to avoid building packages from source -
pip install pyarrow: it’ll check if pyarrow install correctly pip install streamlit streamlit-pandas-profiling
If it still fails at step 3, save and post the whole traceback, especially before the “Failed to build pyarrow” if you encounter it again.
1 Like
Thanks Synode!
I’ve retried all steps from #1 to #3, still failing.
Here’s the traceback: (venv) C:\Users\Charly\Desktop\Profiling NEW 4>pip install pyarrowCollecting p - Pastebin.com
Possibly something to do with runtime error (my laptop is 64 bits).
RuntimeError: Not supported on 32-bit Windows
Thanks,
Charly
Apparently you have a 32 bit version of Python. Could you try to upgrade it to a 64 bit version?
Otherwise, using conda to install pandas-profiling should work as it won’t compile anything.
Wow, THAT was the issue! Thank you SO much Synode!
1 Like
Aha! Glad it finally worked 
1 Like
Hi @okld.
Great Work.
I don’t understand why you can’t get the app bar working in an iframe . It works for me in my implementation at awesome-panel.org.
You might find inspiration to solve the problem here https://github.com/MarcSkovMadsen/awesome-panel-extensions/blob/master/awesome_panel_extensions/pane/pandas_profile_report.py
Or test that it works on binder https://mybinder.org/v2/gh/MarcSkovMadsen/awesome-panel-extensions/master?filepath=examples%2Freference_gallery%2Fpanes%2FPandasProfileReport.ipynb
2 Likes
Hey @Marc,
Thank you for your reply!
In your Panel implementation when you click a button, it scrolls your iframe. In my case, I purposely resize the profile report to have no iframe scrollbar. In consequence the user don’t have to define a height and he gets a seamless component, but this means that buttons need to scroll the parent page to reach an anchor contained inside the report’s iframe (which is contained inside a sandboxed streamlit iframe), and that doesn’t seem to be possible.
That said, even if I had a fixed height and used scrolling instead like your version, because of streamlit’s iframes sandbox, I think I’d run into other kinds of issues unfortunately.
PS: Great work on awesome-panel.org and awesome-streamlit.org btw!
2 Likes
Hi Synode! 
I’ve finally had a proper try - Great integration, bravo! 
Any chance some of the profiling data can be saved, ideally onto raw csv?
Thanks,
Charly
I’m trying to use this with streamlit sharing but st_profile_report won’t display. When I run this locally, everything seems to be okay.
See here: https://share.streamlit.io/felixdasilva/streamlitpandasprofiling/main/app.py
What am I doing wrong that is not displaying the report? this is my first app to share and trying it. Thanks!
Hi Synode!
I’m wondering, is it possible to edit/format elements in the Pandas Profiling module, e.g remove the ‘overview’ bit:
Thanks,
Charly
Hello @felixdasilva, welcome to the forum!
This is weird indeed. Gonna check this out
Hello @Charly_Wargnier
I didn’t see anything to do that unfortunately. I may have missed it though.
I know you can customize many options when generating a report, but styling doesn’t seem to let you customize that part.
1 Like
Thanks for looking into it. I don’t know why it won’t display when I deploy it on streamlit sharing when locally it runs perfectly fine.
Right now, I’ve made it so that it doesn’t display the report and the user just downloads it for now.
Again, appreciate the work and the time you’ve put into this component.
Noticed that. I’ve forked your repo, I’ll do tests on my side.
And I may also release a new reworked pandas profiling version. Stay tuned!
Thanks Synode! Very interesting!
Do you reckon parameter may be amneddable right from the Streamlit app?
At first glance, I’d assume so but because these are in a yaml file and not Py file, I may have to test it
Charly