Failed to import ROOT in jupyter notebook in Scientific Linux 7

ROOT
1

Hi, I have the exact same problem as described in this post, except I use Scientific Linux 7 and I compiled ROOT 6.28.6 from source for our computer with G++ (GCC) 11.2.0 and

cmake -DCMAKE_INSTALL_PREFIX="/cern_root_6.28.6_cpp17" -DCMAKE_C_COMPILER=/usr/local/bin/gcc -DCMAKE_CXX_COMPILER=/usr/local/bin/g++ -DCMAKE_LINKER=/usr/local/bin/g++ -DCMAKE_CXX_STANDARD=17 -DPython3_EXECUTABLE=$(which python3) -DPython2_EXECUTABLE=$(which python2) ../root-6.28.06 
sudo cmake --build . -j16 --target install

The problem is, that I don’t really understand what the answer means in my case. What can I do to make this work?

Thanks in advance!
( @eguiraud perhaps?)

The exact traceback is this:

Welcome to JupyROOT 6.28/06
Error in callback <bound method CaptureDrawnPrimitives._post_execute of <JupyROOT.helpers.utils.CaptureDrawnPrimitives object at 0x7f01c49e85c0>> (for post_execute):
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/cern_root_6.28.6_cpp17/lib/JupyROOT/helpers/utils.py in _post_execute(self)
    456 
    457     def _post_execute(self):
--> 458         NotebookDraw()
    459 
    460     def register(self):

/cern_root_6.28.6_cpp17/lib/JupyROOT/helpers/utils.py in NotebookDraw()
    446     DrawGeometry()
    447     DrawCanvases()
--> 448     DrawRCanvases()
    449 
    450 class CaptureDrawnPrimitives(object):

/cern_root_6.28.6_cpp17/lib/JupyROOT/helpers/utils.py in DrawRCanvases()
    439 
    440 def DrawRCanvases():
--> 441     rdrawers = GetRCanvasDrawers()
    442     for drawer in rdrawers:
    443         drawer.Draw()

/cern_root_6.28.6_cpp17/lib/JupyROOT/helpers/utils.py in GetRCanvasDrawers()
    411 def GetRCanvasDrawers():
    412     if not RCanvasAvailable(): return []
--> 413     lOfC = ROOT.Experimental.RCanvas.GetCanvases()
    414     return [NotebookDrawer(can.__smartptr__().get()) for can in lOfC if can.IsShown()]
    415 

ValueError: static const vector<shared_ptr<ROOT::Experimental::RCanvas> > ROOT::Experimental::RCanvas::GetCanvases() =>
    ValueError: nullptr result where temporary expected
2

Hi,

Can you start normal ROOT session and try:

root [0] auto lst = ROOT::Experimental::RCanvas::GetCanvases()

Regards,
Sergey

3

Thank you for the reply Sergey. This is my output:

$ root
   ------------------------------------------------------------------
  | Welcome to ROOT 6.28/06                        https://root.cern |
  | (c) 1995-2023, The ROOT Team; conception: R. Brun, F. Rademakers |
  | Built for linuxx8664gcc on Aug 28 2023, 11:29:15                 |
  | From tags/v6-28-06@v6-28-06                                      |
  | With g++ (GCC) 11.2.0                                            |
  | Try '.help'/'.?', '.demo', '.license', '.credits', '.quit'/'.q'  |
   ------------------------------------------------------------------

root [0] auto lst = ROOT::Experimental::RCanvas::GetCanvases()
(std::vector<std::shared_ptr<ROOT::Experimental::RCanvas>, std::allocator<std::shared_ptr<ROOT::Experimental::RCanvas> > > &) {}
root [1]
4

Means this should work.

Do you have in $PATH some other ROOT installations?
Do you have system-wide ROOT installed?

5

I have no system-wide installations, but I have older versions on the computer, yes. We change the versions by sourcing the appropriate thisroot.sh
According to the error python picks up the proper version in the notebook too :frowning:
I checked in the notebook with:

%%bash
echo $PATH

and the correct one appears to be the only root directory in the $PATH.

6

Calling the test above in a simple python terminal also works:

$ python3
Python 3.6.8 (default, Nov  9 2023, 05:22:21) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-44)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import ROOT
>>> ROOT.Experimental.RCanvas.GetCanvases()
vector<shared_ptr<ROOT::Experimental::RCanvas> >{}
>>>
7

I am not python expert, but may be python3.6 is too old? @eguiraud @vpadulan?

8

Currently we support all versions of Python 3, so that should not be the direct cause for the problem seen by the user. The next thing I would try is to make sure that the Python executable used in the ROOT installation is the same Python executable that launches the jupyter notebook.

Cheers,
Vincenzo

9

Thank you for the many answers so far!

In the notebook:

import sys
print(sys.version)
print(sys.executable)

returns

3.6.8 (default, Nov  9 2023, 05:22:21) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-44)]
/usr/bin/python3

importing ROOT still meets the same show-stopper error.

In my terminal calling this python:

$ /usr/bin/python3
Python 3.6.8 (default, Nov  9 2023, 05:22:21) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-44)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import ROOT
>>> lOfC = ROOT.Experimental.RCanvas.GetCanvases()
>>> print(lOfC)
{}
>>>
10

As a temporary fix I modified

/cern_root_6.28.6_cpp17/lib/JupyROOT/helpers/utils.py

as

def GetRCanvasDrawers():
    return []

The module now imports and I can read root files and operate on histograms, which was the most important part of the whole thing for me.
However, when I try to do a plot with ROOT the script that works in the interactive python shell gives an error. Maybe this hints at something? The simple script:

h = ROOT.TH1F("gauss","Example histogram",100,-4,4)
h.FillRandom("gaus")
c = ROOT.TCanvas("myCanvasName","The Canvas Title",800,600)
h.Draw()
c.Draw()

The error I am getting in the notebook is:

IncrementalExecutor::executeFunction: symbol '_ZN5cling7runtime6gClingE' unresolved while linking [cling interface function]!
You are probably missing the definition of cling::runtime::gCling
Maybe you need to load the corresponding shared library?
Error in <TH1F::FillRandom>: Unknown function: gaus
IncrementalExecutor::executeFunction: symbol '_ZN5cling7runtime6gClingE' unresolved while linking [cling interface function]!
You are probably missing the definition of cling::runtime::gCling
Maybe you need to load the corresponding shared library?

If in the notebook I echo my path I get

/usr/bin:/home/pmajor/.local/bin:/home/pmajor/.vscode-server/bin/1a5daa3a0231a0fbba4f14db7ec463cf99d7768e/bin/remote-cli:/cern_root_6.28.6_cpp17/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/home/pmajor/.local/bin:/home/pmajor/bin:/home/pmajor/.local/bin:/home/pmajor/bin:/home/pmajor/.vscode-server/bin/1a5daa3a0231a0fbba4f14db7ec463cf99d7768e/bin/remote-cli:/cern_root_6.28.6_cpp17/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/home/pmajor/.local/bin:/home/pmajor/bin:/home/pmajor/.local/bin:/home/pmajor/bin
11

Hi,

You trying to fix symptoms, but not the problem itself.
Something wrong with jupyter and pyROOT.

From your cmake command:

sudo cmake --build . -j16 --target install

Are you really install ROOT libraries in system directories?
Can you try without installing it?

Regards,
Sergey

12

I dont think anything goes into the system directories. I have -DCMAKE_INSTALL_PREFIX="/cern_root_6.28.6_cpp17" so install should just copy the build results here, right?

13

Can you try without install prefix and without --install command?

Just to be sure that this is not kind of broken configuration of jupyter.

14

Yes, sure. So like this?

cmake -DCMAKE_C_COMPILER=/usr/local/bin/gcc -DCMAKE_CXX_COMPILER=/usr/local/bin/g++ -DCMAKE_LINKER=/usr/local/bin/g++ -DCMAKE_CXX_STANDARD=17 -DPython3_EXECUTABLE=$(which python3) -DPython2_EXECUTABLE=$(which python2) ../root-6.28.06 
sudo cmake --build . -j16
15

Yes, try such configuration.
And do not use sudo.