This page compiles a list of common errors that can appear when using Scipion.
- Installing Scipion/Xmipp from precompiled bundles
- Fixing libjbig.so.0 not found in OpenSUSE42.3
- Fixing fonts in Ubuntu 18
- Launching Eman boxer protocol
- Relion3 compilation with CUDA8.0 and g++>6
- Compiling Scipion with OpenCV
- Scipion freezes after click on open bibtex
- Compiling Scipion in Opensuse
- Endless list of CUDA related errors
- Requirement already satisfied
- ImportError: cannot import name HTTPSHandler
- Launching XMIPP3 CL2D protocol
- ImportError: libgfortran.so.3
- bigtiff in Claudio
- Install Xmipp3 in Diocletian
- scikit-learn installation fails
- sh_alignment installation fails
- Deep consensus fail due to index out of run.
We have checked the following Operative Systems for the installation using the precompiled bundles:
- Centos7: works with Scipion_Centos and Xmipp_Centos
- Ubuntu16: works with Scipion_Ubuntu and Xmipp_Debian
- Ubuntu18: workks with Scipion_Ubuntu and Xmipp_Debian
- OpenSUSE42.3: works with Scipion_Ubuntu and Xmipp_Debian
For OpenSUSE15 and Debian, installation from sources is needed, so far.
When Scipion_Ubuntu precompiled bundle is installed, maybe a “libjbig.so.0 not found” error is raised. We have observed that OpenSUSE includes libjbig.so.2 and we have checked that is also valid, thus we propose to link one to the other by
sudo ln -s /usr/lib64/libjbig.so.2 /usr/lib64/libjbig.so.0
The Scipion font is not right in Ubuntu 18. A temporary fix for this is to move all TK and TCL files in software/lib and use the system ones.
If you see an error like ‘Cannot mix incompatible Qt library (version 0x40806) with this library (version 0x40804)’. This means the Qt installed on your computer is conflicting with the Qt distributed with EMAN2. In most cases it gets solved by removing the Qt that comes with EMAN2 from EMAN2DIR/extlib/lib.
If you are getting an error telling that g++ later than 6 is not supported
by nvcc8, you can set a lower g++ compiler in the
CXX=g++-5. To do that you need to have gcc/g++-5 installed.
If you have problems compiling Scipion with OpenCV support (CUDA version >=6.5), e.g. opencv-2.4.9 compilation fails with an error:
Error: target 'software/lib/libopencv_core.so' not built (after running 'make install > /home/user/soft/scipion/software/log/opencv_make_install.log 2>&1')
And log file (software/log/opencv_make.log) shows something like:
[ 9%] Building NVCC (Device) object modules/core/CMakeFiles/cuda_compile.dir/src/cuda/cuda_compile_generated_gpu_mat.cu.o /usr/include/string.h: In function ‘void* __mempcpy_inline(void, const void, size_t)’: /usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope return (char *) memcpy (__dest, __src, __n) + __n; ^ CMake Error at cuda_compile_generated_gpu_mat.cu.o.cmake:264 (message): Error generating file /home/mag/opencv/build_opencv_master/modules/core/CMakeFiles/cuda_compile.dir/src/cuda/./cuda_compile_generated_gpu_mat.cu.o
- Find the micro-architecture name for your GPU card, e.g. Kepler for K40 or Fermi for M2070 card
cmake -DCUDA_GENERATION=Kepler -DWITH_CUDA:BOOL=ON -DCMAKE_INSTALL_PREFIX:PATH=/path/to/scipion/software . > /path/to/scipion/software/log/opencv_cmake.log 2>&1substituting correct path and micro-architecture
- Modify source files in opencv-2.4.9 folder according to this and this fix
scipion install, opencv now should compile cleanly ***
This likely happens because your machine doesn’t have a default program to open bibtex. Type this in your terminal to set gedit as your default program for bibtex files:
xdg-mime default gedit.desktop text/x-bibtex
Scipion instalationin Opensuse sometimes involves a few drawbacks. Once
in the terminal the compilation has been launched,
./scipion install, stop the installation (
Crtl+C). It is
neccesary to change the python version (download python 2.7.13). Copy
the download file to
scipion\software\tmp\ and edit next file
The line in which the python version is specified must be modified by
the downloaded version 2.7.13, it means to substitute the old version
2.7.8 by 2.7.13. Finally we can go to the terminal again and relaunch
the installation by doing
Conditions 1. you had Scipion already installed (from source) 2. later on you installed numpy again (e.g. with pandas) 3. you want to reinstall Scipion (from source)
Building numpy ... python /home/user/Scipion/software/lib/python2.7/site-packages/pip install numpy==1.14.1 Requirement already satisfied: numpy==1.14.1 in /home/user/.local/lib/python2.7/site-packages Error: target '/home/user/Scipion/software/lib/python2.7/site-packages/numpy' not built (after running 'python /home/user/Scipion/software/lib/python2.7/site-packages/pip install numpy==1.14.1')
Numpy version conflict?
uninstall Scipion’s version of numpy
scipion run pip uninstall numpy rm -rf software/lib/python2.7/site-packages/numpy
run install again
scipion install -j 8
Building pip ... python scripts/get-pip.py -I --no-setuptools Traceback (most recent call last): File "scripts/get-pip.py", line 19177, in <module> main() File "scripts/get-pip.py", line 194, in main bootstrap(tmpdir=tmpdir) File "scripts/get-pip.py", line 82, in bootstrap import pip File "/tmp/tmpXJbtSy/pip.zip/pip/__init__.py", line 16, in <module> # * File "/tmp/tmpXJbtSy/pip.zip/pip/vcs/subversion.py", line 9, in <module> File "/tmp/tmpXJbtSy/pip.zip/pip/index.py", line 30, in <module> File "/tmp/tmpXJbtSy/pip.zip/pip/wheel.py", line 39, in <module> File "/tmp/tmpXJbtSy/pip.zip/pip/_vendor/distlib/scripts.py", line 14, in <module> File "/tmp/tmpXJbtSy/pip.zip/pip/_vendor/distlib/compat.py", line 31, in <module> ImportError: cannot import name HTTPSHandler Error: target 'scipion/software/lib/python2.7/site-packages/pip' not built (after running 'python scripts/get-pip.py -I --no-setuptools')
sudo apt-get install libssl-dev rm -rf software/bin/python* software/lib/python2.7/ ./scipion install
Error: libmpi.so - No such file or directory
If executing Xmipp3-cl2d protocol fails with an error:
.../Scipion/Projects/release-1.2.1/scipion/software/em/xmipp/bin/xmipp_mpi_classify_CL2D: error while loading shared libraries: libmpi.so.1: cannot open shared object file: No such file or directory ... ... ... Protocol failed: Command 'mpirun -np 4 -bynode `which xmipp_mpi_classify_CL2D` -i Runs/002697_XmippProtCL2D/tmp/input_particles.xmd --odir Runs/002697_XmippProtCL2D/extra --oroot level --nref 8 --iter 10 --distance correlation --classicalMultiref --nref0 2' returned non-zero exit status 127
This means that the libmpi.so.1 library installed on your computer cannot open.
Create a symbolic link to this library at the location of the libmpi.so library (/usr/lib/ in Ubunut16 or /usr/lib/x86_64-linux-gnu in Ubuntu18).
Assuming that ls /usr/lib/libmpi.so find a file:
ln -s /usr/lib/libmpi.so /usr/lib/libmpi.so.1
We have experimented something similar with libmpi_cxx.so.1
ln -s /usr/lib/libmpi_cxx.so /usr/lib/libmpi_cxx.so.1
This has been reported on an UBUNTU-18 machine using binaries, but may happen at compile time using sources. It was happening when launching scipion. The error reported looked like this:
Traceback (most recent call last): File "/home/xxx/bin/scipion/pyworkflow/apps/pw_manager.py", line 32, in <module> from pyworkflow.gui.project import ProjectManagerWindow File "/home/xxx/bin/scipion/pyworkflow/gui/__init__.py", line 27, in <module> from gui import * File "/home/xxx/bin/scipion/pyworkflow/gui/gui.py", line 34, in <module> from pyworkflow.utils.properties import Message, Color, Icon File "/home/xxx/bin/scipion/pyworkflow/utils/__init__.py", line 30, in <module> from utils import * File "/home/xxx/bin/scipion/pyworkflow/utils/utils.py", line 32, in <module> import numpy as np File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/__init__.py", line 153, in <module> from . import add_newdocs File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/add_newdocs.py", line 13, in <module> from numpy.lib import add_newdoc File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/lib/__init__.py", line 18, in <module> from .polynomial import * File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/lib/polynomial.py", line 19, in <module> from numpy.linalg import eigvals, lstsq, inv File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/linalg/__init__.py", line 50, in <module> from .linalg import * File "/home/xxx/bin/scipion/software/lib/python2.7/site-packages/numpy/linalg/linalg.py", line 29, in <module> from numpy.linalg import lapack_lite, _umath_linalg ImportError: libgfortran.so.3: cannot open shared object file: No such file or directory
Cause: Missing libgfortran.so.3
The missing library can be installed using:
sudo apt-get install libgfortran3
We have updated the tiff library to handle BIGtiff data and it will be available from Scipion version 2.0.0. If you are running Claudio (v1.2.1) there are some steps you can follow to enable Scipion to work with bigtiff data. Please, take into account that this hasn’t been extensively tested but all our tests where successful. Our recommendation would be to wait for v2.0 release (Spring 2019 aprox.).
If you are determined to move forward follow this steps:
- open a terminal and cd to the scipion folder
- backup your old libtiff files:
mkdir software/lib/old_tiff mv software/lib/libtiff* software/lib/old_tiff/
- modify scipion to use libtiff 4.0.10 (bigtiff lib)
sed -i -e s/tiff-3.9.4/tiff-4.0.10/ install/script.py
- Tell scipion to install bigtiff
./scipion install tiff --no-xmipp
Because we haven’t installed xmipp yet, you’ll see a message saying something like this in the terminal:
Scipion v2.0 (2019-03-12) Diocletian (release-2.0.0-fixes 50b9908) >>>>> python /home/yaiza/Desktop/scipion/pyworkflow/apps/pw_manager.py >>> WARNING: Xmipp binaries not found. Ghost active.....BOOOOOO! > Please install Xmipp to get full functionality. (Configuration->Plugins->scipion-em-xmipp in Scipion manager window)
or this one when importing something:
Error: AttributeError Description: 'NoneType' object has no attribute 'isImage' Traceback: File "/home/me/scipion/pyworkflow/protocol/protocol.py", line 1817, in validate childErrors = self._validate() File "/home/me/scipion/pyworkflow/em/protocol/protocol_import/images.py", line 372, in validate errors += self.validateImages() File "/home/me/scipion/pyworkflow/em/protocol/protocol_import/images.py", line 354, in validateImages ih.isImageFile(imgFn))): File "/home/me/scipion/pyworkflow/em/convert/imagehandler.py", line 436, in isImageFile return xmippLib.FileName(imgFn).isImage()
- Open Plugin Manager
- Select Xmipp to install it by clicking on the empty checkbox on the left.
- Add the number of processors you’d like to use (the more, the merrier!). Then click on the install button on the operations tab
- Now we can check the progress on the Output log tab (or go make some coffee, Xmipp installation will take a bit!). You might have to refresh the logs by clicking on the refresh symbol on the right. Please note that messages might not appear in order if we are using more than 1 processor.
- When the operation gets a green check, it’s done!
Note: if xmipp installation fails, you might have to uninstall it with the plugin manager:
And manually remove leftover elements:
rm -rf software/em/xmipp*
- Now when we close and re-launch Scipion, we should get no messages.
./scipion Scipion v2.0 (2019-03-12) Diocletian (release-2.0.0-fixes 50b9908) >>>>> python /home/yaiza/Desktop/scipion/pyworkflow/apps/pw_manager.py
If you are getting error while scipion tries to install scikit-learn python package, something like:
00086: Building scikit-learn ... 00087: /home/fanhc/Programs/scipion/software/bin/python /home/fanhc/Programs/scipion/software/lib/python2.7/site-packages/pip install scikit-learn==0.17 00088: Collecting scikit-learn==0.17 00089: Using cached https://files.pythonhosted.org/packages/60/b8/c420dce3f72d95e06f7c1e50a6e705f4e8b6078d7d6db38425ac77ae3fab/scikit-learn-0.17.tar.gz 00090: Building wheels for collected packages: scikit-learn 00091: Building wheel for scikit-learn (setup.py): started 00092: Building wheel for scikit-learn (setup.py): finished with status 'error' 00093: Running setup.py clean for scikit-learn 00094: Failed to build scikit-learn 00095: Installing collected packages: scikit-learn 00096: Running setup.py install for scikit-learn: started 00097: Running setup.py install for scikit-learn: finished with status 'error'
Try to run:
scipion python -m pip install scikit-learn==0.17.1
Some program in Xmipp use the sh_alignent library. If you get some of the errors below try the following:
- swig: Command not found: Install swig in your computer, ie. sudo apt-get install swig (yum in Centos distros and zypper in OpenSUSE).
We have find a bug reporting the following error:
133 consensusNpixels = consensusRadius* boxSize 134 135 # Add the rest of coordinates 136 Ncurrent = N0 137 for n in range(1, len(coords_files)): 138 for coord in coords[n]: <----------------------- BUG coord = array([2379, 102]) coords = [array([[3543, 222], [3757, 133], ...3935], [3063, 3935], [ 712, 3944]]), array([[1136, 280], [2388, 2416], ... 120], [1788, 624], [2608, 3204]]), array([[ 663, 3811], [ 287, 3688], ... 162], [3048, 159], [2379, 102]])] n = 3 139 if Ncurrent > 0: 140 dist = np.sum((coord - allCoords[0:Ncurrent])**2, axis=1) 141 imin = np.argmin(dist) 142 if sqrt(dist[imin]) < consensusNpixels:
This bug should be fixed for versions after v19.04. However, to fix it in prior versions, please download the bug-fixed file to your Xmipp installation.
wget -O $(scipionBIN run printenv | grep XMIPP_HOME | sed 's/.*=//')/bin/xmipp_coordinates_consensus https://raw.githubusercontent.com/I2PC/xmipp/devel/applications/scripts/coordinates_consensus/coordinates_consensus.py
Please, ensure it has executable permissions
chmod a+x $(scipionBIN run printenv | grep XMIPP_HOME= | sed 's/.*=//')/bin/xmipp_coordinates_consensus