From the command line, can you run
tracker -v
Which should respond with 22.1.0
In the meantime, I am copying your environment to test.
Thanks for the response, let me know what you find!
Ran tracker -v as suggested and can confirm version # is 22.1.0
Also not sure if this matters but I installed Tracker (and dependencies) in the base environment.
Installing in “base” environment is generally not a problem, it just removes the ability to maintain separate versions at the same time, and makes uninstallation a little less convenient, etc.
Small update: just to try something different, I uninstalled my Anaconda distribution, reinstalled (this time using Miniconda), and created a new environment for Tracker (tracker-env).
The previously-mentioned “error on start-up” (QMimeDatabase…) disappeared. However, the other issues still occur (crash with file menu and errors with OpenCV), with the same error messages.
Versions are below (slightly different than before). All packages were downloaded from conda-forge channel (except Tracker of course), not sure if that matters?
conda 23.11.0
# Name Version Build Channel
python 3.12.1 h2628c8c_1_cpython conda-forge
pyqt 5.15.9 py312he09f080_5 conda-forge
qt-main 5.15.8 h9e85ed6_18 conda-forge
qtpy 2.4.1 pyhd8ed1ab_0 conda-forge
opencv 4.9.0 py312haba1e54_6 conda-forge
tracker 22.1.2 py_0 https://mfix.netl.doe.gov/etc.
numpy 1.26.3 py312h8753938_0 conda-forge
scipy 1.12.0 py312h8753938_2 conda-forge
Just thought I would circle back to share that I was able to get Tracker running by installing it into an environment with previous versions below. No more error messages (although I find Tracker frequently crashes if I click on a command before it finished processing the previous one – not sure if that’s just me but anyway not a big deal).
I imagine updates may be forthcoming but in the interim, this works for me and maybe others will benefit too.
python 3.7.16
pyqt 5.12.3
qtpy 2.4.1
opencv 3.4.9
numpy 1.21.6
scipy 1.7.3
Commands I used (channel address for Tracker to be replaced with unique link for the user):
conda create -n trackpy37 python=3.7
conda activate trackpy37
conda config --env --add channels conda-forge
conda config --env --set channel_priority strict
conda install pyqt=5.12 opencv=3.4 qtpy scipy numpy matplotlib tracker -c https://usr:token@mfix.netl.doe.gov/dist
Yes we use the conda-forge channel since it provides packages we need which are not in the main channel.
It’s sometimes a problem using a mix of main and conda-forge packages, but it looks like this is not the issue in your case. 100% conda-forge (plus the stuff from NETL) is good.
These commands prevent the (undesired) mixing of conda-forge and non-conda-forge packages