Thanks to this post, from Greg, K17MT, here's what you do.
sudo add-apt-repository ppa:ki7mt/wspr
# press enter when asked something
sudo apt-get update
sudo apt-get install wspr ntp
Now you can use the Ubuntu launcher to launch wspr, or as I did, drag the icon to the side dock and launch.
Note that I also recommend installing the ntp time sync service as I find my ubuntu laptop clock drifts off and by default Ubuntu only syncs the clock on each boot.
Watch out for me, vk2tpm, on 20m.
Obsolete below here
Building WSPR on linux is often a bit of a puzzle. With the help of the clearly very clever George Smart M1GEO I got most of the way.
He didn't draw a lot of attention to it, but after running ./autogen.sh you do need to edit the Makefile and change /usr/bin/f2py to /usr/bin/f2py3.4
After successful make and make install, running wspr gave me a python exception ending in ImportError: cannot import name '_imagingtk'
The solution is:
sudo apt-get install tk8.5-dev tcl8.5-dev
sudo pip3 install -I pillow
It's now running although I now get a crash shortly after it starts decoding.
So audio in is working. My guess is something is wrong with the fortran to python interface. Any tips would be most welcome.