Welcome to Cadabra Q&A, where you can ask questions and receive answers from other members of the community.
0 votes

After a long absence I've decided to try and restart using Cadabra again.

I'm running OpenSUSE (Leap) for which there doesn't seem to be a recent build on github, so I downloaded Cadabra_2.5.4_x86_64.AppImage instead.

Then I tried to run the schwarzschild.cnb tutorial. At the 2nd code block it complains that it cannot find the sympy module.

I reinstalled Anaconda, then used conda to install and update python3 and sympy. This didn't help. I also tried "pip install sympy", but this didn't help either.

What am I doing wrong?

[BTW: if I run python3 standalone, it imports sympy without complaint.]

in Installation trouble by (460 points)
edited by

2 Answers

0 votes

The AppImage uses a bundled Python (it has to, because it is meant to be standalone). So you cannot resolve this "from the outside", it has to be fixed inside the AppImage itself. I will investigate, this should not be terribly difficult to fix. WIll post here when it's resolved.

(OpenSUSE Leap is sufficiently far behind the current versions of many libraries and tools that it is very difficult to build Cadabra on it. I simply lack the time to investigate that route any further).

by (83.1k points)

No need to worry about OpenSUSE Leap. Their development path seems to be departing from desktop systems so I'll probably have to adopt a different default Linux soon.

I'm happy to use a bundled Python in the Cadabra AppImage. I daresay it would always remain reasonably new.

But does this mean that I don't need my own local up-to-date Python installation at all (nor anaconda) in order to run the AppImage?

0 votes

I have just released 2.5.6, which fixes this issue, and also includes plotting support in the AppImage version. You can find the updated AppImage in the github release assets for this release.

by (83.1k points)

OK, schwarzschild.cnb now runs ok for me.

But the first time I tried to evaluate the derivative of a function for myself, it didn't work, even though I used very-similar syntax to that in schwarzschild.cnb.

But I'll start a separate thread for that.

...