Python mtest module not working as expected?

[mhpierre]

Hi,
I'm trying to implement more advanced things in mtest, namely finding the right increment of driving variables to ensure some condition on some internal state variables. This would require accessing the internal state variable and using something like a Newton algorithm to find the right increments.
The python module works fine for me in a simple case (using MTest.execute()), however when I try to get under the hood I'm faced with some problems.
Firstly, the printOutput function seems to yield no output. The file is empty after calling it.
Secondly, I can't seem to access the underlying elements of the MTestCurrentState class. Trying to access s1 for instance , I get this error:

TypeError: No to_python (by-value) converter found for C++ type: tfel::math::vector<double>

I'm currently working with tfel 4.2.0 against python 3.11.6.

Thanks in advance,
Maxime

[thelfer]

irstly, the printOutput function seems to yield no output. The file is empty after calling it.

The attached example test.zip works fine. Would you test it ?

$ mfront --obuild --interface=castem Tvergaard.mfront
$ python test.py

Secondly, I can't seem to access the underlying elements of the MTestCurrentState class. Trying to access s1 for instance , I get this error

You need to import the std module firts

Best

[mhpierre]

Your example works fine indeed, there seems to be a weird interaction with jupyterlab where it doesn't print to the file but as a python script it works, hence my confusion.

You need to import the std module firts

I had the std module imported already, I still have this error ...

[mhpierre]

For some reason, adding import tfel.math seems to do the trick.

[thelfer]

My bad, I misread the error message. I though we were taking std::vector. For tfel::math::vector, you indeed to need to import the tfel.math model