Hey @StrikerRUS, thanks for reporting this. Can m2cgen be used as a library in a Jupyter Notebook scenario?

@izeigerman Why not?.. I just used it so for preparing #109.

Yeah, I'm just trying to understand whether piping is sufficient after all or not.

Just a comment; I started working on this today and the docs could be upgraded to clarify that CLI use can use piping to save to a file. I "might" try to tackle this as I would like to get my feet wet on contributing to OSS.

@eafpres Do you mean docs of the CLI tool itself or the README? We do mention this in the latter: https://github.com/BayesWitnesses/m2cgen#cli . In any case please feel free to submit a PR. Contributions are welcome!

@izeigerman Initially I was going to propose a change to the readme. Piping works fine on the CLI so I don't see an urgent need to explicitly add the functionality to that tool. Long-term maybe.

Hello,
Is there any example showing how to save the model to a specific path?. This would be very helpful as I try several models to run in the background and would appreciate if models can be saved in a folder after each iteration.

Hey @wbadry!

If you're using m2cgen in CLI you can just redirect its stdout into the file. Eg:

$ m2cgen <path_to_model> --language c > <path_to_output_file>

Alternatively if you're using the Python API:

import m2cgen as m2c

generated_code = m2c.export_to_c(...)
with open('<path_to_output_file', 'w') as fd:
    fd.write(generated_code)

Hope this helps!