Run_PyStan_in_Atom.rst

How to use PyStan in Atom editor?

This is a short doc explaining how PyStan can be run in Atom on a Windows machine.

PyStan doesn't automatically run in Atom even if we use script to pick the correct python.exe file from the environment we have installed Stan in. The reason is that Windows is very sensitive to the activation of an environment. If the environment is not explicitly activated before running the PyStan code in Atom compiling errors will occur.

Steps

• Install PyStan as explained here.

• Install Atom. Atom will be added to your path variable during the installation.

• Launch Atom in your activated Stan environment as follows:

  1. Open Anaconda Prompt or Command Prompt.

     1. First time only on Command Prompt: Initialise anaconda on the Command Prompt by typing conda init. This will allow you to run conda commands with Command Prompt as you would with Anaconda Prompt.

  2. Activate your Stan environment with activate myenv, where myenv is the name of your environment such as stan_env.

  3. Type atom to launch the Atom editor (once launched you can close the comand prompt window if you like).

Test

To check if you can now run PyStan in Atom, copy and run the example code from the installation doc.

>>> import pystan
>>> model_code = 'parameters {real y;} model {y ~ normal(0,1);}'
>>> model = pystan.StanModel(model_code=model_code)
>>> y = model.sampling().extract()['y']
>>> y.mean()  # with luck the result will be near 0

In my case it looked like pystan was running but after a while the program appeard to repeat itself showing the following output:

INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_19a09b474d1901f191444eaf8a6b8ce2 NOW. INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_19a09b474d1901f191444eaf8a6b8ce2 NOW. INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_19a09b474d1901f191444eaf8a6b8ce2 NOW. INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_19a09b474d1901f191444eaf8a6b8ce2 NOW. INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_19a09b474d1901f191444eaf8a6b8ce2 NOW.

Have a look at this comment for more details.

To remedy this we will need to add if __name__ == "__main__": to our program which will look like the below. Jupyter does this by default and the line isn't necessary.

>>> import pystan
>>> model_code = 'parameters {real y;} model {y ~ normal(0,1);}'
>>> if __name__ == "__main__":
>>>     model = pystan.StanModel(model_code=model_code)
>>>     y = model.sampling().extract()['y']
>>>     y.mean()  # with luck the result will be near 0