Open 3sigma opened 7 years ago
Hi!
Sorry for the delayed response! As you may have noticed, boxflow isn't currently under active development right now although I would love a chance to work on it again (perhaps in my free time).
As to answering your question, yes, boxflow builds graphs of parameterized objects and we generally make sure that these objects have valid reprs. This means that it wouldn't be hard to ensure that the repr of a boxflow network would be valid python code that you could then use in a regular python script or in a Jupyter notebook.
Most the examples I have played with use such objects: for instance imagen pattern generators that can take numbergen objects (available with param) which can supply dynamic values (there are some examples on the imagen homepage).
In that example, boxflow could build a pattern generator with a repr that when evaluated gives you a pattern (i.e a numpy array with some structure, which could be time varying, with all sorts of parameters you could play with). Just to be clear, the exact way you evaluate the object built with boxflow is going to depend on the type of object it is.
Hope that helps clarify things! If you are interested in boxflow, and wish to make any contributions I would be very happy to make suggestions and/or review any PRs.
Hello,
No issue, but I wonder: once a diagram is made, what is it used for ? Can you generate Python code from it ? Can you execute it ? And if yes, how ?
Thanks in advance for your answer