This is an interactive python script that can run python files, and preserve the global state between runs. It is helpful for when the initialization step of your code (like loading a large csv or asset) takes a long time, which would make rerunning the script over and over again directly a bit unpractical.
A Jupyter notebook seems to be the accepted solution for this kind of problem, but I don't like it because:
- Your can't work on your code as a collection of python files forming a software module
- You can't edit the code with an editor of your choice unless that editor has a Jupyter Notebook extension
- The Jupyter Notebook extension is not guaranteed to work all the time. Maybe if it had worked for me today, I wouldn't have made this ...
- It is overkill for something that can be achieved quite simply
Download the main.py and put it somewhere.
Now run it like python main.py. You should get an interactive REPL-like command line interface.
Type help and hit enter to see all the commands you have.
An example use:
D:\Some\Directory>python main.py
Started. Enter 'help' for help
> cd testing
changed current directory to : D:\Projects\Python\ComputeGraph\testing
Files:
algorithm.py, init.py
> run algorithm.py
running algorithm.py ...
an error occured when trying to execute the command:
name 'x' is not defined
Note that this will run arbitrary python code, with exec.
Also, try to avoid running code that will ask for input, make sure you are doing that if __name__ == "__main__": thing.
> run init.py
running init.py ...
done
> run algorithm.py
running algorithm.py ...
55
done
> vars
x: 55
function: <function function at 0x0000024B22178AF0>
>
-
Transactional runs, and Undo/Redo - Allow state to be reset to what it was before running a script
-
Change tracking - Keep track of which script run modified which global variable, and when