CODEWITHSUNDEEP
python
Whitehot
Whitehot

Reputation: 497

Identify where a python script was launched from

I was looking through methods of handling arguments when I realised that I'm actually looking to do something slightly different.

Basically I am writing my python3 script in Jupyter as foo.ipynb to test it on small datasets, and to share it with my team easily, but I will also be exporting the script to foo.py and running it on our cluster for the bigger datasets. I would like it to have some different behaviours, like getting a filepath from an argument or not, depending on how the script was ran.

Is there any way for a python script to know if it was ran in a Jupyter notebook or through the command line? Our cluster uses slurm, so foo.py will be called from a bash script, is there any way to handle that as well?

Upvotes: 0

Views: 63

Answers (1)

yvesonline
yvesonline

Reputation: 4837

You can use __file__ (Python doc).

When calling foo.py directly via python foo.py then __file__ will be set, e.g.: foo.py with only a print(__file__) will output foo.py.

If called in Jupyter it won't be set.

Regarding slurm, I'd recommend passing in another argument to your script to distinguish this, e.g. foo.py --slurm with argparse.

Code Addendum:

if '__file__' in locals():
    print(__file__)
else:
    print('__file__ not set')

Upvotes: 1

Related Questions