I am trying to reuse a package that I have written in a new project and I am currently having problems with imports. I recreated a minimal example that shows the problem. The folder structure is as following:
test\
__init__.py
main.py
package\
__init__.py
p1.py
p2.py
The problem stems from the fact that p1 imports p2
from __future__ import print_function
print("P1",__name__,__package__)
import p2
def a():
p2.b()
print("A")
if __name__=="__main__":
a()
p2.b()
And this is the main.py code:
from __future__ import print_function
print("MAIN",__name__,__package__)
from package.p1 import a
from package.p2 import b
if __name__=="__main__":
a()
b()
and here is p2.py:
from __future__ import print_function
print("P2",__name__,__package__)
def b():
print("B")
if __name__=="__main__":
b()
When running main.py I get a ImportError: No module named p2
I can fix this according to the modules documentation by changing the import in p1:
Since the name of the main module is always "main", modules intended for use as the main module of a Python application should always use absolute imports.
My understanding of this sentence is that I should change the import in p1 to from package import p2. However this breaks p1:
ImportError: No module named package
I've found other answers that just suggest to modify sys.path explicitly adding folders but that does not seem clean to me. What am I missing? What is the pythonic way to do this?
