For example, a project where main.py
calls another file ...
- main.py
- sub1.py
- folder
- sub2.py
↓ ↓ ↓ The functions and classes that depend on it ...
onefile.py
Gyu! !! !! I'm surprised.
(The motivation is that I wondered if I could easily execute the python file managed by the project with google colab.)
Here, you can use a python package called stickytape. https://pypi.org/project/stickytape/
stickytape = Adhesive tape
All the experimented code is published on Github, so for reference.
install Now you can use the stickytape command. (No command name ...)
$ pip install stickytape
The configuration is as follows.
- main.py
- sub1.py
- folder
- sub2.py
folder/sub2.py Make an Apple class as you like and give it a value property as you like.
class Apple:
def __init__(self, value):
self.value = value
sub1.py I will make an average function as appropriate.
sub1.py
def mean(a, b):
return (a+b)/2
main.py Import it, calculate it appropriately, and display it appropriately.
from sub1 import mean
from folder.sub2 import Apple
apple1 = Apple(value=100)
apple2 = Apple(value=200)
result = mean(apple1.value, apple2.value)
print(result)
Execute the following command. (Of course, anything in onefile.py is OK)
$ stickytape main.py > onefile.py
The following ʻonefile.py` will be generated.
#!/usr/bin/env python
import contextlib as __stickytape_contextlib
@__stickytape_contextlib.contextmanager
def __stickytape_temporary_dir():
import tempfile
import shutil
dir_path = tempfile.mkdtemp()
try:
yield dir_path
finally:
shutil.rmtree(dir_path)
with __stickytape_temporary_dir() as __stickytape_working_dir:
def __stickytape_write_module(path, contents):
import os, os.path
def make_package(path):
parts = path.split("/")
partial_path = __stickytape_working_dir
for part in parts:
partial_path = os.path.join(partial_path, part)
if not os.path.exists(partial_path):
os.mkdir(partial_path)
open(os.path.join(partial_path, "__init__.py"), "w").write("\n")
make_package(os.path.dirname(path))
full_path = os.path.join(__stickytape_working_dir, path)
with open(full_path, "w") as module_file:
module_file.write(contents)
import sys as __stickytape_sys
__stickytape_sys.path.insert(0, __stickytape_working_dir)
__stickytape_write_module('sub1.py', 'def mean(a, b):\n return (a+b)/2')
__stickytape_write_module('folder/sub2.py', 'class Apple:\n\n def __init__(self, value):\n self.value = value')
from sub1 import mean
from folder.sub2 import Apple
apple1 = Apple(value=100)
apple2 = Apple(value=200)
result = mean(apple1.value, apple2.value)
print(result)
For a moment, it became "What's wrong !?", but when I executed this ...
150.0
The correct calculation result is displayed safely.
I copied the above code to Google Colab and ran it.
As shown below, 150.0 was displayed safely.
(Don't worry about the rainbow-colored cat walking.)
It's a digression from here.
The command stickytape
is long, and I'm lazy to specify the directory of the generated file one by one, so I think it's a good idea to script it as follows.
- main.py
- sub1.py
- folder
- sub2.py
- scripts
- tape.sh
- build
- onefile.py
tape.sh
#initial value
entry="main.py"
output="onefile.py"
#option
while getopts e:o: OPT
do
case $OPT in
"e" ) entry=${OPTARG};;
"o" ) output=${OPTARG};;
esac
done
#Run
stickytape ${entry} > "build/${output}"
The following command will run main.py and generate onefile.py in the build directory.
$ sh scripts/tape.sh
I also prepared options.
Option name | Description |
---|---|
-e | Entry point filename |
-o | File name to output |
$ sh scripts/tape.sh -e <file name> -o <file name>
The generated directory is fixed with build
, so if you don't like it, change it.
If you write it at the beginning, it will get in the way, so let me introduce yourself quietly at the end.
name | Aki Wataoka |
---|---|
school | Kobe University Graduate School |
Undergraduate research | Machine learning,Speech processing |
Graduate study | Machine learning,fairness,Generative model, etc |
@Wataoka_Koki |
Follow us on Twitter!
Recommended Posts