sudo pip install -r requirements.txt

SurgicalOntologist fucked around with this message at 22:29 on Jan 29, 2014

deimos fucked around with this message at 22:45 on Jan 29, 2014

import pickle


def load_pickled_instance(filename):
    with open(filename, 'rb') as f:
        return pickle.load(f)


class ToBePickled():
    def __init__(self):
        self.functions = []

    def add_function(self, func):
        self.functions.append(func)
        return func

    def save(self, filename):
        with open(filename, 'wb') as f:
            pickle.dump(self, f)

from test import ToBePickled

picklee = ToBePickled()

@picklee.add_function
def test():
    pass

picklee.save('test.pkl')

In [1]: from test import load_pickled_instance

In [2]: p = load_pickled_instance('test.pkl')
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-2-cbf107d83358> in <module>()
----> 1 p = load_pickled_instance('test.pkl')

~/projects/working-dir/test.py in load_pickled_instance(filename)
      4 def load_pickled_instance(filename):
      5     with open(filename, 'rb') as f:
----> 6         return pickle.load(f)
      7 
      8 

AttributeError: 'module' object has no attribute 'test'

_pickle.PicklingError: Can't pickle <class 'function'>: attribute lookup builtins.function failed

• Enforce that unpickling can only occur from the script that was originally __main__ (test.py in this example), it works in that case. This has a major drawback of how to deal with the nice command-line interface that my library has. Client code would have to copy/paste the CLI code into their scripts so that they could load their objects using the same commands. Maybe there's a way to easily hook the command-line interface of one script into that of another.
  
• In __getstate__, pickle only the function's __code__ attribute and manually execute in in __setstate__. I think this would work, the only thing I can't wrap my head around is the imports that the function relies on. Not sure how difficult this would be to implement.
  
• Pickle only the function's __name__, and the correct script name (via sys.argv[0]). To unpickle, import it and reattach.
  
• I just thought of this: Save the original sys.argv[0], and in load_pickled_instance, import everything from it into the __main__ namespace.

from original_script import *

SurgicalOntologist fucked around with this message at 06:25 on Jan 31, 2014

SurgicalOntologist fucked around with this message at 06:58 on Jan 31, 2014

SurgicalOntologist fucked around with this message at 07:14 on Jan 31, 2014

def __getstate__(self):
    state = self.__dict__.copy()
    # Save only function names.
    state['callbacks'] = [func.__name__ for func in state['callbacks']]
    # Save original module name.
    state['original_module'] = sys.argv[0][:-3]
    return state

def __setstate__(self, state):
    # Import original module.
    original_module = importlib.import_module(state['original_module'])
    # Replace the references to the original functions.
    state['run_callbacks'] = [getattr(original_module, func) for func in state['run_callbacks']]
    self.__dict__.update(state)

SurgicalOntologist fucked around with this message at 03:29 on Feb 1, 2014

• Powershell "[Environment]::SetEnvironmentVariable("Path", "$env:Path;C:\Python27", "User")"
  
• Adding Python to Windows PATH (both through the Control Panel and terminal).
  
• Copying a python.bat file with "@C:\Python27\python.exe %*" to the system32.
  

cinci zoo sniper fucked around with this message at 18:56 on Feb 2, 2014

cinci zoo sniper fucked around with this message at 00:31 on Feb 4, 2014

import os
from matplotlib import pyplot

X = range(10)
Y = [x**2 for x in X]
pyplot.plot(X,Y)
path = os.path.join(os.getcwd(),'science.png')
pyplot.savefig(path)

FoiledAgain fucked around with this message at 08:51 on Feb 5, 2014