module-hot-reload

module_hot_reload

Package for reloading other packages and modules while Python is running

Installation

pip install module-hot-reload

Usage

Instantiate reloader you need, import module that you want to be reloaded, register it in reloader, start reloader

from module_hot_reload.reloaders import (
    NewModuleAwareAllModulesRecursiveAutomaticReloader,
)

import example

r = NewModuleAwareAllModulesRecursiveAutomaticReloader()
example = r.register(example)
r.start()

while True:
    print('example.e ', example.e)
    input('waiting..........')

There are 2 types of module wrappers: (confusingly)ModuleWrapper - used primarily for reloading of module and its submodules and does it thread-safely; ModuleAttributeAccessor - provides thread-safe . operator and prevents collisions of module and wrapper class attribute names. Reloader's register() method returns module wrapped with ModuleAttributeAccessor.

The "API-methods" accept modules and know how to extract one form ModuleWrapper or ModuleAttributeAccessor. So you can safely do something like

from module_hot_reload.reloaders import (
    NewModuleAwareAllModulesRecursiveAutomaticReloader,
    NewModuleAwareAllModulesRecursiveManualReloader,
)
from module_hot_reload.module_wrappers import (
    ModuleAttributeAccessor,
    NewModuleAwareAllModulesRecursiveStandardModuleWrapper,
)

import example


r = NewModuleAwareAllModulesRecursiveAutomaticReloader()
w = r.module_wrapper_class

example = w(example)
# example is a ModuleWrapperBase instance now

example = r.register(example)
# example is a ModuleAttributeAccessor instance now

example = w(example)
# example is a ModuleWrapperBase instance now

#  \/  code sample continues  \/

As mentioned above, ModuleAttributeAccessor provides thread-safe . operator. ModuleWrapper does provide same functionality as well but with bulkier syntax.

#  /\  code sample continuation /\

example = ModuleAttributeAccessor(example)
print(example.e)  # with ModuleAttributeAccessor

example = w(example)
print(example.locked_get('e'))  # with ModuleWrapperBase

#  \/  code sample continues  \/

The question is: "Why would you need to use ModuleWrapper then?". You really don't need to. Using it you have the ability to manually reload the module and access its attributes; but there are manual reoaders whose register() method still returns ModuleAttributeAccessor.

#  /\  code sample continuation /\

mr = NewModuleAwareAllModulesRecursiveManualReloader()
example = mr.register(example)
mr.reload()
print(example.e)

# is equivalent to

mw = mr.module_wrapper_class
example = mw(example)
example.reload()
print(example.locked_get('e'))

How it works?

Actual reloading of module(s) is done with importlib.reload() so reed the docks to learn about reloaded modules behaviour.

Automatic reloaders use watchdog to watch file system events. It works with Windows as well as Linux.

ModuleWrappers and ModuleAttributeAccessors use sort of singleton pattern but there is an instance of a particular class per wrapped module, so that

from module_hot_reload.module_wrappers import (
    ModuleAttributeAccessor,
    NewModuleAwareAllModulesRecursiveStandardModuleWrapper,
)

import example


maa_1 = ModuleAttributeAccessor(example)
maa_2 = ModuleAttributeAccessor(example)
w_1 = NewModuleAwareAllModulesRecursiveStandardModuleWrapper(example)
w_2 = NewModuleAwareAllModulesRecursiveStandardModuleWrapper(example)

print(maa_1 is maa_2)  # True
print(w_1 is w_2)  # True