A Lite Package focuses on making project's after-post mending pythonic and flexible. Certainly used for expanding finished projects without changing exists source code.

• No Origin Codes Changes, Injection with Licenses
• None Invasive Embed Way, More Flexible
• ✨Easy to use✨

Thr Official Packages ast, inspect, functools used.

Overwrite requires Python 3.6+ to run.

Simply install with one-line pip command.

pip install Mending

Core Import

from Mending import Mend

Set a License Any type of Cipher-text is satisfactory. Set int 0 as example.

_license = 0

Defined Function Mending function, used for overwrite previous method.

def func():
    return

Register Function to Event queue.

1. Register_type (string): Register with the same Event_queue_type will be added to the same event queue (in registration order).
2. times (integer): The times this function calling is affected. (if affected forever, set times to -1)
3. license (any): Used to identify whether the modification is valid and distinguish between different mending events.
4. func (function): Mending function, used for overwriting.

Mend = Mend()
Mend.Butler.register('Register_type', times, _license, func)

# Add entrance to function
@Mend.entrance(catalog='Register_type')
def affected_func():
    return

Add the License Claim If the license is valid, the registered function affected_func will be modified and completely overwritten by func.

# Claim above the function calling
Mend.claim(_license)
affected_func()

The Performance above is same as:

func()

Complex Overwrite

Mend.Butler.register('Event_queue_type_A', 1, license_A, func_A)
Mend.Butler.register('Event_queue_type_A', 2, license_B, func_B)
Mend.Butler.register('Event_queue_type_B', 1, license_B, func_C)

def func_A():
    print('1')
    return
def func_B():
    print('2')
    return
def func_C():
    print('3')
    return

@Mend.entrance(catalog='Event_queue_type_A')
def affected_func_A():
    print('A')
    return
@Mend.entrance(catalog='Event_queue_type_B')
def affected_func_B():
    print('B')
    return

Different license claiming leads to different way of overwriting. After several times calling set before, the performance of modified function will change back to origin again as a result of event queue gets empty. The program works as below:

@Mend.claim(license_B)
affected_func_A() # Output: 2
@Mend.claim(license_A)
affected_func_A() # Output: 1
@Mend.claim(license_B)
affected_func_A() # Output: 2
@Mend.claim(license_A)
affected_func_A() # Output: A
@Mend.claim(license_B)
affected_func_A() # Output: A
@Mend.claim(license_B)
affected_func_B() # Output: 3
@Mend.claim(license_B)
affected_func_B() # Output: B

Get Catalog Name and Available Licenses of Original Function

Mend.Butler.get(affected_func)
-> dict {'catalog_name': str, 'available_license': list[str]}

Browse the Event Queue

Print details for Event Queue:

Mend.Queue.print(catalog: str)

Get length of Event Queue:

Mend.Queue.len(catalog: str)

Get Event Queue as following type:

Mend.Queue.get(catalog: str)

[{'times_left': int, 'license': any, 'function_name': str}, 
 {'times_left': int, 'license': any, 'function_name': str},
 {'times_left': int, 'license': any, 'function_name': str},]

Directly manipulate the Event Queue

Class Queue redefined, making it compatible with Pythonic Style Commands as below:

Mend.Queue.append(catalog: str, event: list)
Mend.Queue.pop(catalog: str, index: int)
Mend.Queue.replace(catalog: str, event: list, index: int)
Mend.Queue.clear(catalog: str)

For multi-threaded function overrides, the event queue is cross-threaded, and the order of event completion depends on the total program time sequence, asynchrony may cause event order confusion. Instantiating Mending for each individual thread is a reasonable solution.

Version: 2021.11.18 (0.0.8)

Author: Zack the White, Qcmcmc

Email Contact: Zack the White Qcmcmc

MIT