Skip to content

Utilities

Overview

From time to time, I usually read internal codebase of public repository, then I come up with some best practice overthere that help me a lot in the time of coding.

Code

To avoid de-duplicated on search, using set for list

Pattern

pool = set() # Define empty set
for element in basket:
    ...
    if <condition>: # Defile condition matching
        pool.update(element)
    ...

# Then we can convert to list or yield from set
for runner in pool:
    # ... do something
    # ...

# or convert back to list
pool = list(pool)

Example:

def _list_folders(self, bkt, prefix, artifact_path):
    results = bkt.list_blobs(prefix=prefix, delimiter="/")
    dir_paths = set()
    for page in results.pages:
        dir_paths.update(page.prefixes)

ref: https://github.com/mlflow/mlflow/blob/fe6434603499bc8e420b75492aa1c0ba64064f39/mlflow/store/artifact/gcs_artifact_repo.py#L30

Validate python version

Valiudate specific version in application. The snippet can put in __init__.py or entrypoint script.

```py indent=4 title="init.py"

!/bin/python3

Gloabal

import sys

Declare

required_major, required_minor, _ = "3.9.x".split(".")

Extract

major, minor, micro, releaselevel, serial = sys.version_info

Required

if not all([major >= int(required_major), minor >= int(required_minor)]): raise RuntimeError( f"Python version required to >= {required_major}.{required_minor}.x release. " f"Got Python version {'.'.join([str(x) for x in sys.version_info])}" ) 

### Detect the system platform identifier

From `sys` library, enumuration of system platforms:

Table 1: System Platform Encoding

| Platform       | Identifier |
| -------------- | ---------- |
| AIX            | 'aix'      |
| Linux          | 'linux'    |
| Windows        | 'win32'    |
| Windows/Cygwin | 'cygwin'   |
| macOS          | 'darwin'   |

```py
import sys

if sys.platform == "linux":
  ...
elif sys.platform.starswith("cy"):
  ...

One Line Punch

One line punch is a 1 line command in python to do various tasks.

python -c "import sys; print(sys.version)"
# 3.12.8 (main, Jan 14 2025, 22:49:36) [MSC v.1942 64 bit (AMD64)]

Hash an target IP

python -c "print('.'.join([val if ind != 2 else '___' for ind, val in enumerate('196.168.5.3'.split('.'))]))"
# yield '196.168.___.3'

Lambda with default argument

To declare lambda with default argument, using following syntax

lambda arg1=value1, arg2=value2, ..., argN=valueN : expression
# with:
# arg1, arg2, argN – the names of the arguments that the lambda expression uses;
# value1, value2, valueN – values that are assigned respectively to the arguments arg1, arg2, argN;
# expression – the lambda expression itself.

For example:

summ = lambda a=1, b=2, c=3: a+b+c

print("summ() = ", summ()) # 1+2+3 = 6
print("summ(10) = ", summ(10)) # 10+2+3 = 15
print("summ(10, 20) = ", summ(10,20)) # 10+20+3 = 33
print("summ(10, 20, 30) = ", summ(10,20,30)) # 10+20+30 = 60

If you do not specify the values of the arguments when calling the lambda expression, then they will be assigned the default values value1, …, valueN. If you specify only part of the arguments, then this part will replace the default value starting from the leftmost argument (arg1).

If you declare a lambda expression that takes 3 arguments by default, and call the lambda expression with 1 argument, then the changes will affect the leftmost argument. The following example demonstrates the rule for overriding default arguments when invoking a lambda expression.

Reference: https://www.bestprog.net/en/2021/06/17/python-default-arguments-in-lambda-expressions/