Skip to content

Python Advanced

Python Advanced#

Regular Expressions#

Basic Patterns#

use of .#

1
- matches any character

use of *#

1
- matches 0 or more repetition of a char/set

use of +#

1
- matches 1 or more repetition of a char/set

use of ?#

1
2
3
4
5
- matches 0 or 1 repetition of a char/set
- also works as a Non-greedy pattern (with repeaters)
    - means in string '<b>Hello</b>' pattern r'<.*>' will match the whole string instead of '<b>'
    - but if pattern is used as r'<.*?>' then will match '<b>' only
    - AKA pcre (Perl Compatible Regular Expression)

use of \#

1
2
- sign of speciality
- used before any special chars, to match that char

use of \t, \r, \n#

1
2
3
4
- \t matches a tab
- \r matches a carriage return (line break) in Mac, \n\r in Windows
- \n matches a line break ( carriage return) in Linux & Windows
- note: On "old" printers, \r sent the print head back to the start of the line, and \n advanced the paper by one line. Both were therefore necessary to start printing on the next line.

use of \b (start and end of word anchors)#

1
2
3
4
5
- matches the boundary between word and non-word chars
- matches the position called word boundaries
    - match has zero length
    - usually before (including start of the line) and after a word
    -e.g. <here>apple<here>

use of \B#

1
2
- opposite of \b
- matches every position where \b does not

use of \s#

use of \d#

use of \D#

use of \w#

use of \W#

use of []#

1
2
- dash - inside []
- dot . inside []

use of ()#

use of ^ (Start of String Anchors)#

1
- with square bracket (set of chars)

use of $ (End of String Anchors)#

use of |#

use of syntax (? ...)#

1
2
3
4
5
6
- Lookarounds
    - +ve Lookahead
    - -ve Lookahead
    - +ve Lookbehind
    - -ve Lookbehind
- Non-Capturing Group

Builtin Functions:#

re.search#

re.group#

re.findall#

1
2
- with files
- with groups

re.sub#

re.compile#

extra options to above functions#

Examples:#

  • Repetitions
  • Leftmost and largest
  • Square Brackets (Set of chars)
    • Email example
  • Group Extraction
  • Greedy vs Non-Greedy
  • Substitution
1
2
3
4
5
6
7
8
9
import re

str = "This#is#$% a&%name%$ #"
r = re.sub(r'(?<=[\w])([\W]+)(?=[\w])', ' ', str)
print(r)
r = re.sub(r'(?<=[A-Za-z0-9])([^A-Za-z0-9]+)(?=[A-Za-z0-9])', ' ', str)
print(r)
r = re.sub(r'q(?=u)', ' ', 'quit')
print(r)
  • remove white spaces from starting of the line but line break
1
2
3
4
regex = r'^[^\S\n]+'

#or if regex supports PCRE
regex = r'^[\h]+'

Credits:

  • https://www.rexegg.com/regex-disambiguation.html#lookarounds
  • www.regular-expressions.info

Multithreading & Multiprocessing#

GIL - Global Interpreter Lock#

  • Is a mutex (mutual exclusion attribute) in python
  • protects python objects from multiple threads
  • i.e. prevents multiple threads to execute python (byte)codes at once inorder to protect access to python objects
  • i.e. provides lock to protect shared mutable state
  • The lock is necessary because CPython's Interpreter or memory management is not thread safe
    • for example, when two threads simultaneously increment the reference count of the same object, the reference count could end up being incremented only once instead of twice.
  • hence, GIL is here to make python thread-safe, wherever needed
  • GIL is controversial because due to it python's multithreading lack few features like
  • python's multithreaded codes cannot utilize multiprocessor system
  • the longer operations like I/O, image processing happens outside the GIL
  • it is only bottleneck for codes which enter GIL for longer time
  • GIL can causes scheduling IO-bound threads ahead of a CPU-bound threads
  • inshort, GIL is only bad for multi-core CPU - bound thread operations
  • each forked process have separate GIL
  • Jython, IronPython does not have GIL
  • writing a C extension needs GIL
  • in Cython the GIL exists, but can be released temporarily using a "with" statement - read more

  • extra: https://stackoverflow.com/questions/1294382/what-is-a-global-interpreter-lock-gil

  • Note : cython & CPython are different

Some terminologies#

Concurrency#
  • When two or more task can start, run & complete in overlapping time periods.
  • It doesn't necessarily mean they'll ever be running at the same instant.
  • e.g. Multi-tasking on a single core
Parallelism#
  • When two are more tasks are executed simultaneously.
process#
  • is an instance of a program running in a computer
  • can contain one or more threads
  • has its independent memory space
  • are spawned by creating a Process() object and then calling its start() method
thread#
  • is a sequence of instructions within a process
  • as a light-weight process
  • all threads shares the same memory space of the process

process block

Multi-threading#

  • can be implemented to speedup the program using module : threading
  • thread-based parallelism concurrency
  • CPython implementation uses a python construct GIL (global interpreter lock). GIL makes sure that at a time only single thread can execute.
  • a thread acquires GIL --> executes just for a while --> passes GIL to next thread
  • happens very quickly that human cannot detect, and creates illusion of multiple thread running simultaneously.
  • in reality all threads works turn by turn into the same core of CPU
  • parallel CPU computation not possible due to GIL
  • but parallel IO operations are possible (it releases GIL on IO)
  • GIL passing is an overhead here.
Note:#
  • can turn off GIL - dirty practice
  • this will result in messed memory management
  • need to be very careful while writting semaphores & mutex properly
Use of thread#
  • in GUI apps to keep UI threads responsive
  • IO tasks (network IO or filesystem IO)
Facts#
  • using multiple-threads for CPU bound tasks will result in worse performance than a single thread

Multi-processing#

  • used to speedup CPU bound tasks using module: multiprocessing
  • process-based parallelism
  • module results in full CPU utilization
  • Inter-process communication can be achieved using queues and pipes *
pipe#
  • is a duplex(two way) communication channel
1
2
3
4
5
6
7
8
9
from multiprocessing import Process

def f(name):
    print('hello', name)

if __name__ == '__main__':
    p = Process(target=f, args=('bob',))
    p.start()
    p.join()

Out: hello bob

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
from multiprocessing import Process
import os

def info(title):
    print(title)
    print('module name:', __name__)
    print('parent process:', os.getppid())
    print('process id:', os.getpid())

def f(name):
    info('function f')
    print('hello', name)

if __name__ == '__main__':
    info('main line')
    p = Process(target=f, args=('bob',))
    p.start()
    p.join()

Out: 

1
2
3
4
5
6
7
8
9
main line
module name: __main__
parent process: 65
process id: 66
function f
module name: __main__
parent process: 66
process id: 80
hello bob

What is .join()?#

  • The join() method, when used with threading or multiprocessing, is not related to str.join()
  • it's not actually concatenating anything together
  • It just means "wait for this [thread/process] to complete"
  • The name join is used because the multiprocessing module's API is meant to look as similar to the threading module's API
  • The reason why is called join is that is joining the processes into a single one.

Note: * By default, when the main process is ready to exit, it will implicitly call join() on all running multiprocessing Process instances * This isn't as clearly stated in the multiprocessing docs as it should be, but it is mentioned in the Programming Guidelines section. * non-daemonic processes will be automatically be joined. * can override this behavior by setting the daemon flag on the Process to True prior to starting the process:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
from multiprocessing import Process
import os

def say_hello():
    print("Hello")

p = Process(target=say_hello)
p.daemon = True
p.start()

# Both parent and child will exit here, since the main process has completed.
# the child process will be terminated as soon as the main process completes:

Out: Hello

daemon#

In Linux: A daemon is a long-running background process that answers requests for services. There are also some other processes like orphan & zombie

  • The process has daemon flag
  • a Boolean value
  • This must be set before start() is called
  • The initial value is inherited from the creating process
  • When a process exits, it attempts to terminate all of its daemonic child processes

Files#

File Handling#

  • How
    • Using try, except, finally
    • Using with (pythonic way)
  • Why
    • to manage resources effeciently
    • file descriptors (fd) are limited at OS level, so we can save fd by closing after use

Using try, except, finally#

  • the step finally is important here for us
    • to use finally we need to write whole try, except, finally block in each languages
  • finally is important because we need f.close() here to close the file descriptor with 100% guarantee
  • file descriptors (fd) are limited at OS level, so we can save fd by closing after use
  • code: 
    1
2
3
4
5
6
7
    try:
    f = open('csv.txt', 'r', encoding='utf-8')
    f.write('abc')
except:
    pass
finally:
    f.close()

Using with#

  • Read here
  • code: 
    1
2
    with open('csv.txt', 'r', encoding='utf-8') as f:
    f.readline()

JSON Operations#

import json#

flask: jsonify()#

Shipping Python Solution#

  • never ship .py files
    • find /path/to/your/files -type f -name "*.py" -delete
  • can ship .pyc or .pyo files

.pyc#

  • compiled
  • size(.py) < size(.pyc)

.pyo#

  • compiled + optimized
    • removed doc strings
    • does not contain "set the current line to ..." bytecode instructions
    • for faster performance
    • renaming .pyo to .pyc works fine
      • find "ETHEREAL_DIR"/Ray/src/ -type f -name "*.pyo" -exec bash -c 'mv $0 ${0/.pyo/.pyc}' {} \;
  • size(.pyo) < size(.pyc)
  • python -O -m compileall /path/to/your/files
  • Note - After this, you may get some import issues

.exe#

  • Source

    • http://docs.python-guide.org/en/latest/shipping/freezing/
    • https://pyinstaller.readthedocs.io/en/v3.3.1/

  • Note:

    • Freezing Python code on Linux into a Windows executable was only once supported in PyInstaller and later dropped
    • All solutions need MS Visual C++ dll to be installed on target machine, except py2app. Only Pyinstaller makes self-executable exe that bundles the dll when passing --onefile to Configure.py.

PyInstaller#

  • Uses the OS support to load the dynamic libraries, thus ensures full compatibility
  • available for windows, linux, macOS etc.
1
pyinstaller -i <icon.ico> --windowed --onefile <python_file.py>

py2exe#

Python Version Management#

Setting default Python Version#

  • not recommanded
  • will affect other applications using another version of python
  • can be done like:
1
2
3
ln -l /usr/bin/python /usr/bin/python3.6

ln -l /usr/bin/python /home/toran/anaconda/bin/python3.6

By aliasing python#

  • not recommanded
  • will conflict when python will encounter in other commands like
    • which python
    • python manage.py runserver
  • can be done like:
1
alias python='/home/toran/anaconda/bin/python3.6`

By setting PATH env variable#

  • recommanded with anaconda
  • works on particular shell only
  • for persistency; include the command in .bashrc or .zshrc file
  • can be done like:
1
export PATH="/home/toran/anaconda3/bin:<dollar>PATH"

Anaconda#

  • install it from official site
  • you can also choose miniconda over regular anaconda
    • can be used in production env/ deployments etc.

Linux#

  • set anaconda python as default

    1
    export PATH="/home/toran/anaconda3/bin:<dollar>PATH"

Windows#

  • set anaconda python as default
1
2
3
4
C:\ProgramData\Miniconda3\Scripts\activate

#if using git bash
. /c/ProgramData/Miniconda3/Scripts/activate

virtualenv#

pipenv#

  • officially recommended dependency management system
  • automagically
    • installs dependencies in venv
    • keep record in Pipfile
      • packages section
      • dev-packages section
    • keeps record of dependencies with specific version in Pipfile.lock file
    • if requirements.txt already exists, installs dep from there
    • installs & uninstalls deps

Installation#

1
pip install pipenv

Usage#

create venv#
1
pipenv install
create venv with specific python version#
1
2
3
pipenv --two install 

pipenv --three install
  • will initialize the venv with that python version
install/uninstall dependencies#
1
2
3
pipenv install nose2

pipenv uninstall nose2
lock dependencies for production#

1
pipenv lock
- will keep record of all the deps in Pipfile.lock file with their specific versions

install dependencies only for development#
1
pipenv install --dev  nose2
  • will keep this record in [dev-packages] section
  • and will do not install it by default
  • to install dev packages; need to do:
1
pipenv install --dev
install dependencies with other options#
  • --dev: Install both develop and default packages from Pipfile.lock.
  • --system: Use the system pip command rather than the one from your virtualenv.
  • --ignore-pipfile: Ignore the Pipfile and install from the Pipfile.lock.
  • --skip-lock: Ignore the Pipfile.lock and install from the Pipfile. In addition, do not write out a Pipfile.lock reflecting changes to the Pipfile.
run command within venv#
1
pipenv run which python
remove venv#
1
pipenv --rm
see venv details/path#
1
pipenv --venv
enter venv shell#
1
pipenv shell
clean venv#
  • uninstall all the packages not recorded in Pipfile
1
pipenv clean
update venv#
  • locks & then update all the packages
1
pipenv update
security check#
1
pipenv check

Python Package Distribution (PyPi)#

Project Structure#

There are various guidelines:

  • https://packaging.python.org/en/latest/tutorials/packaging-projects/
  • https://py-pkgs.org/03-how-to-package-a-python#how-to-package-a-python (A very good read)
  • https://docs.python.org/3/reference/import.html#namespace-packages
  • https://peps.python.org/pep-0420/
1
2
3
4
5
6
7
8
/example_pkg
  /pkg1
    __init__.py
  /pkg2
    /pkg2_1
  setup.py
  LICENSE
  README.md

where - __init__.py should contain a variable name='example_pkg' - setup.py should look like this gist - where classifiers can be like this or gist - README.md should be present, which will define the long_description about the package - LICENSE should be accurate with the list - packages - if want to includes all the packages/subpackages - use defaultone packages=setuptools.find_packages(), - else define manually like - packages=['pkg1', 'pkg2', 'pkg2.pkg2_1']

Note: After installing the package, you able to import packages with names listed in packages var.

Pre-requisites#

1
2
3
4
5
6
#setuptools & wheel - to create build in .whl as well as .tar.gz. format
python3 -m pip install --user --upgrade setuptools wheel


#twine - to upload build in pypi server
python3 -m pip install --user --upgrade twine

Build#

  • change directory where setup.py is & run
1
python3 setup.py sdist bdist_wheel

output will be like

1
2
3
dist/
  example_pkg-0.0.1-py3-none-any.whl
  example_pkg-0.0.1.tar.gz

Test Distribution#

1
twine upload --repository-url https://test.pypi.org/legacy/ dist/*

Final Distribution#

1
2
3
4
5
twine upload --repository-url https://upload.test.pypi.org/legacy/ dist/*

#OR

twine upload dist/*

Rebuild#

  • delete build, dist, & *.egg-info files then build again

Test#

(Test) Package#

1
python3 -m pip install --index-url https://test.pypi.org/simple/ example_pkg

(Final) Package#

1
python3 -m pip install example_pkg

Advanced - Run Package in Command Line#

Source: http://python-packaging.readthedocs.io/en/latest/command-line-scripts.html

FAQ#

  • You're not allowed

  • File already exists

  • File already exists, change version

  • particular line is not know/right

Python Linter & PEP-8#

  • I found flake8, autopep8 & yapf similar
  • but I found black best among them
    • quotes: double
    • no extra line change
    • only draw back is bad list slice formating
      • create issue: https://github.com/python/black/issues

Python Templates#

PyCharm Live & File Templates#

https://gist.github.com/toransahu/6080cbf2cc1a8808f19bd0eccafc5ef0

Vim Template#

https://github.com/toransahu/vim-template/blob/master/templates/%3Dtemplate%3D.py