ong-utils 0.4.2

Common utilities for python projects

Ong_Utils

Description

Simple package with some utils to import in any project:

• Class to manage configuration files in yaml or json. It also uses keyring to store and retrieve passwords
• logger and a timer to record elapsed times for optimizing some processes
• a create_pool_manager function to create instances of urllib3.PoolManager with retries and timeouts and checking of https connections
• a TZ_LOCAL variable with the local timezone of the computer).
• a is_debugging function that returns True when debugging code
• a cookies2header that converts cookies in dict to header field 'Cookie' for use in urllib3
• a get_cookies function to extract a dict of cookies from a response of a urllib3 request
• a class to store any data into keyring (e.g. strings, dicts...)

Optional dependencies

Installing pip install ong_utils[shortcuts]:

• functions to create desktop shortcuts for packages installed with pip

Installing pip install ong_utils[xlsx]:

• function to export a pandas dataframe into a xlsx excel sheet

Installing pip install ong_utils[jwt]:

• functions to decode access tokens

General usage

Simple example of an init.py in a package ("mypackage") using ong_utils:

import pandas as pd
from ong_utils import OngConfig, LOCAL_TZ, OngTimer
_cfg = OngConfig("mypackage")
config = _cfg.config
logger = _cfg.logger

And usage in other parts of the code of "mypackage":

import pandas as pd
from mypackage import config, logger, LOCAL_TZ, http
local_now = pd.Timestamp.now(tz=LOCAL_TZ)
res = http.request("GET", config('url'))
logger.info("Sample log")

Configuration files

Config files are yaml/json files located (by default) in ~/.config/ongpi/{project_name}.{extension}. The file extension can be yaml, yml, json or js. File can have this form:

my_project:
  sample_key1: sample_value1
  sample_key2: sample_value2
log:
  optional_log_config: values

or this

{
  "my_project": {
    "sample_key1": sample_value1,
    "sample_key2": "sample_value2"
  },
  "log": {
    "optional_log_config": value
  }
}

Config files must have a section with the name the project. OngConfig.config("key") will raise an Exception if "key" is undefined, unless a default value is used (OngConfig.config("key", "default_value"))
If a file_path is supplied as second argument to the constructor, that file will be used. In that case many projects can share the same config file. E.g. if config file is /etc/myconfig.yaml, ti can has this form:

project_name1:
  key_for_project1: value
project_name2:
  key_for_project2: value

and config method can only access to configuration of current project.

New values can be added to the configuration in execution time by calling add_app_config. That will persist the new values in the configuration file.

Passwords

Module uses keyring to store passwords

from ong_utils import OngConfig
# configuration file should have "service" and "user" keys
_cfg = OngConfig("mypackage")
config = _cfg.config
get_password = _cfg.get_password
set_password = _cfg.set_password

# Sets password (prompts user)
set_password("service", "user")
# Equivalent to keyring.set_password(config("service"), config("user"), input())
# Gets password
pwd = get_password("service", "user")
# Equivalent to keyring.get_password(config("service"), config("user"))

Storing arbitrary data in keyring

For storing long passwords (e.g. a jwt_token) or non string data (e.g. a dictionary of cookies), use ong_utils.InternalStorage class.

from ong_utils import InternalStorage

internal_storage = InternalStorage("your app name")
for value to store in [
        "long string" * 120,
        {"an": "example", "of": "dictionary"}
        ]:
  key_name = "sample_key"
  internal_storage.store_value(key_name, value)
  stored = internal_storage.get_value("key name")
  assert value == stored
  internal_storage.remove_stored_value()

Timers

OngTimer class uses tic(msg) to start timer and toc(msg) to stop timer and show a message with the elapsed time. Several timers can be created with different msg. The parameter msg is used to link methods tic(msg) and toc(msg), so the time is measured from tic to toc.Before a toc there must be a tic, so if there is not a tic(msg) with the same msg as a toc(msg) an exception is risen. Additionally, it can be used as a context manager

Example Usage:

    from ong_utils import OngTimer
    from time import sleep

    #########################################################################################################
    # Standard use (defining an instance and using tic, toc and toc_loop methods, changing decimal places)
    #########################################################################################################
    tic = OngTimer()  # if used OngTimer(False), all prints would be disabled
    more_precise_tic = OngTimer(decimal_places=6)     # Use decimals parameter to increase decimals (defaults to 3)

    tic.tic("Starting")
    more_precise_tic.tic("Starting (6 decimals)")
    for i in range(10):
        tic.tic("Without loop")
        sleep(0.15)
        tic.toc("Without loop")
        tic.tic("Loop")
        sleep(0.1)
        if i != 5:
            tic.toc_loop("Loop")  # Will print elapsed time up to iter #5
        else:
            tic.toc("Loop")  # Will print in this case
    sleep(1)
    tic.print_loop("Loop")  # Forces print In any case it would be printed in destruction of tic instance
    tic.toc("Starting")  # Will print total time of the whole loop
    more_precise_tic.toc("Starting (6 decimals)")  # Will print total time with 6 decimals

    ########################################################################################
    # Using toc/toc_loop with a non previously defined msg will raise a ValueError Exception
    ########################################################################################
    try:
        tic.toc("This msg has not been defined in a previous tick so ValueError Exception will be risen")
    except ValueError as ve:
        print(ve)

    #############################################################
    # Use as a context manager. Won't work accumulating in a loop
    #############################################################
    with OngTimer(msg="Testing sleep"):
        print("hello context manager")
        sleep(0.27)
    with OngTimer().context_manager("Testing sleep"):  # Exactly same as above
        print("hello context manager")
        sleep(0.27)
    # Use context manager (but testing that it can be disabled)
    with OngTimer(msg="Testing sleep disabled", enabled=False):
        print("hello disabled context manager")
        sleep(0.22)
    # use global timer as context manager
    existing_instance = OngTimer()
    with existing_instance.context_manager("Example using an existing context manager instance"):
        sleep(.19)

    # Optionally: write also tick using a logger
    import logging
    logging.basicConfig(level=logging.DEBUG)
    with OngTimer(msg="Using a logger", logger=logging, log_level=logging.DEBUG):
        sleep(0.2)

    ##############################################################
    # When a timer is deleted, any tic without toc will be printed
    ##############################################################
    forgoten_toc_timer = OngTimer()             # This timer will have tics without corresponding toc
    standard_timer = OngTimer(decimals=6)
    forgoten_toc_timer_disabled = OngTimer(enabled=False)
    forgoten_toc_timer.tic("forgotten timer1")
    forgoten_toc_timer.tic("forgotten timer2")
    standard_timer.tic("unforgotten timer")
    forgoten_toc_timer_disabled.tic("forgotten disabled timer")
    sleep(0.1)
    standard_timer.toc("unforgotten timer")
    del forgoten_toc_timer   # Will print elapsed time, as are pending tocs
    del standard_timer   # Prints nothing (as there is not pending tic)
    del forgoten_toc_timer_disabled     # Prints nothing (is disabled)

    #####################################################
    # Use .msgs property to iterate over all named timers
    #####################################################
    loop_timer = OngTimer()
    for _ in range(10):
        loop_timer.tic("hello1")
        loop_timer.tic("hello2")
        sleep(0.1)
        loop_timer.toc_loop("hello1")
        loop_timer.toc_loop("hello2")
    for msg in loop_timer.msgs:
        loop_timer.print_loop(msg)

Urllib3 utils

Module ong_utils.urllib3 includes simple functions to treat cookies in urllib3.

Example:

from ong_utils import create_pool_manager, cookies2header, get_cookies
url = "whichevervalidurl"
http = create_pool_manager()    # Creates a PoolManager with retries
req = http.request("get", url)
# get cookies (as a dict)
cookies = get_cookies(req)
headers = {"Accept": "text/html;application/json"}
# append cookies to the headers dict
headers.update(cookies2header(cookies))
req.http.request("get", url, headers=headers)       # Using cookies from previous response

Make shortcuts for entry points

You can create desktop shortcuts for each entry point in the script to easily launch them in your system.

You have to install optional dependency pip install ong_utils[shortcuts]

NOTE: for the shortcut to work, each entry point defined in e.g. script_file of package package must be executable with python -m package.script_file

There are two ways to create shortcuts:

• Create shortcuts when installing with pip: valid when installing from git (e.g. pip install git+https://github.com/someone/somerepo.git). Uses a custom postinstall script that creates the desktop launcher and modifies the wheel file (so they can be uninstalled) after building the wheel file from sources.
• Create a script and run it manually after install: valid for any other case. Create a entry_point e.g. post_install and call it manually after installation. That scritp will create the shortcut(s) and add it/them to the RECORDfile so the shortcut will be later uninstalled

Create the shortcut when installing with PIP

Create a setup.py in your root directory and add the following code:

from setuptools import setup
from ong_utils.desktop_shortcut import PipCreateShortcut

setup(cmdclass={'bdist_wheel': PipCreateShortcut})

In your pyproject.toml add the following:

[build-system]
requires = [
    "setuptools",
    "wheel",
    "ong_utils[shortcuts]"
]
[project.scripts]
script1 = "package.file:function"

Then the program will install the wheel from pip and create a desktop shortcut for script1.

Create a manual script

Provided that you have the following entry point in your pyproject.toml:

[project.scripts]
script1 = "mypackage.myscript:myfunction"

You'll have to create a script in your code. Let's call it post_install.py with the following content:

from ong_utils.desktop_shortcut import PostInstallCreateShortcut


def main():
    PostInstallCreateShortcut("your_library_name").make_shortcuts()


if __name__ == '__main__':
    main()

Assuming that post_install.py script is in the mypackage folder, then you have to ask the user to run it manually after installation:

pip install mypackage
python -m mypackage.post_install

NOTE: optionally, you can add icons to the shorcut with png format or icns format (for mac), provided that the icons have the same name as the entry_point. The program will use the first icon that matches the name of the entry point.

Nicer output pandas DataFrame to Excel

You can export a pandas DataFrame to Excel nicely formated (converted to an Excel Table, with autofilter enabled and columns widths autofitted)

Example:

import pandas as pd
from ong_utils import df_to_excel

with pd.ExcelWriter(filename) as writer:
    df_to_excel(df, writer, sheet_name)

Decoding jwt tokens

Needs install extra packages with pip install ong_utils[jwt]

Use ong_utils.decode_jwt_token to decode a jwt token into a dict.

Use ong_utils.decode_jwt_token_expiry to decode expiration as a datetime object.