Abseil Python Common Libraries

Warnings

• breaking Abseil Python has progressively dropped support for older Python versions. Version 2.4.0 dropped support for Python 3.8 and 3.9. Version 2.2.2 removed support for Python 3.7. Version 2.0.0 dropped support for Python 3.6.
  Fix: Ensure your Python environment is version 3.10 or newer to use the latest `absl-py` versions.
• gotcha Directly evaluating `absl.flags.Flag` instances in a boolean context (e.g., `if FLAGS.myflag:`) will raise an error. This prevents accidental usage of the Flag object itself rather than its value.
  Fix: Always access the flag's value using the `.value` attribute: `if FLAGS.myflag.value:`.
• breaking The `absl.logging.exception` function's behavior regarding the `exc_info` argument changed in v2.0.0. Previously, passing `exc_info` could raise a `KeyError`. It now correctly accepts `exc_info` as an argument (with a default value of `True`).
  Fix: If your code explicitly passes `exc_info` to `logging.exception`, ensure `absl-py` is version 2.0.0 or higher. Otherwise, verify behavior on older versions if `exc_info` is relied upon.
• gotcha Flag names defined with `absl.flags.DEFINE_*` are globally registered. This means that if multiple modules or libraries define flags with the same name, a `DuplicateFlagError` will occur upon import or flag definition.
  Fix: Choose unique and descriptive flag names, especially in larger applications or when integrating multiple libraries. For more advanced isolation, consider managing flags with separate `flags.FlagValues` instances, though this is less common for standard usage.

Install

• pip install absl-py Install stable version

Imports

• app

  from absl import app

  Used for managing application lifecycle and parsing command-line flags. The main entry point is typically `app.run(main_function)`.
• flags

  from absl import flags

  Provides functions like `flags.DEFINE_string`, `flags.DEFINE_integer`, and `flags.FLAGS` to define and access command-line flags.
• logging

  from absl import logging

  Offers a custom logging module built on top of Python's standard logging, with features like various verbosity levels and conditional logging.
• absltest

  from absl.testing import absltest

  The base class for Abseil-style Python tests, similar to `unittest.TestCase`.
• flagsaver

  from absl.testing import flagsaver

  A utility for testing code that relies on `absl.flags`, allowing flags to be temporarily overridden for test isolation.

Quickstart

This example demonstrates defining and using command-line flags, basic logging, and integrating with `absl.app.run()` to handle program execution. You can run this script and pass flags like `--greeting_target=Pythonista` or `--repeat_count=5`. The `os.environ` usage is for demonstrating how to programmatically set flags, especially useful in testing contexts where direct command-line arguments might not be feasible.

from absl import app
from absl import flags
from absl import logging
import os

FLAGS = flags.FLAGS

flags.DEFINE_string('greeting_target', 'World', 'The entity to greet.')
flags.DEFINE_integer('repeat_count', 1, 'How many times to repeat the greeting.', lower_bound=1)
flags.DEFINE_boolean('verbose', False, 'If true, print verbose output.')

def main(argv):
  if FLAGS.verbose:
    logging.set_verbosity(logging.DEBUG)
    logging.debug('Verbose mode enabled. Non-flag arguments: %s', argv[1:])
  else:
    logging.set_verbosity(logging.INFO)

  for _ in range(FLAGS.repeat_count):
    logging.info('Hello, %s!', FLAGS.greeting_target)
  
  # Example of an argument passed directly, not as a flag
  if len(argv) > 1:
    logging.warning('Unrecognized positional arguments: %s', argv[1:])

if __name__ == '__main__':
  # To run: python your_script.py --greeting_target=User --repeat_count=3 --verbose
  # Or with environment variable for testing:
  # os.environ['greeting_target'] = 'TestUser'
  # os.environ['repeat_count'] = '2'
  # Then run: python your_script.py --verbose
  app.run(main)