wipac-dev-tools

Common, basic, and reusable development tools


Keywords
python, tools, utilities, WIPAC, IceCube
License
MIT
Install
pip install wipac-dev-tools==1.9.1

Documentation

PyPI GitHub release (latest by date including pre-releases) PyPI - License Lines of code GitHub issues GitHub pull requests

wipac-dev-tools

Common, basic, and reusable development tools

Utilities

Logging Tools

wipac_dev_tools.logging_tools.set_level()

Available for Python 3.6+

def set_level(
    level: LoggerLevel,
    first_party_loggers: Union[
        None, LogggerObjectOrName, List[LogggerObjectOrName]
    ] = None,
    third_party_level: LoggerLevel = "WARNING",
    future_third_parties: Union[None, str, List[str]] = None,
    specialty_loggers: Optional[Dict[LogggerObjectOrName, LoggerLevel]] = None,
    use_coloredlogs: bool = False,
) -> None:
    """Set the level of loggers of various precedence.

    The root logger and first-party logger(s) are set to the same level (`level`).

    Args:
        `level`
            the desired logging level (first-party), case-insensitive
        `first_party_loggers`
            a list (or a single instance) of `logging.Logger` or the loggers' names
        `third_party_level`
            the desired logging level for any other (currently) available loggers, case-insensitive
        `future_third_parties`
            additional third party logger(s) which have not yet been created (at call time)
        `specialty_loggers`
            additional loggers, each paired with a logging level, which are not
            considered first-party nor third-party loggers. **These have the highest precedence**
        `use_coloredlogs`
            if True, will import and use the `coloredlogs` package.
            This will set the logger format and use colored text.
    """

wipac_dev_tools.logging_tools.log_argparse_args()

Available for Python 3.6+

def log_argparse_args(
    args: argparse.Namespace,
    logger: Union[None, str, logging.Logger] = None,
    level: LoggerLevel = "WARNING",
) -> argparse.Namespace:
    """Log the argparse args and their values at the given level.

    Return the args (Namespace) unchanged.

    Example:
        2022-05-13 22:37:21 fv-az136-643 my-logs[61] WARNING in_file: in_msg.pkl
        2022-05-13 22:37:21 fv-az136-643 my-logs[61] WARNING out_file: out_msg.pkl
        2022-05-13 22:37:21 fv-az136-643 my-logs[61] WARNING log: DEBUG
        2022-05-13 22:37:21 fv-az136-643 my-logs[61] WARNING log_third_party: WARNING
    """

Environment Variable Tool(s)

wipac_dev_tools.from_environment()

Available for Python 3.6+

def from_environment(keys: KeySpec) -> Dict[str, RetVal]:
    """Obtain configuration values from the OS environment.

    Parsing Details:
    Types are inferred from the default values, and casted as such:
    `bool`: *(case-insensitive)*:
        - `True`  => ("y", "yes", "t", "true", "on", or "1")
        - `False` => ("n", "no", "f", "false", "off", or "0")
        - `Error` => any other string
    `int`: normal cast (`int(str)`)
    `float`: normal cast (`float(str)`)
    `other`: no change (`str`)

    Arguments:
        keys - Specify the configuration values to obtain.

               This can be a string, specifying a single key, such as:

                   config_dict = from_environment("LANGUAGE")

               This can be a list of strings, specifying multiple keys,
               such as:

                   config_dict = from_environment(["HOME", "LANGUAGE"])

               This can be a dictionary that provides some default values,
               and will accept overrides from the environment:

                   default_config = {
                       "HOST": "localhost",
                       "PORT": 8080,
                       "REQUIRED_FROM_ENVIRONMENT": None
                   }
                   config_dict = from_environment(default_config)

               Note in this case that if 'HOST' or 'PORT' were defined in the
               environment, those values would be returned in config_dict. If
               the values were not defined in the environment, the default values
               from default_config would be returned in config_dict.

               Also note, that if 'REQUIRED_FROM_ENVIRONMENT' is not defined,
               an OSError will be raised. The sentinel value of None indicates
               that the configuration parameter MUST be sourced from the
               environment.

    Returns:
        a dictionary mapping configuration keys to configuration values

    Raises:
        OSError - If a configuration value is requested and no default
                  value is provided (via a dict), to indicate that the
                  component's configuration is incomplete due to missing
                  data from the OS.
        ValueError - If a type-indicated value is not a legal value
    """

wipac_dev_tools.from_environment_as_dataclass()

Available for Python 3.7+

def from_environment_as_dataclass(
    dclass: Type[T],
    collection_sep: Optional[str] = None,
    dict_kv_joiner: str = "=",
    log_vars: Optional[logging_tools.LoggerLevel] = "WARNING",
) -> T:
    """Obtain configuration values from the OS environment formatted in a
    dataclass.

    Environment variables are matched to a dataclass field's name. The
    matching environment string is cast using the dataclass field's type
    (there are some special cases for built-in types, see below). Then,
    the values are used to create a dataclass instance. All normal
    dataclass init-behavior is expected, like required fields
    (positional arguments), optional fields with defaults, default
    factories, post-init processing, etc.

    If a field's type is a bool, `wipac_dev_tools.strtobool` is applied.

    If a field's type is a `list`, `dict`, `set`, `frozenset`, or
    an analogous type alias from the 'typing' module, then a conversion
    is made (see `collection_sep` and `dict_kv_joiner`). Sub-types
    are cast if using a typing-module type alias. The typing-module's
    alias types must resolve to `type` within 1 nesting (eg: List[bool]
    and Dict[int, float] are okay; List[Dict[int, float]] is not), or
    2 if using 'Final' or 'Optional' (ex: Final[Dict[int, float]]).

    If a field's type is a class that accepts 1 argument, it is
    instantiated as such.

    Arguments:
        dclass - a (non-instantiated) dataclass, aka a type
        collection_sep - the delimiter to split collections on ("1 2 5")
        dict_kv_joiner - the delimiter that joins key-value pairs ("a=1 b=2 c=1")
        log_vars - what level to log the collected environment variables (set to `None` to not log)

    Returns:
        a dataclass instance mapping configuration keys to configuration values

    Example:
        env:
            FPATH=/home/example/path
            PORT=9999
            HOST=localhost
            MSGS_PER_CLIENTS=alpha=0 beta=55 delta=3
            USE_EVEN=22
            RETRIES=3

        python:
            @dataclasses.dataclass(frozen=True)
            class Config:
                FPATH: pathlib.Path
                PORT: int
                HOST: str
                MSGS_PER_CLIENTS: Dict[str, int]
                USE_EVEN: EvenState
                RETRIES: Optional[int] = None
                TIMEOUT: int = 30

                def __post_init__(self) -> None:
                    if self.PORT <= 0:
                        raise ValueError("'PORT' is non-positive")

            class EvenState:
                def __init__(self, arg: str):
                    self.is_even = not bool(int(arg) % 2)  # 1%2 -> 1 -> T -> F
                def __repr__(self) -> str:
                    return f"EvenState(is_even={self.is_even})"

            config = from_environment_as_dataclass(Config)
            print(config)

        stdout:
            Config(
                FPATH=PosixPath('/home/example/path'),
                PORT=9999,
                HOST='localhost',
                MSGS_PER_CLIENTS={'alpha': 0, 'beta': 55, 'delta': 3},
                USE_EVEN=EvenState(is_even=True),
                RETRIES=3,
                TIMEOUT=30)


    Raises:
        OSError - If a configuration value is requested and no default
                  value is provided, to indicate that the component's
                  configuration is incomplete due to missing data from
                  the OS.
        ValueError - If an indicated value is not a legal value
        TypeError - If an argument or indicated value is not a legal type
    """

wipac_dev_tools.strtobool()

def strtobool(string: str) -> bool:
    """Smart-cast a string to a bool using common-sense interpretations.

    Unlike the since deprecated `distutils.util.strtobool`, this
    returns an actual bool.

    True: 'y', 'yes', 't', 'true', 'on', '1'
    False: 'n', 'no', 'f', 'false', 'off', '0'

    Raises:
        ValueError: if the string does not match any of the about
    """