Metadata-Version: 2.3
Name: tremors
Version: 0.7.0
Summary: Tremors is a library for logging while collecting metrics.
Keywords: logging,log,logger,metrics
Author: Narvin Singh
Author-email: Narvin Singh <Narvin.A.Singh@gmail.com>
License: Tremors is a library for logging with metrics.
         Copyright (C) 2025 Narvin Singh
         
         This program is free software: you can redistribute it and/or modify
         it under the terms of the GNU General Public License as published by
         the Free Software Foundation, either version 3 of the License, or
         (at your option) any later version.
         
         This program is distributed in the hope that it will be useful,
         but WITHOUT ANY WARRANTY; without even the implied warranty of
         MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
         GNU General Public License for more details.
         
         You should have received a copy of the GNU General Public License
         along with this program.  If not, see <https://www.gnu.org/licenses/>.
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)
Classifier: Operating System :: OS Independent
Requires-Dist: python-docs-theme~=2025.12 ; extra == 'doc'
Requires-Dist: sphinx~=9.1 ; extra == 'doc'
Requires-Python: >=3.12
Project-URL: Homepage, https://tremors.readthedocs.io
Project-URL: Documentation, https://tremors.readthedocs.io
Project-URL: Repository, https://codeberg.org/narvin/tremors
Project-URL: Issues, https://codeberg.org/narvin/tremors/issues
Provides-Extra: doc
Description-Content-Type: text/x-rst

Tremors
#######

Tremors is a library for logging while collecting metrics. Tremors loggers are
drop-in replacements for standard loggers. But Tremors loggers have metrics
collectors that run when messages are logged. The loggers are also context
managers. The library maintains a hierarchy of nested contexts, where all
logs and metrics are grouped together. You can create a new hierarchy at
anytime to group related logs.

Installation
************

.. code-block:: shell

   pip install tremors

Usage
*****

A function can be wrapped in a logger context with the ``logged`` decorator. If
you call the function without a logger argument, one will automatically be
injected into it.

.. code-block:: python

    import logging

    import tremors
    from tremors import collector


    @tremors.logged
    def fn(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.info("hello")


    logging.basicConfig(
        format="Tremors > %(levelname)s:%(name)s:%(message)s",
        level=logging.INFO,
    )
    fn()

The context automatically logs ``entered``, and ``exited`` messages before,
and after each function call. The logger uses the configured standard root
logger by default to log the messages.

.. code-block:: shell

    Tremors > INFO:root:entered: fn
    Tremors > INFO:root:hello
    Tremors > INFO:root:exited: fn

You may specify a standard logger by name for the Tremors logger to use as
its underlying logger.

.. code-block:: python

    @tremors.logged(logger_name=__name__)
    def fn2(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.info("hello")


    fn2()

The messages are logged by the specified underlying logger. Based on our
standard logging configuration, the messages propagate from the underlying
logger to the standard root logger, which emits them.

.. code-block:: shell

    Tremors > INFO:__main__:entered: fn2
    Tremors > INFO:__main__:hello
    Tremors > INFO:__main__:exited: fn2

Next let's use a collector to measure the elapsed time since the function
started each time a message is logged. When a message is logged, the logger
runs the collector, and adds its updated state to the message's LogRecord. We
use a standard logging filter to inspect and modify the record before it is
emitted. We format the collector state, then add the formatted state to the
``elapsed`` custom attribute of the record. Finally, we configure the root
logger's formatter to incorporate the elapsed attribute.

.. note::

    The elapsed collector bundle included with Tremors has a factory for
    creating a collector. It also has a formatter that we use in ``flt``
    to extract the state from the record, and format it. In ``flt`` we
    make a copy of the record, then modify and return the copy, instead of
    modifying the original record, so as not to have side effects on other
    loggers that may process the messasge. We also make sure to attach
    ``flt`` to the root logger's handler, and not to the logger itself;
    messages that originate from descendant loggers will not go through
    logger filters when they are propagated, but they will go through handler
    filters before they are emitted.

.. code-block:: python

    import copy
    import time


    def flt(record: logging.LogRecord) -> logging.LogRecord:
        record = copy.copy(record)
        elapsed = collector.elapsed.formatter(record)
        record.elapsed = f"{elapsed} " if elapsed else ""
        return record


    @tremors.logged(collector.elapsed.factory())
    def fn3(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.info("sleeping for 1s...")
        time.sleep(1)


    logging.basicConfig(
        format="%(elapsed)s%(levelname)s:%(name)s:%(message)s",
        level=logging.INFO,
        force=True,
    )
    logging.root.handlers[0].addFilter(flt)
    fn3()

The messages contain elapsed information, according to the formatter
configuration, that is sourced from the record's elapsed custom attribute.

.. code-block:: shell

    0.000 INFO:root:entered: fn3
    0.000 INFO:root:sleeping for 1s...
    1.000 INFO:root:exited: fn3

A Logger can have any number of collectors. Here, in addition to the elapsed
collector from the previous example, we add a counter collector. A collector
has a level, and will only run if the message is being logged at that level or
higher. Our counter level is ``ERROR``. We can also control which custom record
attribute has the formatted collector state via the collector's name. This is
useful if you have multiple of the same collector on a single logger. Here,
we name the counter ``errors``, so ``record.errors`` will contain a formatted
string with the running total number of errors that have been logged by a
single function call. Finally, we an control the format of the counter state
via the ``fmt`` argument of the counter's formatter.

.. code-block:: python

    def flt2(record: logging.LogRecord) -> logging.LogRecord:
        record = copy.copy(record)
        errors = collector.counter.formatter(
            record, name="errors", fmt="errors={counter}"
        )
        record.errors = f"{errors} " if errors else ""
        elapsed = collector.elapsed.formatter(record)
        record.elapsed = f"{elapsed} " if elapsed else ""
        return record


    @tremors.logged(
        collector.elapsed.factory(),
        collector.counter.factory(name="errors", level=logging.ERROR),
    )
    def fn4(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.info("hello")
        time.sleep(1)
        logger.error("uh-ho!")


    logging.basicConfig(
        format="%(elapsed)s%(errors)s%(levelname)s:%(name)s:%(message)s",
        level=logging.INFO,
        force=True,
    )
    logging.root.handlers[0].addFilter(flt2)
    fn4()

The messages contain information from both collectors.

.. code-block:: shell

    0.000 errors=0 INFO:root:entered: fn4
    0.000 errors=0 INFO:root:hello
    1.001 errors=1 ERROR:root:uh-ho!
    1.001 errors=1 INFO:root:exited: fn4

Passing a collector factory, as in the previous example, will result in a new
counter collector being used each time the function is called. Let's reuse the
same collector to keep a tally of errors across *all* calls to the function by
passing a collector instance that we get by calling the factory's ``create``
method.

.. code-block:: python

    fn_errors = collector.counter.factory(
        name="errors", level=logging.ERROR
    ).create()


    @tremors.logged(fn_errors)
    def fn5(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.error("uh-ho!")


    fn5()
    fn5()

The error count doesn't reset in the second function call.

.. code-block:: shell

    errors=0 INFO:root:entered: fn5
    errors=1 ERROR:root:uh-ho!
    errors=1 INFO:root:exited: fn5
    errors=1 INFO:root:entered: fn5
    errors=2 ERROR:root:uh-ho!
    errors=2 INFO:root:exited: fn5

Another way we can tally the count across all function calls is to pass the
same logger with each call.

.. code-block:: python

    def fn6(*, logger: tremors.Logger) -> None:
        logger.error("uh-ho!")


    with tremors.Logger(
        collector.counter.factory(name="errors", level=logging.ERROR),
        name="context",
    ) as logger:
        fn6(logger=logger)
        fn6(logger=logger)

We only get entered and exited messages for the context block. But the single
logger used in both function calls maintains its state between calls.

.. code-block:: shell

    errors=0 INFO:root:entered: context
    errors=1 ERROR:root:uh-ho!
    errors=2 ERROR:root:uh-ho!
    errors=2 INFO:root:exited: context

Collectors may be inherited by descendant loggers. Let's count errors across
nested loggers.

.. code-block:: python

    @tremors.logged(
        collector.counter.factory(
            name="errors", level=logging.ERROR, inherit=True
        ),
        enter_msg=False,
        exit_msg=False,
    )
    def parent(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.error("uh-ho!")
        child()
        child()


    @tremors.logged(enter_msg=False, exit_msg=False)
    def child(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.error("doh!")
        grandchild()


    @tremors.logged(enter_msg=False, exit_msg=False)
    def grandchild(*, logger: tremors.Logger = tremors.from_logged) -> None:
        logger.info("so far, so good")
        logger.error("spoke too soon!")


    parent()

We've disabled the entered and exited messages with the ``enter_msg`` and
``exit_msg`` parameters. The ``parent`` counter is used in the ``child`` and
``grandchild`` functions.

.. code-block:: shell

    errors=1 ERROR:root:uh-ho!
    errors=2 ERROR:root:doh!
    errors=2 INFO:root:so far, so good
    errors=3 ERROR:root:spoke too soon!
    errors=4 ERROR:root:doh!
    errors=4 INFO:root:so far, so good
    errors=5 ERROR:root:spoke too soon!

Asynchronous functions and methods may be decorated with ``async_logged``.

.. code-block:: python

    import asyncio


    @tremors.async_logged(collector.elapsed.factory())
    async def async_work_long(
        *, logger: tremors.Logger = tremors.from_logged
    ) -> None:
        logger.info("long work starting")
        await asyncio.sleep(1)


    @tremors.async_logged(collector.elapsed.factory())
    async def async_work_short(
        *, logger: tremors.Logger = tremors.from_logged
    ) -> None:
        logger.info("short work starting")
        await asyncio.sleep(0.1)


    @tremors.async_logged(collector.elapsed.factory())
    async def async_fn(*, logger: tremors.Logger = tremors.from_logged) -> None:
        coros = async_work_long(), async_work_short()
        logger.info("awaiting %d works", len(coros))
        await asyncio.gather(*coros)


    logging.basicConfig(
        format="%(elapsed)s%(levelname)s:%(name)s:%(message)s",
        level=logging.INFO,
        force=True,
    )
    logging.root.handlers[0].addFilter(flt)
    asyncio.run(async_fn())

.. code-block:: shell

    0.000 INFO:root:entered: async_fn
    0.000 INFO:root:awaiting 2 works
    0.000 INFO:root:entered: async_work_long
    0.000 INFO:root:long work starting
    0.000 INFO:root:entered: async_work_short
    0.000 INFO:root:short work starting
    0.101 INFO:root:exited: async_work_short
    1.002 INFO:root:exited: async_work_long
    1.003 INFO:root:exited: async_fn

See the `collector module`_ in the full `documentation`_ for how you can
define your own collectors, and bundles.

.. _documentation: https://tremors.readthedocs.io/en/latest
.. _collector module: https://tremors.readthedocs.io/en/latest/#module-tremors.collector
