More filters

This release adds filters for selecting specific data series based on
criteria and also a filter to calculate basic transformations on
individual quantities.

Furthermore, Matplotlib is fixed to versions below 3.5 because 3.6
breaks polt currently. This is not a long-term fix though.

MacOS Fixes and a MosquittoParser

This release fixes polt for MacOS and adds a simple parser for
Mosquitto's cli client `mosquitto_sub`.

Changelog in the Sphinx Documenatation

From this release on, a changelog generated from git version tags is
available in the Sphinx documentation.

Keyboard Shortcut for Pretty Xaxis Date Format

This release adds a keyboard shortcut 'x' to the `TimeLinesAnimator` to
run autofmt_xdate.

Fix automated PyPI upload

This release fixes the automated package upload to PyPI. pyOpenSSL was
not installed in the runner.

Tight Layout Keyboard Shurtcut

This release adds a keyboard shortcut 't' to run
matplotlib.pyplot.tight_layout() manually.

Introduce Filters

This release introduces the concept of Filters for modifying incoming
data. With custom Filters it is possible to adjust (e.g. smooth) data or
even to add new derived data (e.g.  the maximum, a derivative, an
integral, etc...).

A slight speedup of the SpectrumAnimator was achieved by using scipy's
FFT implementation, not numpy's. Furthermore, displaying the current FPS
is now possible for some Animators.

Structural Improvements

This release reduces duplicate code and improves the overall code
structure. Furthermore, CI tests are now performed for Python versions
back to 3.4.

Live Histograms and Spectra and Data Generation

This release adds a lot of useful new features:

- Live Histograms
- Live FFTs (spectra)
- Data Generation Tool

Pausing, Custom Animators, Rate Limiting and More Documentation

This release introduces some structural changes that allow for using
custom Animators (#14), which can now be specified similar to the custom
parsers.

Furthermore, the incoming data rate can now be limited per streamer (#27)
which can be handy when the plotting can't catch up the the input.

Also, the animation can now be paused (and resumed) by hitting the
spacebar. During pause, the graph can be moved, zoomed and modified with
the standard Matplotlib toolbar without further data updates.

The documentation has recieved a big update with a clearer structure and
more information on various topics.

Prettier Subplots Arrangement

Bugs Fixed:

- #44: Repeated values were deleted when using --extrapolate

New Features:

- #45: Option for automatic tight_layout(), i.e. less whitespace
- #37: Column blacklist for CsvParser

Improved Timeseries Plotting

New Features:

- #4: make the temporal time frame configurable
- #41: clear the buffer at each frame, better preventing indefinite
  buffer growth
- #23: enable sharing time axis between subplots
- #26: use autofmt_xdate to display time axis more readable
- #30, #31: proper documentation (deployed to GitLab Pages)

Multiple Improvements

Fixed issues:

- #35: add-source command failed due to importlib not being imported
- #36: csv parsing failed when using a command as input

New Features:

- #34: setuptools' entry_points can now be used to add parsers do polt
- #32: the csv delimiter used by CsvParser can now be specified via cli and the
       configuration
- #36, #39: an --encoding option exists now for the add-source command to
            specify the input encoding for parsers

Improvements:

- #40: the left-free subplot area is now minimized more aggressively

Code style:

- #33: parsers now reside in separate modules

Improved Command-Line Interface

- use click as command-line parser
- use console_scripts entry point to make cli available via a simple `polt`

Improve Parsers

This release improves the available parsers for numbers and CSV input.
Several bugs were also fixed.

Initial Version

This is the initial release with a basic set of features:

- reading numbers from multiple sources
- rudimentary timeseries plotting