Technical

What's Changed

* Don't fillna whole dataframe as part of indicator calculation
* simplify usage for `vwmacd` and `td_sequential` (they return the original dataframe with new columns now).


**Full Changelog**: https://github.com/freqtrade/technical/compare/1.4.2...1.4.3

Add compatibility alias for zema (will raise a warning when called for the first time).

Mostly a maintenance release, increasing compatibility with uptodate dependencies.

* Remove zema, as it's implementation was actually a dema

New indicators:
* `tv_wma`, `tv_hma`, `chopiness`.
* disable `qtpylib.vwap()` as that's implicitly forward-looking. Use `qtpylib.rolling_vwap()` instead.
* fix some bugs in qtpylib, causing certain cases to fail.
* update master-branch to main branch.
* Remove some unused dependencies (arrow), which will improve compatibility with more systems
*

Convert the project to use `pyproject.toml`.
Use the black formatter to format code (so the code-style is somewhat aligned throughout the repository)

Improve resample / merge logic to use data as soon as it'll become available (as soon as the long candle closes).

We did some maintenance on technical - fixing some bugs and cleaning up some dependencies.

Due to this, some custom implementations (mainly indicators wrapping pyti) have been removed, as these indicators are very slow to calulate (due to the way pyti does their calculations).

This removes the following indicators (all are replaced by their ta-lib equivalent)
* `indicators.aroon` -> `ta.AROON`
* `indicators.cci` -> `ta.CCI`
* `indicators.cmo` -> `ta.CMO`
* `indicators.momentum` -> `ta.MOM`
* `indicators.ultimate_oscilator` -> `ta.ULTOSC`
* `indicators.accumulation_distribution` -> `ta.AD`

The above results in significant speed improvements for these indicators, as well as for all Consensus methods relying on the above methods (`Consensus`, `MovingAverageConsensus`, `OscillatorConsensus`).

:warning: **WARNING** Some of the above changes will now return a slightly different result, so make sure to reevaluate your strategy / backtest your strategy again before upgrading to this version.