Trading-strategy

- Fix `get_by_chain_and_factory()` was returning `ExchangeNotFoundError` instead of raising it

Binance data downloading added. Python 3.11, Python 3.12 and Pandas 2.x compatibility fixes.

- Add: `BinanceDownloader.fetch_approx_asset_trading_start_date()` to figure out when asset was listed on Binance
- Fix: `BinanceDownloader` timestamps to be UTC
- Add: Example script how to download Binance data for multiple assets and write in a single Parquet file
- Upgrade to support Python 3.12 with breaking `datetime.datetime.utcnow()` changes
- Upgrade to support Pandas 2.x with breaking `pd.Timestamp.utcnow()` changes
- Upgrade to web3-ethereum-defi latest with Python 3.12 compatibilty
- Upgrade to `pyarrow` 14.0 for Python 3.12 compatibility
- `datetime.datetime.utcnow()` and `datetime.datetime.utcfromtimestamp` workarounds - we do not need timezones that unnecessarily bloat data size
- Add `naive_utcnow` for Python 3.12 compatibility
- Add `naive_utcfromtimestamp` for Python 3.12 compatibility
- Change from `pd.Timestamp.utcfromtimestamp()` to `pd.Timestamp.utcfromtimestamp().tz_localize(None)` when needed
-

- Fix `estimate_accrued_interest` crashing when there are gaps in data
- Fix `resample_candles` to retain `pair_id` column if one is present in data

- Add functionality for downloading and manipulating Binance candle and lending data
to have alternative datasets to benchmark the strategies for overfit and data quality issues
- Add `resample_series()` method to work for single column pd.Series price data
- Fix `CandleUniverse.get_candles_by_pair()` (add return value)
- Fix `resample_candles()` method

- API update: `CandleUniverse.get_candles_by_pair()` accepts `DEXPair` besides `pair_id` as an argument
- `forward_fill()` can handle `volume` column

- Fix: DataFrame copy warning in `estimate_accrued_interest`
- Add: `get_price_with_tolerance(pair_name_hint)` to have better exception messages
- Fix: `CandleSampleUnavailable` exception message was tuple instead of newline separated string