Trading-strategy

- Escape hatch for interactive API key setup loop
- Added `Candle.generate_synthetic_sample` for mocking test data
- Added `GroupedCandleUniverse.get_price_with_tolerance` to safely get the latest price
in backtesting

- Make `matplotlib` optional in `Client.setup_notebook`
- Added `DEXPair.exchange_type` to support Uniswap v3 pairs

- Change: List `requests` as an explicit dependency for the client
(Might have been causing issues with Pyodide builds)

- Change: Remove a lot of dependencies and cleaning up old code.
The package with its dependencies should be now much more compact.
Any old Backtrader or QSTrader code is behind an extra dependency in
the Python package definition. There is no longer dependency to
Ethereum or crypto packages that had C and were problematic
install.
- Change: preparing for Pyodide support.
- Change: Bumped Python 3.10 because 3.10 is minimum for Pyodide

- Fix: `PairGroupedUniverse` did not have OHLCV data sorted by `timestamp`, but instead of was
`pair_id`, `timestamp` causing some failures of managing expectations when accessing data.
- Fix: Update tests to reflect new datasets from Trading Strategy API
- Added `Client.fetch_candles_by_pair_ids` that loads trading data for certain pairs
data over JSONL endpoint and avoids loading large Parquest candle and liquidity files.
This makes it possible to run backtests on low memory environments.
- Added `Client.fetch_candle_dataset` and `read_parquest(filters)`
to filter Parquest files when loading candle datasets to optimise memory usage.
(At the moment this path is not used outside the tests.)
- Change `ChainId` base class to `IntEnum` instead of `Enum` for better type hinting
- Change `PandasPairUniverse.build_index` to use Pandas DataFrame frame transpose,
greatly speeding up the creation of the index
- Fix: `get_closest_price` gets a closest price even if the timestamp and
and time bucket are not in the same internal
- Added `environment/jupyterlite` for the upcoming WebAssembly integration

- Added `is_candle_green()`, `is_candle_red()` helpers
- Added `TimeBucket.to_pandas_timedelta()` helper
- Added `upsample_candles()` helper
- Added `unknown` and `osmosis` blockchains