A very simple fast fetcher for Binance historical market data.

By using a combination of historical CSV files and API calls this tiny library can avoid rate limits and fetch data significantly faster than alternatives that fetch just from the API.

here is a benchmark for you :)

Downloaded tickers	Time range	HistoricalBinance	Freqtrade
BTC/USDT:USDT, ETH/USDT:USDT	20230201-20240315	5.139s	5min 11.78s
BTC/USDT:USDT, ETH/USDT:USDT, OP/USDT:USDT, FET/USDT:USDT, SOL/USDT:USDT, RUNE/USDT:USDT	20230201-20240315	9.68s	14min 28.22s

Usage

The library includes two classes - BinanceDataDownloader and BinanceDataDownloader

BinanceDataDownloader is a very simple class that just fetches the available tickers to download and upon requesting the download_one_ticker(ticker, start_date, end_date, timeframe) function will return a polars dataframe requested. If necessary it can be easily converted to pandas.

 downloader = BinanceDataDownloader()
 data = await downloader.download_one_ticker("BTCUSDT", datetime(2020, 1, 1), datetime(2024, 3, 10), "1d")

BinanceDataProvider is served as an example of how to effectively use the library - caching and asynchronously fetching multiple tickers.

Here is an example of how to use BinanceDataProvider:

from binance_data import BinanceDataProvider 
 
async def binance_data_provider_example(self):
    provider = BinanceDataProvider(["ETH/USDT:USDT", "BTC/USDT:USDT", "FET/USDT:USDT", "RUNE/USDT:USDT", "SOL/USDT:USDT", "OP/USDT:USDT"], ["1m"])
    await provider.update_tickers_async(["ETH/USDT:USDT", "BTC/USDT:USDT", "FET/USDT:USDT", "RUNE/USDT:USDT", "SOL/USDT:USDT", "OP/USDT:USDT"], ["1m"], datetime(2023, 2, 1))
    print(provider.cached_dataframes["1m"]["ETHUSDT"])

This will print out a polars dataframe with columns 'date', 'open', 'high', 'low', 'close', 'volume', 'count', 'taker_buy_volume', however if you wish you can modify the code to keep the 'ignore', 'taker_buy_quote_volume' and 'quote_volume' fetched from Binance.