Quickstart

There are mainly 6 ways to interact with data in pfeed:

1. get_historical_data() retrieves stored data from storage (if any) or downloads it from data sources. The data is then cached and returned as a cleaned DataFrame or LazyFrame.

import pfeed as pe
feed = pe.BybitFeed(data_tool='polars', use_ray=True)
df = feed.get_historical_data('BTC_USDT_PERP', rollback_period='1w', resolution='1d')

2. download() saves data locally or to the cloud.

df = feed.download(
    product='BTC_USDT_PERP', 
    resolution='1m',
    start_date='2025-01-01', 
    end_date='2025-01-03', 
    to_storage='local',  # or 'minio', if MinIO is running
)

3. retrieve() loads previously stored data from local or cloud storage.

df = feed.retrieve(
    product='BTC_USDT_PERP', 
    resolution='1m',
    start_date='2025-01-01', 
    end_date='2025-01-03', 
)

4. 🚧 get_realtime_data() fetches real-time data directly from the broker/exchange’s API.

5. 🚧 stream() listens to real-time data via websocket and can optionally store it .

6. 🚧 fetch() pulls data from a data source’s API without storing it.