Star Digital Financial Data

Problem it addresses

When building pipelines for equity, fund, convertible bond, futures, and option data, teams often face scattered sources, inconsistent definitions, and repeated historical pulls. AmazingData wraps financial data from China Galaxy Securities into a Python API, using BaseData, MarketData, and InfoData to separate reference data, market data, and fundamentals, which fits notebook, backtesting, and factor work.

Core objects and workflow

• Three objects: ad.BaseData() covers securities code lists, trading calendars, adjustment factors, and ETF creation/redemption lists; ad.MarketData(calendar) covers historical K-lines and snapshots and requires a calendar; ad.InfoData() covers financial statements, earnings notices, shareholders, margin trading, dragon-tiger lists, options, ETFs, indexes, convertible bonds, bonds, and announcements.
• Parameters and returns: dates usually use eight-digit YYYYMMDD integers, securities codes use exchange-suffixed formats such as 000001.SZ and 600000.SH; most endpoints return pandas.DataFrame, while some return a dict keyed by code.
• Fetching modes: local caching uses local_path plus is_local to read local HDF5 first and refill from the server when missing; date-range mode uses begin_date plus end_date and fetches from the server without local caching.
• Authentication: set AD_USERNAME, AD_PASSWORD, AD_HOST, and AD_PORT, then call ad.login() before querying; credentials and server details are usually issued by the account-opening branch.

Boundaries

• Historical depth: equities trace back to 2013, futures to 2010, and options to 2015.
• MarketData must be initialized with a trading calendar rather than empty arguments.
• Local caching and date-range fetching are mutually exclusive; local caching may require substantial disk space.
• The service limits single active login, so one account can maintain only one concurrent session.