Improvement: Concurrent Batch Processing for Financials

[thedrunkfishh]

Summary

This PR implements concurrent batch processing for retrieving financial statements (financials, balance_sheet, cash_flow, earnings) in the Tickers class.

Currently, fetching these attributes for a list of tickers happens serially, which is a bottleneck for users analyzing multiple stocks. This change uses concurrent.futures.ThreadPoolExecutor to parallelize these requests.

Changes

• Updated Tickers class (tickers.py):
  • Added _download_attribute helper method.
  • Added public batch methods:
    • download_financials(threads=None)
    • download_balance_sheet(threads=None)
    • download_cash_flow(threads=None)
    • download_earnings(threads=None)
    • download_info(threads=None)

Usage

import yfinance as yf

symbols = "AAPL MSFT GOOG AMZN META"
tickers = yf.Tickers(symbols)

# previously (serial, slower):
# data = {s: yf.Ticker(s).financials for s in symbols.split()}

# now (concurrent, faster):
financials = tickers.download_financials()
# returns dict: {'AAPL': pd.DataFrame, 'MSFT': pd.DataFrame, ...}

balance_sheets = tickers.download_balance_sheet()

Performance Benchmark

1. Benchmarked with 10 popular tickers (AAPL, MSFT, GOOG, AMZN, META, TSLA, NVDA, JPM, V, WMT).

Results:

• Serial Fetching: ~1.81 seconds
• Batch Fetching: ~0.08 seconds
• Speedup: ~22.88x

2. Benchmarked with an isolated, cold-cache environment.

Results:

• Serial Fetching (Cold): ~3.01 seconds
• Batch Fetching (Cold): ~0.54 seconds
• Speedup: ~5.56x

(Note: Speedups can be significantly higher in warm-cache scenarios)

[ValueRaider]

Do you really need to frequently bulk-fetch financials? Try https://pypi.org/project/yfinance-cache