jaymd96-fmp

Python client for the Financial Modeling Prep API with automatic DuckDB caching.

Features

• 150+ endpoints covering quotes, financials, charts, news, SEC filings, insider trades, and more
• DuckDB cache with configurable TTLs per endpoint category — avoid redundant API calls
• Concurrent fetching — fetch_many() pulls data for hundreds of symbols in parallel
• Auto-pagination — paginate_all() walks through paginated endpoints automatically
• Rate limiting — built-in token-bucket rate limiter, thread-safe for concurrent use
• Context manager support for clean resource cleanup

Installation

pip install jaymd96-fmp

Quick Start

from fmp import FMPClient

# API key from argument or FMP_API_KEY env var
client = FMPClient(api_key="your-api-key")

# Real-time quote
client.quote("AAPL")

# Financial statements
client.income_statement("AAPL", period="quarter", limit=4)

# Historical prices
client.historical_price_eod_full("AAPL", from_date="2024-01-01", to_date="2024-12-31")

# Stock screener
client.screener(sector="Technology", market_cap_more_than=1_000_000_000)

# Technical indicators
client.sma("AAPL", period_length=20, timeframe="1day")

Concurrent Fetching

# Fetch quotes for many symbols in parallel
results = client.fetch_many(client.quote, ["AAPL", "MSFT", "GOOG", "AMZN"], max_workers=10)
# Returns: {"AAPL": [...], "MSFT": [...], ...}

Auto-Pagination

# Fetch all pages of stock news
all_news = client.paginate_all(client.stock_news_latest, limit=100, max_pages=50)

Cache

All responses are cached in a local DuckDB database (~/.fmp/cache.db by default). TTLs are set per endpoint category (e.g., 60s for quotes, 7 days for financial statements).

# In-memory cache (no persistence)
client = FMPClient(api_key="...", cache_path=None)

# Custom TTLs
client = FMPClient(api_key="...", ttl_overrides={"realtime_quotes": 30})

# Force refresh (bypass cache)
client.quote("AAPL", force_refresh=True)

# Run SQL queries on cached data
client.sql("SELECT * FROM _raw_cache WHERE endpoint = 'quote'")

# Clear cache
client.clear_cache()
client.clear_cache(endpoint="quote")

Rate Limiting

# Limit to 5 requests per second (for free-tier plans)
client = FMPClient(api_key="...", rate_limit=5)

Context Manager

with FMPClient(api_key="...") as client:
    data = client.quote("AAPL")
# Connections closed automatically

Configuration

License

MIT