Quant Buffet API

Data API

load_daily_prices and price panel conventions.

Market data lives in `backtest.data`. The lab uses adjusted daily closes from Yahoo Finance, cached under backtest/data_cache/ for repeat runs.

load_daily_prices

prices = load_daily_prices(
    ["SPY", "TLT", "GLD"],
    start="2010-01-01",
    end=None,          # optional end date "YYYY-MM-DD"
    use_cache=True,    # default: read/write backtest/data_cache/
    strict=False,      # if True, raise on first symbol failure
)
# prices: pd.DataFrame — DatetimeIndex rows, one column per ticker (adjusted close)

Parameter	Default	Description
`symbols`	—	List of ticker strings, e.g. `["SPY", "TLT"]`.
`start`	`"2000-01-01"`	Inclusive start date (ISO string).
`end`	`None`	Optional exclusive end; `None` means latest available.
`use_cache`	`True`	Read/write CSV cache files keyed by symbol and date range.
`strict`	`False`	If `True`, raise on first symbol failure instead of skipping.

Return value

`pd.DataFrame` indexed by `DatetimeIndex` (timezone-naive).
One float column per symbol that loaded successfully.
Missing days within a series are forward-filled; pre-IPO history stays NaN until first print.
Failed symbols are omitted unless strict=True. Partial failures attach prices.attrs["load_errors"] when some symbols fail.

load_price_panel (optional)

from backtest.data import load_price_panel

prices, missing = load_price_panel(["SPY", "XYZ"], start="2010-01-01")
# missing: list of symbols with no data

Note. In the lab sandbox you normally do not call load_daily_prices yourself — the runner loads ASSETS for you. Use this API when running scripts locally or in batch backtests.

Working with the panel

Filter columns: cols = [c for c in ASSETS if c in prices.columns].
Rolling windows: prices[cols].rolling(20, min_periods=20).mean().
Cross-sectional ranks: rets.rank(axis=1, ascending=False).
Align to engine dates: series.reindex(engine_index).ffill().