Indicators Reference¶
Technical indicator facade for pandas-ta-classic, TA-Lib, TDX, and TradingView styles.
Generic technical indicator namespace.
The default ta.* functions are thin wrappers over pandas-ta-classic.
Same-name indicators may differ from future ta.tdx and ta.tv namespaces.
Choose the namespace for the market convention you need.
FunctionIndicator ¶
Callable indicator wrapper with batch and streaming entry points.
For streaming use (on_bar), pass bar_columns to declare which OHLCV
fields the function needs. Each call to on_bar appends to a streaming
state buffer and recomputes over the full history — correct but O(n).
Module-level instances such as ta.sma are shared. Use new_state()
and pass that state to on_bar when a streaming indicator is used by a
long-lived strategy or event loop.
__call__ ¶
__call__(*args: Any, **kwargs: Any) -> pd.Series | pd.DataFrame
Compute the indicator over a batch of data.
compute ¶
compute(*args: Any, **kwargs: Any) -> pd.Series | pd.DataFrame
Compute the indicator over a batch of data.
on_bar ¶
on_bar(bar: Any, *, state: IndicatorState | None = None) -> float | tuple[Any, ...]
Update from a single bar and return the current indicator value.
bar may be a StreamBar dataclass or any mapping with the
required column keys. Recomputes over the full buffer each call.
reset ¶
reset(*, state: IndicatorState | None = None) -> None
Clear streaming buffers.
Passing a state clears only that isolated state. Without a state, this
clears the backwards-compatible default buffer used by direct
on_bar calls.
Indicator ¶
Protocol implemented by function-style indicators.