Principle:TA Lib Ta lib python Abstract Function Execution

Knowledge Sources	TA-Lib Documentation
Domains	Technical_Analysis, Software_Architecture
Last Updated	2026-02-09 22:00 GMT

Overview

An execution pattern for computing indicator values through a Function object interface, supporting both explicit method calls and implicit callable invocation.

Description

The Abstract API provides three ways to execute a configured Function and retrieve results:

1. outputs property: Returns cached results (recomputes if inputs/parameters changed)
2. run(input_arrays): Sets input arrays and returns outputs in one call
3. __call__(*args, **kwargs): Allows calling the Function object like a function with the same signature as the Function API

The execution caches results and only recomputes when inputs or parameters change (tracked via the outputs_valid flag). Output format adapts to input type — pandas DataFrame input yields pandas Series/DataFrame output.

Usage

Use this principle when you need to execute indicators through the Abstract API. Choose between the three execution methods based on your workflow:

• outputs — when data and params are already set
• run() — when you want to set data and execute in one step
• __call__() — when you want the same syntax as the Function API

Theoretical Basis

The execution pattern implements lazy evaluation with caching:

# Abstract execution pattern
if not outputs_valid:
    results = call_c_function(input_arrays, parameters)
    cache_results(results)
    outputs_valid = True
return cached_results

Output conversion preserves input types:

• dict input → numpy array/list output
• pandas.DataFrame input → pandas.Series/DataFrame output
• polars.DataFrame input → polars.Series/DataFrame output

Related Pages

Implemented By

• Implementation:TA_Lib_Ta_lib_python_Abstract_Function_Run

Uses Heuristic

• Heuristic:TA_Lib_Ta_lib_python_Thread_Safety_With_Abstract_API