Processing Flows

The runtime pipelines for indicator, strategy, and library scripts — how each declaration type flows from inputs through to execution or rendering.

Indicator, Strategy, and Library Processing Flows#

Indicator flow#

Use this when the script computes analysis series and then renders them through static declarations or a render-only build_visuals(...).

indicator(...)
  -> input.* defaults register into module context
  -> build_indicator_frame(df, params)
  -> frame columns + optional static attrs config
  -> static plot declarations and/or build_visuals(frame, params, ctx)
  -> managed visual snapshot
  -> ATK visual router and adapters

Strategy flow#

Use this when the script emits canonical trade-frame fields that ATK will normalize into execution intents for backtest, replay, and live runtimes.

strategy(...)
  -> input.* defaults register into module context
  -> build_signal_frame(df, params)
  -> entry_* / sl / tp / trail_* prepared
  -> build_trade_frame(signal_df, params, styles)
  -> build_mapped_trade_frame(signal_df) or equivalent mapping-only normalization
  -> normalized trade frame
  -> execution intent model
  -> backtest / replay / live execution path

Library flow#

Use this when the script exists only to export reusable functions, constants, or enum types for other PyneScript modules.

library(...)
  -> exported callables and enum types discovered from the module
  -> publish metadata and dependency reports generated
  -> importing script calls import_library(...) or library_import(...)
  -> selected exports resolve into a namespace object
  -> indicator or strategy script reuses the helpers in its own builders

Builder Contracts

Imports & Declarations