NumPy array metadata (shape, dtype, ndim)

Type: Feature Request

Hello VS Code team,

Feature request: please add an option to show NumPy array metadata (shape, dtype, ndim) directly in the Variables/Locals view when debugging Python. Current workarounds (Data Viewer, Watch, Debug Console) are helpful but a compact shape/dtype column would greatly speed up debugging numeric code.

Use case:

• Quickly inspect array shapes/dtypes (e.g., frame.shape, frame.dtype) while paused at breakpoints without opening the Data Viewer or adding watches.
• Helpful for image-processing / ML workflows where many arrays are inspected frequently.

Suggested design:

• Add optional compact badges or columns in Variables: name | type | shape | dtype | value-preview.
• Or provide a toggle Python: Show Array Shape/DType in Variables.

Repro steps:

1. Open a Python file using NumPy arrays.
2. Start debugger and pause on a line where arrays exist.
3. Observe Variables view — shape/dtype not visible by default.

Environment:

• VS Code version: 1.109.4
• Python extension: 2026.2.0
• OS: linux

Thanks for considering.

Regards,
Uri ZACKHEM urizackhem@gmail.com

VS Code version: Code 1.109.4 (c3a26841a84f20dfe0850d0a5a9bd01da4f003ea, 2026-02-16T15:35:57.932Z)
OS version: Linux x64 6.17.0-14-generic snap
Modes:

Comienza localizando el punto de entrada de renderizado de Variables/Locals en la extensión de depuración de TypeScript y revisando cómo se representan los valores de Python cuando la ejecución está pausada. Determina dónde podrían exponerse shape, dtype y ndim de NumPy y cómo debería comportarse una visualización opcional; se considera completado cuando los metadatos solicitados son visibles mientras la ejecución está pausada sin requerir Data Viewer ni watches.