Interferometry

This page describes ALMASim’s interferometric simulation layer.

Location

The main code is in src/almasim/services/interferometry.

Main Modules

core.py
antenna.py
baselines.py
frequency.py
imaging.py
multiconfig.py
noise.py
total_power.py
utils.py

Current Scope

ALMASim currently focuses on single-pointing observation fidelity.

It supports:

interferometric simulation for ALMA-like antenna arrays
multi-configuration single-pointing observation plans
total-power simulation
combined interferometric result aggregation
UV mask generation
dirty beam and dirty image products

It does not currently target mosaics as a first-class workflow.

Observation Planning

Single-pointing observation plans are built in src/almasim/services/observation_plan.py.

Plans can be:

constructed directly
inferred from metadata rows
split automatically into 12m, 7m, and TP components when metadata contains mixed antenna prefixes

Interferometric Products

The interferometry layer can produce:

dirty image cube
dirty visibility cube
beam cube
total sampling cube
UV mask cube
U cube
V cube

These products are useful both for astronomy workflows and for inverse-problem or ML workflows.

Multi-Configuration Support

The multi-configuration aggregation code is in multiconfig.py.

It combines:

per-config interferometric runs
per-config metadata
per-config dirty and beam products
combined interferometric products

Total Power

Total-power simulation is implemented in total_power.py.

It provides:

diffraction-limited TP beam estimation
TP smoothing of model cubes
TP dirty products
multi-config TP combination

Fidelity Direction

The current design goal is to move ALMASim closer to simalma for single-pointing fidelity while preserving library-first and ML-oriented access to intermediate products.