Fix memory copies getz

[steven-murray]

Stacks on #119

Fixes an issue where memory copies were made when using multiple beams.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.88%. Comparing base (f6f93ee) to head (f5cfd85).
⚠️ Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #120      +/-   ##
==========================================
+ Coverage   98.46%   98.88%   +0.42%     
==========================================
  Files          22       22              
  Lines         975      989      +14     
  Branches       98      103       +5     
==========================================
+ Hits          960      978      +18     
+ Misses          8        6       -2     
+ Partials        7        5       -2     

Flag	Coverage Δ
unittests	78.46% <64.70%> (+0.20%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR addresses excess memory usage when simulating with per-antenna beam selection by avoiding fancy-indexing copies in Z-matrix formation, and extends the chunk-sizing logic with a configurable free-memory safety margin.

Changes:

• Avoid fancy-indexing beam selection in ZMatrixCalc to prevent large temporary allocations when beam_idx is used.
• Add a memory_buffer parameter to CPU/GPU simulate() and propagate it into chunk estimation utilities.
• Adjust chunk-estimation thresholds to account for memory_buffer, and tweak docstring composition in the GPU entrypoint.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 8 comments.

File	Description
src/matvis/gpu/gpu.py	Adds memory_buffer to GPU simulate and passes it into chunk sizing.
src/matvis/cpu/cpu.py	Adds memory_buffer to CPU simulate and documents the new parameter.
src/matvis/core/getz.py	Replaces beam fancy-indexing with a per-antenna loop to avoid memory copies.
src/matvis/_utils.py	Extends chunk-sizing helpers with memory_buffer and uses it in the stopping criterion.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[steven-murray]

@copilot please re-review given the rebase

[Copilot]

@copilot please re-review given the rebase

Re-reviewed after the rebase. The remaining open items were:

1. Missing input validation for source_buffer and memory_buffer in cpu/cpu.py
2. Incomplete validation in gpu/gpu.py (only checked source_buffer > 1.0, missing lower-bound and memory_buffer validation)

Both are now addressed in commit 945ac01 — both functions now raise ValueError when source_buffer or memory_buffer is outside (0, 1], consistent with the validation already in _utils.get_required_chunks. The previously-resolved items (_utils.py validation, docstring fixes, and memory_buffer tests) were already in place from prior commits.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• data.astropy.org
  • Triggering command: /usr/bin/python3 python3 -m pytest tests/test_utils.py tests/test_matvis_cpu.py -x -v _TOKEN"; }; f (dns block)
• www.astropy.org
  • Triggering command: /usr/bin/python3 python3 -m pytest tests/test_utils.py tests/test_matvis_cpu.py -x -v _TOKEN"; }; f (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[tyler-a-cox]

This is all looking good to me. Thanks @steven-murray!