[AUTO-MERGE] Pass a deepcopy of config to verbs and Hyrax methods

[drewoldag]

Pass a deepcopy of the config to the various verbs and Hyrax methods that take a config as input. Added unit tests to check for config leakage and prevent regressions.

Change Description

Closes #703
This change will create a deep copy of the top level config before passing it to the verb (or quasi-verb that is called from a Hyrax method)

Perhaps we were trying to do this in the base Verb class with the self.config = config line, but that was allowing config changes to leak back up to the top level where they weren't appropriate.

This change allows user to pass in a config, and update the config interactively (i.e. in a notebook) but when a verb is called i.e. h.infer(), the updates to the config that the verb makes will not be seen by subsequent verb calls.

Issue #703 describes the scenario where this was first noticed.

[drewoldag]

Note, a larger change could involve cleaning up the config stuff throughout the various verbs. There is inconsistency about how the input configs are handled. In a later PR it might be nice to consolidate on one approach.

[codecov]

Codecov Report

❌ Patch coverage is 45.45455% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 64.69%. Comparing base (d4e6908) to head (666c490).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/hyrax/hyrax.py	45.45%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #757      +/-   ##
==========================================
- Coverage   64.70%   64.69%   -0.02%     
==========================================
  Files          61       61              
  Lines        5888     5894       +6     
==========================================
+ Hits         3810     3813       +3     
- Misses       2078     2081       +3     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Copilot]

Pull request overview

This PR aims to prevent Hyrax runtime config “leakage” by ensuring verbs (and verb-like Hyrax methods) receive an isolated deep-copied config, addressing the regression described in issue #703 where sequential train()/infer() cycles could reuse stale weights due to mutated config.

Changes:

• Pass deepcopy(self.config) into verb dispatch (Hyrax.__getattr__) and into download/prepare/rebuild_manifest/raw_data_dimensions call paths.
• Add regression tests ensuring sequential verb calls do not mutate h.config.
• Add a unit test demonstrating that load_model_weights() mutations don’t affect the original config when a deepcopy is used.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
src/hyrax/hyrax.py	Deep-copies self.config before dispatching to verbs and quasi-verbs to prevent mutation leaks into Hyrax.config.
tests/hyrax/test_infer.py	Adds a unit test asserting config isolation when load_model_weights() is called with a deep-copied config.
tests/hyrax/test_config_isolation.py	Adds regression tests to ensure train()/infer()/prepare() do not mutate the Hyrax instance’s config across calls.

---

You can also share your feedback on Copilot code review. Take the survey.

[github-actions]

Before [d4e6908]	After [0b88925]	Ratio	Benchmark (Parameter)
failed	failed	n/a	data_cache_benchmarks.DataCacheBenchmarks.time_preload_cache_hsc1k
failed	failed	n/a	data_cache_benchmarks.DataCacheBenchmarks.track_cache_hsc1k_hyrax_size_undercount
failed	failed	n/a	data_request_benchmarks.DatasetRequestBenchmarks.time_request_all_data
40.4±0.4ms	41.0±2ms	1.02	benchmarks.time_nb_obj_dir
2.00±0.01s	2.03±0.01s	1.01	benchmarks.time_prepare_help
2.00±0.01s	2.01±0.01s	1.01	benchmarks.time_visualize_help
3.71G	3.75G	1.01	vector_db_benchmarks.VectorDBInsertBenchmarks.peakmem_load_vector_db(16384, 'qdrant')
285±4ms	285±2ms	1.00	benchmarks.time_import
2.01±0.01s	2.00±0s	1.00	benchmarks.time_infer_help
2.00±0.01s	2.01±0.02s	1.00	benchmarks.time_lookup_help

Click here to view all benchmarks.

[mtauraso]

🚢 it