Allow no-copy construction from SciPy COO arrays.

[hameerabbasi]

What type of PR is this? (check all applicable)

• 💾 Refactor
• 🪄 Feature
• 🐞 Bug Fix
• 🔧 Optimization
• 📚 Documentation
• 🧪 Test
• 🛠️ Other

Related issues

• Related issue #
• Closes #

Checklist

• Code follows style guide
• Tests added
• Documented the changes

---

Please explain your changes below.

• Hold reference to converted scipy.sparse.coo_*.
• Allow comparison against NumPy dtypes.
• Add a sentence explaining the sparse_vector failure.

[hameerabbasi]

@mtsokol This turned out to be a bug on the SciPy side -- SciPy deallocates the underlying data as soon as the coo_array is gone, instead of holding a reference to it in the constituent NumPy arrays. This is also why the tests succeeded before.

[codspeed-hq]

CodSpeed Performance Report

Merging #822 will degrade performances by 12.24%

_{Comparing hameerabbasi:allow-coo-copy (08a65b2) with main (13dda8e)}

Summary

⚡ 1 improvements
❌ 1 regressions
✅ 338 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	hameerabbasi:allow-coo-copy	Change
❌	test_index_fancy[side=100-rank=1-format='coo']	1.3 ms	1.5 ms	-12.24%
⚡	test_index_slice[side=100-rank=2-format='gcxs']	2.7 ms	2.3 ms	+19.09%