Use positional argument in np.reshape to keep working with newer numpy

[sdvillal]

Since numpy version 2.4.0, newshape is now called shape in np.reshape.

[padix-key]

Thanks for finding this 👍.

[codspeed-hq]

CodSpeed Performance Report

Merging #850 will degrade performance by 13.58%

_{Comparing sdvillal:fix-numpy-reshape-compat (3f70f84) with main (693a9cb)}

Summary

⚡ 1 improvement
❌ 1 (👁 1) regression
✅ 61 untouched

Benchmarks breakdown

	Benchmark	BASE	HEAD	Efficiency
👁	benchmark_superimpose[superimpose_homologs]	24.4 ms	28.3 ms	-13.58%
⚡	benchmark_get_assembly	900.8 ms	663.9 ms	+35.67%

[padix-key]

The changes in performance are probably simply caused by the newer NumPy version that is installed in the benchmark job now.

[sdvillal]

Thanks for the quick turnaround @padix-key, and happy new year!