Handle NumPy 2.x hash compatibility for WebGL data

[Copilot]

WebGL data hashing failed on NumPy ≥2.2 because ndarray.tostring() was removed, causing WebGL view initialization to crash when hashing brain data arrays.

• Hashing fix: Coerce inputs to NumPy arrays and hash via tobytes() in braindata._hash to restore compatibility with modern NumPy.
• Regression coverage: Added a focused test in cortex/tests/test_braindata.py that asserts _hash matches hashlib.sha1(array.tobytes()) for NumPy arrays.

Example:

import numpy as np
from cortex.dataset.braindata import _hash

arr = np.arange(12, dtype=np.float32).reshape(3, 4)
print(_hash(arr))  # hashes using arr.tobytes()

Original prompt

---

This section details on the original issue you should resolve

<issue_title>numpy.ndarray has no attribute 'tostring' when using webgl</issue_title>
<issue_description>Hi,
I encountered this error while attempting to use the WebGL view with Pycortex v1.2.11 and NumPy 2.3.4:

  File "/home/ftravi/Documents/Projects/fmri-context/decoding/plot_rois.py", line 30, in <module>
    webgl.show(vol, with_colorbar=True, linewidth=4, thick=1, with_curvature=True, roi_list=ROIS_TO_PLOT, open_browser=True)
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/utils.py", line 40, in __call__
    return self._load()(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/webgl/view.py", line 399, in show
    package = Package(data)
              ^^^^^^^^^^^^^
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/webgl/data.py", line 24, in __init__
    self.uniques = list(data.uniques(collapse=True))
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/dataset/dataset.py", line 120, in uniques
    uniques.add(sv)
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/dataset/braindata.py", line 63, in __hash__
    return hash(_hash(self.data))
                ^^^^^^^^^^^^^^^^
  File "/home/ftravi/.local/lib/python3.12/site-packages/cortex/dataset/braindata.py", line 631, in _hash
    return hashlib.sha1(array.tostring()).hexdigest()
                        ^^^^^^^^^^^^^^
AttributeError: 'numpy.ndarray' object has no attribute 'tostring'

As stated in the traceback, it occurs in line 631 of dataset/braindata.py.

Minimal code to reproduce the error:

• Have Numpy > 2.2 installed

import cortex
vol = cortex.Volume.random(subject='S1', xfmname='fullhead')
cortex.webshow(vol)

Quick fix:
Simply replacing line 631 in dataset/braindata.py from:
return hashlib.sha1(array.tostring()).hexdigest()
to:
return hashlib.sha1(array.tobytes()).hexdigest()

If needed, I can open a pull request with this fix.

Thank you!</issue_description>

<agent_instructions>Make sure to check other places that have .tostring() and ensure compatibility with recent numpy versions.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes numpy.ndarray has no attribute 'tostring' when using webgl #583

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[mvdoc]

Rename this to test_braindata.py; hopefully we can add more tests in the future

[Copilot]

Renamed to test_braindata.py in bfb0e22 to keep room for additional cases.

[Copilot]

Pull request overview

This PR fixes a compatibility issue with NumPy ≥2.2 where ndarray.tostring() was removed, causing WebGL view initialization to crash when hashing brain data arrays.

• Updated _hash() function to use tobytes() instead of the deprecated tostring() method
• Added np.asarray() coercion to ensure inputs are converted to NumPy arrays before hashing
• Added regression test to verify correct hashing behavior

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
cortex/dataset/braindata.py	Fixed _hash() function to use tobytes() and added array coercion with np.asarray() for NumPy 2.x compatibility
cortex/tests/test_braindata.py	Added test to verify _hash() correctly uses tobytes() and matches expected SHA1 hash output

---

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