numpy 2

[casperdcl]

• misc numpy>=2 safety (fixes support numpy v2 #1915)
• test numpy 2.*
• test astra-toolbox 2.*
• bump ccpi-regulariser 25.0.0
• bump tomophantom Tomophantom plugin #832
  • setup conda correctly dkazanc/TomoPhantom#130
  • suitale version of TomoPhantom on httomo, probably v3.0.3
  • move from ccpi::tomophantom to httomo::tomophantom updates to TomoPhantom plugin to use v3.0 #2287
• bump tigre Bump TIGRE version #2270
• move from conda-forge::astra-toolbox to astra-toolbox::astra-toolbox
• depends on ccpi-regulariser build with numpy>=2 Remove pinning to numpy versions CCPi-Regularisation-Toolkit#215

unsure about

• CIL/Wrappers/Python/cil/optimisation/algorithms/FISTA.py

  Line 234 in 4b038a7

  return self.f(x) + self.g(x)

• cil.plugins.tigre.Atb

• CIL/Wrappers/Python/cil/optimisation/algorithms/PDHG.py

  Line 516 in 4b038a7

  d1 = f_eval_d + g_eval_d

• CIL/Wrappers/Python/cil/optimisation/algorithms/APGD.py

  Line 285 in 4b038a7

  return self.f(x) + self.g(x)

upstream problems

• test tomophantom with numpy>=2
• test tigre with numpy>=2
• test astra-toolbox=*=cuda* with numpy>=2
• test_CIL_CVXPy -> cvxpy -> scipy:

test_cil_vs_cvxpy_totalvariation_anisotropic (test_CIL_CVXPy.Test_CIL_vs_CVXPy.test_cil_vs_cvxpy_totalvariation_anisotropic) ... 
...
$CONDA_PREFIX/lib/python3.12/site-packages/scipy/sparse/_dia.py:140: UserWarning: result dtype changed due to the removal of value-based promotion from NumPy. Changed from int64 to int32.
  nnz += min(M+k,N)
...
$CONDA_PREFIX/lib/python3.12/site-packages/cvxpy/utilities/coeff_extractor.py:296: UserWarning: result dtype changed due to the removal of value-based promotion from NumPy. Changed from int32 to int64.
  P.row += offset

[paskino]

I'm assuming that all the filters in this file return the same type as the input.

[lauramurgatroyd]

For tomophantom, the next step is to test with the latest version on conda:

conda install -c httomo tomophantom

These are the recommended instructions on the readme. It may support numpy 2, we don't know.

If possible we would like to continue to support tomophantom for now, unless their code interface has changed.

If it has changed we will rebuild and release on ccpi channel tomophantom v2.0.0 with numpy 2

[gfardell]

These tests fail using Tigre's FBP for the usual data type promotion problem.

CIL/Wrappers/Python/test/test_PluginsTigre_FBP.py

Lines 52 to 67 in 3065ebd

	@unittest.skipUnless(has_tigre and has_nvidia, "Requires TIGRE GPU")
	class Test_Parallel3D_FBP_tigre(unittest.TestCase, TestCommon_FBP_SIM):
	def setUp(self):
	setup_parameters(self)
	self.Parallel3D()
	self.tolerance_fbp = 1e-3
	self.tolerance_fbp_roi = 1e-3
	@unittest.skipUnless(has_tigre and has_nvidia, "Requires TIGRE GPU")
	class Test_Parallel2D_FBP_tigre(unittest.TestCase, TestCommon_FBP_SIM):
	def setUp(self):
	setup_parameters(self)
	self.Parallel2D()
	self.tolerance_fbp = 1e-3
	self.tolerance_fbp_roi = 1e-3

I have opened an issue on TIGRE.
CERN/TIGRE#730

We should discuss how to handle it in the mean time, we don't really want CIL to just recast it back and be happy.