Create tests for functions in qutip #64

rochisha0 · 2024-08-07T15:55:57Z

In this task, we aim to create comprehensive tests for the functions within the QuTiP library to check their compatibility with jax.grad and jax.jit. These tests will ensure the correctness and robustness of the implemented functions and will cover a wide range of scenarios.

Ericgig · 2024-08-07T16:41:29Z

tests/test_qutip/test_metrics.py

+
+with qutip.CoreOptions(default_dtype="jax"):
+    X = qutip.sigmax()
+    I = qutip.qeye(2)
+    CNOT = qutip.tensor(qutip.basis(2, 0) * qutip.basis(2, 0).dag(), I) + qutip.tensor(qutip.basis(2, 1) * qutip.basis(2, 1).dag(), X)


CNOT is not used, let's remove it.

Ericgig · 2024-08-07T16:46:20Z

tests/test_qutip/test_entropy.py

+    print(f"{name} (original):", result)
+    print(f"{name} (JIT):", result_jit)


No print in tests.

Suggested change

print(f"{name} (original):", result)

print(f"{name} (JIT):", result_jit)

Ericgig · 2024-08-07T16:46:54Z

tests/test_qutip/test_entropy.py

+with qutip.CoreOptions(default_dtype="jax"):
+    basis_0 = qutip.basis(2, 0)
+    basis_1 = qutip.basis(2, 1)
+    bell_state = (qutip.tensor(basis_0, basis_1) + qutip.tensor(basis_1, basis_0)).unit()


Suggested change

bell_state = (qutip.tensor(basis_0, basis_1) + qutip.tensor(basis_1, basis_0)).unit()

bell_state = qutip.bell_state("10")

Ericgig · 2024-08-07T16:48:52Z

tests/test_qutip/test_entropy.py

+    density_matrix = bell_state * bell_state.dag()
+    dm = qutip.rand_dm([5, 5], distribution="pure")


bell_dm and rand_dm?

Ericgig · 2024-08-07T16:49:59Z

tests/test_qutip/test_entropy.py

+with qutip.CoreOptions(default_dtype="jax"):
+    X = qutip.sigmax()
+    I = qutip.qeye(2)
+    CNOT = qutip.tensor(qutip.basis(2, 0) * qutip.basis(2, 0).dag(), I) + qutip.tensor(qutip.basis(2, 1) * qutip.basis(2, 1).dag(), X)


Not used. We have it as qutip.gates.cnot instead of building it yourself.

Sampreet · 2024-08-19T06:58:52Z

tests/test_qutip/test_mcsolve.py

+# Pytest test case for gradient computation
+@pytest.mark.parametrize("omega_val", [1.0, 2.0, 3.0])
+def test_gradient_mcsolve(omega_val):
+    H, state, tlist, c_ops, e_ops = setup_system(size=2)


Since the system has a continuous variable harmonic oscillator model (defined by the ladder operator a inside setup_system), would it be better to use a Hilbert space size (defined by the parameter size here) greater than or equal to 10?

Sampreet · 2024-08-19T07:19:32Z

tests/test_qutip/test_mcsolve.py

+
+# Test setup for gradient calculation
+def setup_system(size=2):
+    a = qt.destroy(size).to("jax")  


Since the Hamiltonian contains two sub-systems, I think the operators would be tensor products with corresponding identities for the other operators:

a = qt.tensor(qt.destroy(size), qt.qeye(2)).to('jaxdia')

Or alternatively,

a = qt.destroy(size).to('jaxdia') & qt.qeye(2).to('jaxdia')

Same goes for sm:

sm = qt.qeye(size).to('jaxdia') & qt.sigmax().to('jaxdia')

Accordingly, the initial state would be:

state = qt.basis(size, size - 1).to('jax') & qutip.basis(2, 1).to('jax')

Or alternatively,

state = qt.basis([size, 2], [size - 1, 1]).to('jax')

coveralls · 2024-08-23T13:57:37Z

Pull Request Test Coverage Report for Build 10554916907

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

For more information on this, see Tracking coverage changes with pull request builds.
To avoid this issue with future PRs, see these Recommended CI Configurations.
For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

0 of 0 changed or added relevant lines in 0 files are covered.
6 unchanged lines in 2 files lost coverage.
Overall coverage decreased (-0.2%) to 90.303%

Files with Coverage Reduction	New Missed Lines	%
src/qutip_jax/binops.py	1	97.0%
src/qutip_jax/settings.py	5	70.59%

Totals
Change from base Build 9842191187:	-0.2%
Covered Lines:	1220
Relevant Lines:	1351

💛 - Coveralls

Ericgig

Overall look good.
Let's just make the mcsolve test faster.

Ericgig · 2024-08-23T14:18:00Z

tests/test_qutip/test_mcsolve.py

+    return result.expect[0][-1].real
+
+# Pytest test case for gradient computation
+@pytest.mark.parametrize("omega_val", [1.0, 2.0, 3.0])


This test is the slowest and grad functionality is not affected by the value of omega, so let's run it only once.
We could loop over options (improved_sampling, store_states, keep_runs_results) or mixed state input later.

Suggested change

@pytest.mark.parametrize("omega_val", [1.0, 2.0, 3.0])

@pytest.mark.parametrize("omega_val", [2.0])

Ericgig · 2024-08-23T14:20:58Z

tests/test_qutip/test_mcsolve.py

+    e_ops = [a.dag() * a, sm.dag() * sm]
+
+    # Time list
+    tlist = jnp.linspace(0.0, 10.0, 101)


Let's speed up the tests, the actual range does not affect the validity.

Suggested change

tlist = jnp.linspace(0.0, 10.0, 101)

tlist = jnp.linspace(0.0, 1.0, 101)

Ericgig · 2024-08-23T14:33:35Z

tests/test_qutip/test_mcsolve.py

+    H[1][1] = qt.coefficient(H_1_coeff, args={"omega": omega})
+
+    result = mcsolve(H, state, tlist, c_ops, e_ops, ntraj=10, options={"method": "diffrax"})


Passing args to a solver should overwrite the existing values.

Suggested change

H[1][1] = qt.coefficient(H_1_coeff, args={"omega": omega})

result = mcsolve(H, state, tlist, c_ops, e_ops, ntraj=10, options={"method": "diffrax"})

result = mcsolve(

H, state, tlist, c_ops, e_ops, ntraj=10,

args={"omega": omega},

options={"method": "diffrax"}

)

create tests for metrics and entropy

b5d0ff7

rochisha0 marked this pull request as draft August 7, 2024 15:56

Ericgig reviewed Aug 7, 2024

View reviewed changes

rochisha0 added 4 commits August 8, 2024 13:26

made required changes to test_entropy

dcda470

made required changes to test_metrics

379bb16

fixing bell state

946888b

add test for mcsolve and qobj

ae21d43

rochisha0 changed the title ~~Create tests for metrics and entropy~~ Create tests for functions in qutip Aug 14, 2024

Sampreet reviewed Aug 19, 2024

View reviewed changes

update test for qobj

0f561cf

Sampreet reviewed Aug 19, 2024

View reviewed changes

rochisha0 added 2 commits August 19, 2024 15:26

incorporated suggestions

0aee90a

update qobj according to changes

931a3f1

rochisha0 marked this pull request as ready for review August 22, 2024 19:53

rochisha0 added 2 commits August 23, 2024 01:28

replace use_jax_backend with set_as_default

9e1a28c

replace use_jax_backend with set_as_default-2

250e900

Ericgig reviewed Aug 23, 2024

View reviewed changes

make mcsolve tests faster

e7fd203

Ericgig approved these changes Aug 26, 2024

View reviewed changes

Ericgig merged commit 0cc6256 into qutip:master Aug 26, 2024
3 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Create tests for functions in qutip #64

Create tests for functions in qutip #64

rochisha0 commented Aug 7, 2024 •

edited

Loading

Ericgig Aug 7, 2024

Ericgig Aug 7, 2024

Ericgig Aug 7, 2024

Ericgig Aug 7, 2024

Ericgig Aug 7, 2024

Sampreet Aug 19, 2024 •

edited

Loading

Sampreet Aug 19, 2024

coveralls commented Aug 23, 2024 •

edited

Loading

Ericgig left a comment

Ericgig Aug 23, 2024

Ericgig Aug 23, 2024

Ericgig Aug 23, 2024

		print(f"{name} (original):", result)
		print(f"{name} (JIT):", result_jit)

	bell_state = (qutip.tensor(basis_0, basis_1) + qutip.tensor(basis_1, basis_0)).unit()
	bell_state = qutip.bell_state("10")

		density_matrix = bell_state * bell_state.dag()
		dm = qutip.rand_dm([5, 5], distribution="pure")

	@pytest.mark.parametrize("omega_val", [1.0, 2.0, 3.0])
	@pytest.mark.parametrize("omega_val", [2.0])

	tlist = jnp.linspace(0.0, 10.0, 101)
	tlist = jnp.linspace(0.0, 1.0, 101)

		H[1][1] = qt.coefficient(H_1_coeff, args={"omega": omega})

		result = mcsolve(H, state, tlist, c_ops, e_ops, ntraj=10, options={"method": "diffrax"})

Create tests for functions in qutip #64

Create tests for functions in qutip #64

Conversation

rochisha0 commented Aug 7, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Sampreet Aug 19, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

coveralls commented Aug 23, 2024 • edited Loading

Pull Request Test Coverage Report for Build 10554916907

Warning: This coverage report may be inaccurate.

Details

💛 - Coveralls

Ericgig left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rochisha0 commented Aug 7, 2024 •

edited

Loading

Sampreet Aug 19, 2024 •

edited

Loading

coveralls commented Aug 23, 2024 •

edited

Loading