Qutip-QOC 0.1.2 release new

.github/workflows/build_documentation.yml

@@ -8,7 +8,7 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - uses: actions/setup-python@v4
         name: Install Python
@@ -42,7 +42,7 @@ jobs:
           #   -T : display a full traceback if a Python exception occurs
 
       - name: Upload built files
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: qutip_qoc_html_docs
           path: doc/_build/html/*

.github/workflows/test.yml

@@ -47,7 +47,7 @@ jobs:
             python-version: "3.12"
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - name: Set up Python
         uses: actions/setup-python@v4
         with:
@@ -92,7 +92,7 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - name: Set up Python
         uses: actions/setup-python@v4
         with:

README.md

@@ -26,9 +26,12 @@ To install the package, use
 pip install qutip-qoc
 ```
 
+By default, the dependencies required for JOPT and for the RL (reinforcement learning) algorithm are omitted.
+They can be included by using the targets `qutip-qoc[jopt]` and `qutip-qoc[rl]`, respectively (or `qutip-qoc[full]` for all dependencies).
+
 ## Documentation and tutorials
 
-The documentation of `qutip-qoc` updated to the latest development version is hosted at [qutip-qoc.readthedocs.io](https://qutip-qoc.readthedocs.io/en/latest/).
+The documentation of `qutip-qoc` updated to the latest development version is hosted at [qutip-qoc.readthedocs.io](https://qutip-qoc.readthedocs.io/latest/).
 Tutorials related to using quantum optimal control in `qutip-qoc` can be found [_here_](https://qutip.org/qutip-tutorials/#optimal-control).
 
 ## Installation from source
@@ -40,7 +43,7 @@ pip install --upgrade pip
 pip install -e .
 ```
 
-which makes sure that you are up to date with the latest `pip` version. Contribution guidelines are available [_here_](https://qutip-qoc.readthedocs.io/en/latest/contribution-code.html).
+which makes sure that you are up to date with the latest `pip` version. Contribution guidelines are available [_here_](https://qutip-qoc.readthedocs.io/latest/contribution/code.html).
 
 To build and test the documentation, additional packages need to be installed:
 

VERSION

@@ -1 +1 @@
-0.1.1
+0.1.2

doc/changelog.rst

@@ -10,6 +10,11 @@ This is an update to the beta release of ``qutip-qoc``.
 
 It mainly introduces the new reinforcement learning algorithm ``qutip_qoc._rl``.
 
+- Non-public facing functions have been renamed to start with an underscore.
+- As with other QuTiP functions, ``optimize_pulses`` now takes a ``tlist`` argument instead of ``_TimeInterval``.
+- The structure for the control guess and bounds has changed and now takes in an optional ``__time__`` keyword.
+- The ``result`` does no longer return ``optimized_objectives`` but instead ``optimized_H``.
+
 Features
 --------
 

doc/conf.py

@@ -62,6 +62,10 @@ def qutip_qoc_version():
 
 html_theme = "sphinx_rtd_theme"
 html_static_path = []
+html_js_files = [
+    'https://code.jquery.com/jquery-3.6.0.min.js',
+]
+
 
 # -- Options for numpydoc ---------------------------------------
 

doc/contribution/code.rst

@@ -7,7 +7,7 @@ Contributing to the source code
 Build up an development environment
 ===================================
 
-Please follow the instruction on the `QuTiP contribution guide <https://qutip.org/docs/latest/development/contributing.html#building>`_ to
+Please follow the instruction on the `QuTiP contribution guide <https://qutip.readthedocs.io/en/latest/development/contributing.html>`_ to
 build a conda environment.
 
 You don't need to build ``qutip`` in the editable mode unless you also want to contribute to `qutip`.

doc/installation.rst

@@ -23,7 +23,19 @@ In particular, the following packages are necessary for running ``qutip-qoc``:
 
 .. code-block:: bash
 
-    numpy scipy jax jaxlib cython qutip qutip-jax qutip-qtrl
+    numpy scipy cython qutip qutip-qtrl
+
+The following packages are required for using the JOPT algorithm:
+
+.. code-block:: bash
+
+    jax jaxlib qutip-jax
+
+The following packages are required for the RL (reinforcement learning) algorithm:
+
+.. code-block:: bash
+
+    gymnasium stable-baselines3
 
 The following package is used for testing:
 

doc/requirements.txt

@@ -9,18 +9,18 @@ matplotlib>=3.6.1
 docutils==0.18.1
 alabaster==0.7.12
 Babel==2.9.1
-certifi==2020.12.5
+certifi==2024.7.4
 chardet==4.0.0
 colorama==0.4.4
-idna==2.10
+idna==3.7
 imagesize==1.4.1
-Jinja2==3.0.1
+Jinja2==3.1.6
 MarkupSafe==2.0.1
 packaging==23.2
 Pygments==2.17.2
 pyparsing==2.4.7
 pytz==2021.1
-requests==2.25.1
+requests==2.32.2
 snowballstemmer==2.1.0
 Sphinx==6.1.3
 sphinx-gallery==0.12.2
@@ -32,4 +32,4 @@ sphinxcontrib-htmlhelp==2.0.1
 sphinxcontrib-jsmath==1.0.1
 sphinxcontrib-qthelp==1.0.3
 sphinxcontrib-serializinghtml==1.1.5
-urllib3==1.26.4
+urllib3==1.26.19

requirements.txt

@@ -1,11 +1,8 @@
 cython>=1.0
 numpy>=1.16.6,<2.0
 scipy>=1.10.1
-jax==0.4.28
-jaxlib==0.4.28
 qutip>=5.0.1
 qutip-qtrl
-qutip-jax
 pre-commit
 gymnasium>=0.29.1
 stable-baselines3>=2.3.2

setup.cfg

@@ -28,24 +28,35 @@ package_dir=
 packages = find:
 include_package_data = True
 install_requires =
-    jax
-    jaxlib
     packaging
     qutip
     qutip-qtrl
-    qutip-jax
     numpy>=1.16.6,<2.0
-    gymnasium>=0.29.1
-    stable-baselines3>=2.3.2
 setup_requires =
     cython>=1.0
     packaging
 
 [options.packages.find]
 where = src
 
+[options.entry_points]
+qutip.family =
+    qutip_qoc = qutip_qoc.family
+
 [options.extras_require]
+jopt =
+    jax
+    jaxlib
+    qutip-jax
+rl =
+    gymnasium>=0.29.1
+    stable-baselines3>=2.3.2
 tests =
     pytest>=6.0
+    jupytext
+    nbconvert
+    ipykernel
 full =
+    %(jopt)s
+    %(rl)s
     %(tests)s

src/qutip_qoc/_jopt.py

@@ -5,38 +5,47 @@
 import qutip as qt
 from qutip import Qobj, QobjEvo
 
-from diffrax import Dopri5, PIDController
+try:
+    import jax
+    from jax import custom_jvp
+    import jax.numpy as jnp
+    import qutip_jax  # noqa: F401
 
-import jax
-from jax import custom_jvp
-import jax.numpy as jnp
-import qutip_jax  # noqa: F401
+    import jaxlib  # noqa: F401
 
+    from diffrax import Dopri5, PIDController
 
-@custom_jvp
-def _abs(x):
-    return jnp.abs(x)
+    _jax_available = True
+except ImportError:
+    _jax_available = False
 
 
-def _abs_jvp(primals, tangents):
-    """
-    Custom jvp for absolute value of complex functions
-    """
-    (x,) = primals
-    (t,) = tangents
+if _jax_available:
+
+    @custom_jvp
+    def _abs(x):
+        return jnp.abs(x)
+
 
-    abs_x = _abs(x)
-    res = jnp.where(
-        abs_x == 0,
-        0.0,  # prevent division by zero
-        jnp.real(jnp.multiply(jnp.conj(x), t)) / abs_x,
-    )
+    def _abs_jvp(primals, tangents):
+        """
+        Custom jvp for absolute value of complex functions
+        """
+        (x,) = primals
+        (t,) = tangents
+
+        abs_x = _abs(x)
+        res = jnp.where(
+            abs_x == 0,
+            0.0,  # prevent division by zero
+            jnp.real(jnp.multiply(jnp.conj(x), t)) / abs_x,
+        )
 
-    return abs_x, res
+        return abs_x, res
 
 
-# register custom jvp for absolut value of complex functions
-_abs.defjvp(_abs_jvp)
+    # register custom jvp for absolut value of complex functions
+    _abs.defjvp(_abs_jvp)
 
 
 class _JOPT:
@@ -55,6 +64,10 @@ def __init__(
         guess_params,
         **integrator_kwargs,
     ):
+        if not _jax_available:
+            raise ImportError("The JOPT algorithm requires the modules jax, "
+                              "jaxlib, and qutip_jax to be installed.")
+
         self._Hd = objective.H[0]
         self._Hc_lst = objective.H[1:]
 
@@ -137,8 +150,8 @@ def _infid(self, params):
         if self._fid_type == "TRACEDIFF":
             diff = X - self._target
             # to prevent if/else in qobj.dag() and qobj.tr()
-            diff_dag = Qobj(diff.data.adjoint(), dims=diff.dims)
-            g = 1 / 2 * (diff_dag * diff).data.trace()
+            diff_dag = diff.dag() # direct access to JAX array, no fallback!
+            g = 1 / 2 * jnp.trace(diff_dag.data._jxa @ diff.data._jxa)
             infid = jnp.real(self._norm_fac * g)
         else:
             g = self._norm_fac * self._target.overlap(X)
@@ -147,4 +160,4 @@ def _infid(self, params):
             elif self._fid_type == "SU":  # f_SU (incl global phase)
                 infid = 1 - jnp.real(g)
 
-        return infid
+        return infid

src/qutip_qoc/_rl.py

@@ -89,7 +89,7 @@ def create_pulse_func(idx):
         self._result = Result(
             objectives=objectives,
             time_interval=time_interval,
-            start_local_time=time.localtime(),  # initial optimization time
+            start_local_time=time.time(),  # initial optimization time
             n_iters=0,  # Number of iterations(episodes) until convergence
             iter_seconds=[],  # list containing the time taken for each iteration(episode) of the optimization
             var_time=True,  # Whether the optimization was performed with variable time
@@ -99,7 +99,7 @@ def create_pulse_func(idx):
         self._backup_result = Result(  # used as a backup in case the algorithm with shorter_pulses does not find an episode with infid<target_infid
             objectives=objectives,
             time_interval=time_interval,
-            start_local_time=time.localtime(),
+            start_local_time=time.time(),
             n_iters=0,
             iter_seconds=[],
             var_time=True,
@@ -180,7 +180,7 @@ def _save_episode_info(self):
             "terminated": self.terminated,
             "truncated": self.truncated,
             "steps_used": self._current_step,
-            "elapsed_time": time.mktime(time.localtime()),
+            "elapsed_time": time.time(),
         }
         self._episode_info.append(episode_data)
 
@@ -256,7 +256,7 @@ def reset(self, seed=None):
         time_diff = self._episode_info[-1]["elapsed_time"] - (
             self._episode_info[-2]["elapsed_time"]
             if len(self._episode_info) > 1
-            else time.mktime(self._result.start_local_time)
+            else self._result.start_local_time
         )
         self._result.iter_seconds.append(time_diff)
         self._current_step = 0  # Reset the step counter
@@ -281,7 +281,7 @@ def _save_result(self):
             self._backup_result._final_states = self._result._final_states.copy()
             self._backup_result.infidelity = self._result.infidelity
 
-        result_obj.end_local_time = time.localtime()
+        result_obj.end_local_time = time.time()
         result_obj.n_iters = len(self._result.iter_seconds)
         result_obj.optimized_params = self._actions.copy() + [
             self._result.total_seconds
@@ -296,20 +296,20 @@ def result(self):
         """
         if self._use_backup_result:
             self._backup_result.start_local_time = time.strftime(
-                "%Y-%m-%d %H:%M:%S", self._backup_result.start_local_time
-            )  # Convert to a string
+                "%Y-%m-%d %H:%M:%S", time.localtime(self._backup_result.start_local_time)
+            )
             self._backup_result.end_local_time = time.strftime(
-                "%Y-%m-%d %H:%M:%S", self._backup_result.end_local_time
-            )  # Convert to a string
+                "%Y-%m-%d %H:%M:%S", time.localtime(self._backup_result.end_local_time)
+            )
             return self._backup_result
         else:
             self._save_result()
             self._result.start_local_time = time.strftime(
-                "%Y-%m-%d %H:%M:%S", self._result.start_local_time
-            )  # Convert to a string
+                "%Y-%m-%d %H:%M:%S", time.localtime(self._result.start_local_time)
+            )
             self._result.end_local_time = time.strftime(
-                "%Y-%m-%d %H:%M:%S", self._result.end_local_time
-            )  # Convert to a string
+                "%Y-%m-%d %H:%M:%S", time.localtime(self._result.end_local_time)
+            )
             return self._result
 
     def train(self):

src/qutip_qoc/family.py

@@ -0,0 +1,8 @@
+"""QuTiP family package entry point."""
+
+from . import __version__
+
+
+def version():
+    """Return information to include in qutip.about()."""
+    return "qutip-qoc", __version__