Merge branch 'master' into integrate_on_boundary

.github/workflows/cache_dependencies.yml

@@ -34,12 +34,14 @@ jobs:
             gh cache delete "$cache_key"
           done
 
+      # Update the matplotlib version if needed later
       - name: Set up virtual environment
         run: |
           python -m venv .venv-${{ matrix.python-version }}
           source .venv-${{ matrix.python-version }}/bin/activate
           python -m pip install --upgrade pip
           pip install -r devtools/dev-requirements.txt
+          pip install matplotlib==3.7.2
 
       - name: Cache Python environment
         id: cache-env

.github/workflows/regression_tests.yml

@@ -66,7 +66,6 @@ jobs:
           source .venv-${{ matrix.python-version }}/bin/activate
           python -m pip install --upgrade pip
           pip install -r devtools/dev-requirements.txt
-          pip install matplotlib==3.7.2
 
       - name: Set Swap Space
         if: env.has_changes == 'true'
@@ -78,7 +77,6 @@ jobs:
         if: env.has_changes == 'true'
         run: |
           source .venv-${{ matrix.python-version }}/bin/activate
-          pip install matplotlib==3.7.2
           pwd
           lscpu
           python -m pytest -v -m regression\
@@ -107,9 +105,9 @@ jobs:
       - name: Upload coverage
         if: env.has_changes == 'true'
         id : codecov
-        uses: Wandalen/wretry.action@v1.3.0
+        uses: Wandalen/wretry.action@v3.5.0
         with:
-          action: codecov/codecov-action@v3
+          action: codecov/codecov-action@v4
           with: |
             token: ${{ secrets.CODECOV_TOKEN }}
             name: codecov-umbrella

.github/workflows/unit_tests.yml

@@ -72,7 +72,6 @@ jobs:
           source .venv-${{ matrix.combos.python_version }}/bin/activate
           python -m pip install --upgrade pip
           pip install -r devtools/dev-requirements.txt
-          pip install matplotlib==3.7.2
 
       - name: Set Swap Space
         if: env.has_changes == 'true'
@@ -84,7 +83,6 @@ jobs:
         if: env.has_changes == 'true'
         run: |
           source .venv-${{ matrix.combos.python_version }}/bin/activate
-          pip install matplotlib==3.7.2
           pwd
           lscpu
           python -m pytest -v -m unit \
@@ -113,9 +111,9 @@ jobs:
       - name: Upload coverage
         if: env.has_changes == 'true'
         id : codecov
-        uses: Wandalen/wretry.action@v1.3.0
+        uses: Wandalen/wretry.action@v3.5.0
         with:
-          action: codecov/codecov-action@v3
+          action: codecov/codecov-action@v4
           with: |
             token: ${{ secrets.CODECOV_TOKEN }}
             name: codecov-umbrella

.gitignore

@@ -20,4 +20,6 @@ tests/benchmarks/.benchmarks/
 *.output
 *.pdf
 *.pyc
+*.swp
+*.swo
 .pymon

CHANGELOG.md

@@ -12,12 +12,22 @@ New Features
 - Changes ``ToroidalFlux`` objective to default using a 1D loop integral of the vector potential
 to compute the toroidal flux when possible, as opposed to a 2D surface integral of the magnetic field dotted with ``n_zeta``.
 - Allow specification of Nyquist spectrum maximum modenumbers when using ``VMECIO.save`` to save a DESC .h5 file as a VMEC-format wout file
+- Added and tested infinite-n ideal-ballooning stability solver implemented as a part of the BallooningStability Objective. DESC can use reverse-mode AD to now optimize equilibria against infinite-n ideal ballooning modes.
+- Add ``jac_chunk_size`` to ``ObjectiveFunction`` and ``_Objective`` to control the above chunk size for the ``fwd`` mode Jacobian calculation
+  - if ``None``, the chunk size is equal to ``dim_x``, so no chunking is done
+  - if an ``int``, this is the chunk size to be used.
+  - if ``"auto"`` for the ``ObjectiveFunction``, will use a heuristic for the maximum ``jac_chunk_size`` needed to fit the jacobian calculation on the available device memory, according to the formula: ``max_jac_chunk_size = (desc_config.get("avail_mem") / estimated_memory_usage - 0.22)  / 0.85  * self.dim_x`` with ``estimated_memory_usage = 2.4e-7 * self.dim_f * self.dim_x + 1``
+- the ``ObjectiveFunction`` ``jac_chunk_size`` is used if ``deriv_mode="batched"``, and the ``_Objective`` ``jac_chunk_size`` will be used if ``deriv_mode="blocked"``
+
 
 Bug Fixes
 
 - Fixes bugs that occur when saving asymmetric equilibria as wout files
 - Fixes bug that occurs when using ``VMECIO.plot_vmec_comparison`` to compare to an asymmetric wout file
 
+Deprecations
+
+- ``deriv_mode="looped"`` in ``ObjectiveFunction`` is deprecated and will be removed in a future version in favored of ``deriv_mode="batched"`` with ``jac_chunk_size=1``,
 
 
 v0.12.1

CONTRIBUTING.rst

@@ -156,46 +156,43 @@ Opening a PR will trigger a suite of tests and style/formatting checks that must
 We also require approval from at least one (ideally multiple) of the main DESC developers, who may have suggested changes
 or edits to your PR.
 
-What if the ``test_compute_everything`` test fails, or there is a conflict in ``master_compute_data.pkl``?
+What if the ``test_compute_everything`` test fails, or there is a conflict in ``master_compute_data_rpz.pkl``?
 ----------------------------------------------------------------------------------------------------------
-When the outputs of the compute quantities tested by the`test_compute_everything` [test](https://github.com/PlasmaControl/DESC/blob/master/tests/test_compute_funs.py) are changed in a PR, that test will fail.
+When the outputs of the compute quantities tested by the `test_compute_everything` [test](https://github.com/PlasmaControl/DESC/blob/master/tests/test_compute_everything.py) are changed in a PR, that test will fail.
 The three main reasons this could occur are:
 
 -  The PR was not intended to change how things are computed, but messed up something unexpected and now the compute quantities are incorrect, if you did not expect these changes in the PR then look into why these differences are happening and fix the PR.
 -  The PR updated the way one of the existing compute index quantities are computed (either by a redefinition or perhaps fixing an error present in ``master``)
 -  The PR added a new class parametrization (such as a new subclass of ``Curve`` like ``LinearCurve`` etc)
 
-If the 2nd case is the reason, then you must update the ``master_compute_data.pkl`` file with the correct quantities being computed by your PR:
+If the 2nd case is the reason, then you must update the ``master_compute_data_rpz.pkl`` file with the correct quantities being computed by your PR:
 
 -  First, run the test with ``pytest tests -k test_compute_everything`` and inspect the compute quantities whose values are in error, to ensure that only the quantities you expect to be different are shown (and that the new values are indeed the correct ones, you should have a test elsewhere for that though).
 -  If the values are as expected and only the expected compute quantities are different, then replace the block
 
 .. code-block:: python
 
-   except AssertionError as e:
-      error = True
-      print(e)
+   if not error_rpz and update_master_data_rpz:
+        # then update the master compute data
 
 with
 
 .. code-block:: python
 
-   except AssertionError as e:
-      error = False
-      update_master_data = True
-      print(e)
+   if True or (not error_rpz and update_master_data_rpz):
+        # then update the master compute data
 
 
 -  rerun the test ``pytest tests -k test_compute_everything``, now any compute quantity that is different between the PR and master will be updated with the PR value
--  ``git restore tests/test_compute_funs.py`` to remove the change you made to the test
--  ``git add tests/inputs/master_compute_data.pkl`` and commit to commit the new data file
+-  ``git restore tests/test_compute_everything.py`` to remove the change you made to the test
+-  ``git add tests/inputs/master_compute_data_rpz.pkl`` and commit to commit the new data file
 
-If the 3rd case is the reason, then you must simply add the new parametrization to the ``test_compute_everything`` [test](https://github.com/PlasmaControl/DESC/blob/master/tests/test_compute_funs.py)
+If the 3rd case is the reason, then you must simply add the new parametrization to the ``test_compute_everything`` [test](https://github.com/PlasmaControl/DESC/blob/master/tests/test_compute_everything.py)
 
 -  ``things`` dictionary with a sensible example instance of the class to use for the test, and
 -  to the ``grid`` dictionary with a sensible default grid to use when computing the compute quantities for the new class
 -  Then, rerunning the test  ``pytest tests -k test_compute_everything`` will add the compute quantities for the new class and save them to the ``.pkl`` file
--  ``git add tests/inputs/master_compute_data.pkl`` and commit to commit the new data file
+-  ``git add tests/inputs/master_compute_data_rpz.pkl`` and commit to commit the new data file
 
 Styleguides
 ^^^^^^^^^^^

desc/backend.py

@@ -75,14 +75,7 @@
     from jax.numpy import bincount, flatnonzero, repeat, take
     from jax.numpy.fft import irfft, rfft, rfft2
     from jax.scipy.fft import dct, idct
-    from jax.scipy.linalg import (
-        block_diag,
-        cho_factor,
-        cho_solve,
-        eigh_tridiagonal,
-        qr,
-        solve_triangular,
-    )
+    from jax.scipy.linalg import block_diag, cho_factor, cho_solve, qr, solve_triangular
     from jax.scipy.special import gammaln, logsumexp
     from jax.tree_util import (
         register_pytree_node,
@@ -98,6 +91,31 @@
         jnp.trapezoid if hasattr(jnp, "trapezoid") else jax.scipy.integrate.trapezoid
     )
 
+    def execute_on_cpu(func):
+        """Decorator to set default device to CPU for a function.
+
+        Parameters
+        ----------
+        func : callable
+            Function to decorate
+
+        Returns
+        -------
+        wrapper : callable
+            Decorated function that will always run on CPU even if
+            there are available GPUs.
+        """
+
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            with jax.default_device(jax.devices("cpu")[0]):
+                return func(*args, **kwargs)
+
+        return wrapper
+
+    # JAX implementation is not differentiable on gpu.
+    eigh_tridiagonal = execute_on_cpu(jax.scipy.linalg.eigh_tridiagonal)
+
     def put(arr, inds, vals):
         """Functional interface for array "fancy indexing".
 
@@ -123,28 +141,6 @@ def put(arr, inds, vals):
             return arr
         return jnp.asarray(arr).at[inds].set(vals)
 
-    def execute_on_cpu(func):
-        """Decorator to set default device to CPU for a function.
-
-        Parameters
-        ----------
-        func : callable
-            Function to decorate
-
-        Returns
-        -------
-        wrapper : callable
-            Decorated function that will run always on CPU even if
-            there are available GPUs.
-        """
-
-        @functools.wraps(func)
-        def wrapper(*args, **kwargs):
-            with jax.default_device(jax.devices("cpu")[0]):
-                return func(*args, **kwargs)
-
-        return wrapper
-
     def sign(x):
         """Sign function, but returns 1 for x==0.
 
@@ -427,7 +423,7 @@ def tangent_solve(g, y):
 
     trapezoid = np.trapezoid if hasattr(np, "trapezoid") else np.trapz
 
-    def imap(f, xs, batch_size=None, in_axes=0, out_axes=0):
+    def imap(f, xs, *, batch_size=None, in_axes=0, out_axes=0):
         """Generalizes jax.lax.map; uses numpy."""
         if not isinstance(xs, np.ndarray):
             raise NotImplementedError(