ManifoldSubset and ChartFunction objects corresponding to a CVXPY expressions

Question

ManifoldSubset and ChartFunction objects corresponding to a CVXPY expressions

mkoeppe opened this issue 4 years ago · 29 comments

CVXPY is a system for disciplined convex programming in which finite dimensional optimization problems are expressed using expression trees.

We monkey-patch caching _sage_ methods into the CVXPY classes, along the same lines as it is done for SymPy in sage.interfaces.sympy.

Each leaf is a scalar, vector or matrix, subject to a domain restriction such as nonneg or PSD. _sage_ methods return symbolic variables or expressions or vectors/matrices thereof. _sage_domain_ methods return the domain as a set. (For the PSD domain, see #30172.) _sage_manifold_ method returns the pullback (#31688), defining a manifold subset / submanifold etc.
A subclass of cvxpy.Problem would provide a method feasible_set an instance of class ConditionSet_cvxpy (see #24176) or ConvexSet_cvxpy, representing sets of points that satisfy a list of cvxpy.Constraints.

Depends on #31962
Depends on #31688
Depends on #30172
Depends on #31982

CC: @dimpase @yuan-zhou @egourgoulhon

Component: manifolds

Author: Matthias Koeppe, ...

Branch/Commit: u/mkoeppe/manifoldsubset_corresponding_to_a_cvxpy_leaf__variable_parameter_ @ 84a5868

Issue created by migration from https://trac.sagemath.org/ticket/31981

Answer 1 · 2021-06-13T02:40:55.000Z

Description changed:

--- 
+++ 
@@ -6,3 +6,4 @@
 
 The PSD domain can be obtained as pullback (#31688) of #30172.
 
+We may want to monkey-patch caching `_sage_` methods into the CVXPY classes, along the same lines as it is done for SymPy in `sage.interfaces.sympy`.

Answer 2 · 2021-06-13T16:40:36.000Z

Changed dependencies from #31962 to #31962, #31688

Answer 3 · 2021-06-13T19:53:36.000Z

Changed dependencies from #31962, #31688 to #31962, #31688, #30172

Answer 4 · 2021-06-13T19:55:11.000Z

Branch: u/mkoeppe/manifoldsubset_corresponding_to_a_cvxpy_leaf__variable_parameter_

Answer 5 · 2021-06-13T20:58:42.000Z

Commit: 84cffde

Answer 6 · 2021-06-13T20:58:42.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`84cffde`	`_cvxpy_Variable_sage_: Actually return chart functions`

Answer 7 · 2021-06-13T20:59:03.000Z

Author: Matthias Koeppe

Answer 8 · 2021-06-13T21:23:57.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`d4e19aa`	`_cvxpy_MulExpression_sage_: New`

Answer 9 · 2021-06-13T21:23:57.000Z

Changed commit from 84cffde to d4e19aa

Answer 10 · 2021-06-13T21:51:01.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`3889cf5`	`_cvxpy_AddExpression_sage_: New`

Answer 11 · 2021-06-13T21:51:01.000Z

Changed commit from d4e19aa to 3889cf5

Answer 12 · 2021-06-13T22:11:29.000Z

Changed commit from 3889cf5 to 53c849a

Answer 13 · 2021-06-13T22:11:29.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`53c849a`	`_cvxpy_log1p_sage_: New`

Answer 14 · 2021-06-14T23:56:25.000Z

Changed dependencies from #31962, #31688, #30172 to #31962, #31688, #30172, #31982

Answer 15 · 2021-06-15T00:02:30.000Z

Changed commit from 53c849a to 802a1ff

Answer 16 · 2021-06-15T00:02:30.000Z

Branch pushed to git repo; I updated commit sha1. Last 10 new commits:

`e8d3427`	`clarify comment and doctest for multiple inheritance`
`b873bcb`	`some doc`
`27e3889`	`Merge #31721`
`3c36bd2`	`Eliminate use of MultiCoordFunction._functions`
`49f63a4`	`Vector_symbolic_dense: Fix pickling`
`dcdd412`	`src/sage/manifolds/continuous_map.py: Update doctest output`
`fb31802`	`src/sage/manifolds/chart.py: Update doctest output`
`c06130b`	`Merge #31982`
`271943d`	`_cvxpy_log_det_sage_: New`
`802a1ff`	`src/sage/interfaces/cvxpy.py: Update doctest output`

Answer 17 · 2021-06-15T00:45:14.000Z

Changed commit from 802a1ff to 6e6b8cd

Answer 18 · 2021-06-15T00:45:14.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`6e6b8cd`	`_cvxpy_Constant_sage_: New`

Answer 19 · 2021-06-15T03:32:35.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`760b762`	`_cvxpy_power_sage_: New; fix up 'pos' domain`

Answer 20 · 2021-06-15T03:32:35.000Z

Changed commit from 6e6b8cd to 760b762

Answer 21 · 2021-06-15T03:36:27.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`7323b10`	`ConvexSet_base._test_contains: Only test extension to AA for exact base rings`
`e0d2eeb`	`Merge #31959`
`e6c3ed5`	`RelativeInterior.__hash__: New`
`4c7d57f`	`src/sage/geometry/relative_interior.py: Fix doctest output`
`9ea0cea`	`ManifoldSubsetPullback: Do not trip over relative interiors of cones`
`a60e225`	`Merge #31688`

Answer 22 · 2021-06-15T03:36:27.000Z

Changed commit from 760b762 to a60e225

Answer 23 · 2021-06-15T16:47:14.000Z

Changed commit from a60e225 to bb624fb

Answer 24 · 2021-06-15T16:47:14.000Z

Branch pushed to git repo; I updated commit sha1. New commits:

`bb624fb`	`_cvxpy_{NegExpression,quad_over_lin,NonPos,NonNeg,Inequality}_sage_: New`

Answer 25 · 2021-06-18T21:15:59.000Z

Changed commit from bb624fb to 84a5868

Answer 26 · 2021-06-18T21:15:59.000Z

Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:

`2514586`	`Merge #31644`
`3ca139a`	`src/doc/en/reference/manifolds: Add pullback`
`1d430ee`	`src/sage/manifolds/continuous_map.py, src/sage/manifolds/subsets: Use qualified class name in docstrings`
`1266de9`	`src/sage/geometry/convex_set.py: Fix docstring markup`
`df64125`	`ScalarField.preimage: Make pullback an alias; expand documentation`
`3877c5f`	`ManifoldSubsetPullback.is_closed: Remove unfinished NNC_Polyhedron example`
`0c549f0`	`Chart.pullback, preimage: New`
`7bb720e`	`ManifoldSubsetPullback.__contains__: Handle the case of charts, which return tuples of coordinates, not vectors`
`60ea044`	`ContinuousMap.preimage, ScalarField.preimage: Document that output can also be a manifold`
`84a5868`	`Merge #31688`

Answer 27 · 2021-06-19T04:13:37.000Z

Description changed:

--- 
+++ 
@@ -1,9 +1,9 @@
-https://www.cvxpy.org/api_reference/cvxpy.expressions.html#leaf
+CVXPY is a system for disciplined convex programming in which finite dimensional optimization problems are expressed using expression trees.
 
-Each leaf is a scalar, vector or matrix, subject to a domain restriction such as `nonneg` or `PSD`.
+We monkey-patch caching `_sage_` methods into the CVXPY classes, along the same lines as it is done for SymPy in `sage.interfaces.sympy`.
 
-We model the interior of the domain as an embedded submanifold.
+- Each [leaf](https://www.cvxpy.org/api_reference/cvxpy.expressions.html#leaf) is a scalar, vector or matrix, subject to a domain restriction such as `nonneg` or `PSD`. `_sage_` methods return symbolic variables or expressions or vectors/matrices thereof. `_sage_domain_` methods return the domain as a set. (For the PSD domain, see #30172.) `_sage_manifold_` method returns the pullback (#31688), defining a manifold subset / submanifold etc.
 
-The PSD domain can be obtained as pullback (#31688) of #30172.
+- A subclass of `cvxpy.Problem` would provide a method `feasible_set` an instance of class `ConditionSet_cvxpy` (see #24176) or `ConvexSet_cvxpy`, representing sets of points that satisfy a list of `cvxpy.Constraint`s. 
 
-We may want to monkey-patch caching `_sage_` methods into the CVXPY classes, along the same lines as it is done for SymPy in `sage.interfaces.sympy`.
+

Answer 28 · 2022-03-12T21:52:08.000Z

Changed author from Matthias Koeppe to Matthias Koeppe, Andrey Belgorodski

Answer 29 · 2022-03-12T21:52:21.000Z

Changed author from Matthias Koeppe, Andrey Belgorodski to Matthias Koeppe, ...