get rid of CartesianProduct

[videlec]

The features of sage.combinat.cartesian_product.CartesianProduct are now completely integrated into the category framework (see #18290). We remove all occurrences of CartesianProduct to either cartesian_product or itertools.product. We deprecate the CartesianProduct from sage.combinat.cartesian_product.

In order to support all features of the old class we also:

• move the __iter__ from EnumeratedSets.CartesianProducts.ParentMethods to Sets.CartesianProducts.ParentMethods
• allows cartesian_product to be called with list, tuple, set, frozenset
• make cartesian_product([]) works
• introduce a function some_tuples in sage.misc.misc that is intensively used in the testing framework (and incidentally speed up some doc test)
• refine the category of Set([1,2,3]) to be finite
• implement a (very naive) random_element for Set([1,2,3])
• implement a (naive) hash for EnumeratedSetFromIterator

see also : #15425, #19195 one that can be closed as duplicate: #14224, #19192

Depends on #17411

CC: @nthiery @nathanncohen

Component: combinatorics

Author: Vincent Delecroix

Branch/Commit: 6e27dde

Reviewer: Nicolas M. Thiéry

Issue created by migration from https://trac.sagemath.org/ticket/18411

[videlec]

Description changed:

--- 
+++ 
@@ -1,3 +1,3 @@
-The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework. We just make `CartesianProduct` an alias for `sage.sets.cartesian_product.CartesianProduct` (with some tweak to support the behavior).
+The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We just make `CartesianProduct` an alias for `sage.sets.cartesian_product.CartesianProduct` (with some tweak to support the previous behavior which considered the cartesian product as a set of lists).

 `sage.combinat.cartesian_product.CartesianProduct` is deprecated.

[videlec]

Dependencies: #18290, #12955

[videlec]

Branch: public/18411

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Commit: e703eb4

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

e703eb4

Trac 18411: remove occurrences of CartesianProduct

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from e703eb4 to 10df218

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

635279f	Trac 18411: a bounded_number_of_tuples function
1b60903	Trac 18411: remove occurrences of CartesianProduct
bec57ae	Trac 18411: cartesian_product works with tuple/list/set/frozenset
7dbb663	Trac 18411: move iteration of cartesian products to Sets
4fbb4ae	Trac 18411: refine category of Set(...)
10df218	Trac 18411: fix combinat tutorial

[videlec]

Author: Vincent Delecroix

[videlec]

Description changed:

--- 
+++ 
@@ -1,3 +1,5 @@
-The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We just make `CartesianProduct` an alias for `sage.sets.cartesian_product.CartesianProduct` (with some tweak to support the previous behavior which considered the cartesian product as a set of lists).
+The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove most occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`.

-`sage.combinat.cartesian_product.CartesianProduct` is deprecated.
+follow up: #19192
+
+

[videlec]

Changed dependencies from #18290, #12955 to none

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

57b82e3

trax #18411 fixing a typo and doctest continuation

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 10df218 to 57b82e3

[videlec]

comment:7

Thanks Frédéric for fixing a doctest. Though, I do not see the point of modifying src/sage/rings/quotient_ring.py. This file is completely unrelated to this ticket.

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 57b82e3 to a640c24

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

a640c24

trac #18411 fixing some doctests

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

b650e69	trac #18411 fixing some doctests
0dcfed9	Trac 18411: fix combinat/partition_tuple
613cef5	Trac 18411: fix testing issues
3cc8070	Trac 18411: fix composition signed

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from a640c24 to 3cc8070

[videlec]

comment:10

Hi Frédéric,

This was clearly not the way to fix composition_signed.py. I removed this change from your commit. The rest of my commits should fix all issues with timed out testing (ie root system and padic) as well as the issue with tableau tuples.

Vincent

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

70949c7

Trac 18411: fix two doctests

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 3cc8070 to 70949c7

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

9beb023

Trac 18411: bound number of tests for Domains

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 70949c7 to 9beb023

[videlec]

Description changed:

--- 
+++ 
@@ -1,5 +1,10 @@
 The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove most occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`.

+We also:
+- move the `__iter__` from `EnumeratedSets.CartesianProducts.ParentMethods` category to `Sets`
+- allows `cartesian_product` to be called with `list`, `tuple`, `set`, `frozenset`
+- introduce a function `bounded_number_of_tuples` in `sage.misc.misc` that is intensively used in the testing framework
+- refine the category of `Set([1,2,3])` to be finite
+- implement a (very naive) `random_element` for `Set([1,2,3])`
+
 follow up: #19192
-
-

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

6580825

Trac 18411: fix is_empty

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 9beb023 to 6580825

[videlec]

comment:15

The only failing doctest is fine when run independently... I am not sure what may have happend.

Vincent

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 6580825 to 8a5ca01

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

af2c289	Trac 18411: use cartesian_product in CombinatorialFreeModule
8a5ca01	Trac 18411: remove two final occurrences

[videlec]

New commits:

af2c289	Trac 18411: use cartesian_product in CombinatorialFreeModule
8a5ca01	Trac 18411: remove two final occurrences

[videlec]

Description changed:

--- 
+++ 
@@ -1,6 +1,6 @@
-The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove most occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`.
+The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove all occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`. We also deprecate the `CartesianProduct` from `sage.combinat.cartesian_product`.

-We also:
+In order to support all features of the old class we also:
 - move the `__iter__` from `EnumeratedSets.CartesianProducts.ParentMethods` category to `Sets`
 - allows `cartesian_product` to be called with `list`, `tuple`, `set`, `frozenset`
 - introduce a function `bounded_number_of_tuples` in `sage.misc.misc` that is intensively used in the testing framework

[videlec]

comment:18

I was worried with CombinatorialFreeModule but it went smoothly. I just need to add a commit for the deprecation!

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 8a5ca01 to 4eacb22

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

cc3c250	Trac 18411: hash for EnumeratedSetFromIterator
4eacb22	Trac 18411: deprecate CartesianProduct

[videlec]

Description changed:

--- 
+++ 
@@ -6,5 +6,6 @@
 - introduce a function `bounded_number_of_tuples` in `sage.misc.misc` that is intensively used in the testing framework
 - refine the category of `Set([1,2,3])` to be finite
 - implement a (very naive) `random_element` for `Set([1,2,3])`
+- implement a (naive) hash for `EnumeratedSetFromIterator`

 follow up: #19192

[videlec]

Description changed:

--- 
+++ 
@@ -1,4 +1,4 @@
-The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove all occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`. We also deprecate the `CartesianProduct` from `sage.combinat.cartesian_product`.
+The features of `sage.combinat.cartesian_product.CartesianProduct` are now completely integrated into the category framework (see #18290). We remove all occurrences of `CartesianProduct` to either `cartesian_product` or `itertools.product`. We deprecate the `CartesianProduct` from `sage.combinat.cartesian_product`.

 In order to support all features of the old class we also:
 - move the `__iter__` from `EnumeratedSets.CartesianProducts.ParentMethods` category to `Sets`
@@ -8,4 +8,5 @@
 - implement a (very naive) `random_element` for `Set([1,2,3])`
 - implement a (naive) hash for `EnumeratedSetFromIterator`

-follow up: #19192
+see also : #15425
+one that can be closed as duplicate: #14224, #19192

[videlec]

comment:22

Annoying bug:

sage: F = FiniteEnumeratedSet([int(1), int(2)])
sage: F(1)
Traceback (most recent call last):
...
TypeError: Cannot convert int to sage.structure.element.Element

And the following is also a bug IMHO

sage: FiniteEnumeratedSet([1,2]) is FiniteEnumeratedSet([int(1),int(2)])
True

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 4eacb22 to 9f29074

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

f0dbc10	Trac 18411: deprecate CartesianProduct
596ad95	Trac 18411: a comment in combinat/free_module.py
9f29074	Trac 18411: a simplification in schur algebra

[videlec]

comment:24

I got into troubles. Combinatorial free modules are too laxist and you discover mistakes only very lately... in the following an element g is built

sage: M = SchurTensorModule(QQ, 2, 2)
sage: A = M._schur
sage: f = M.basis()[(1,1)]
sage: e = A.basis().values()[0]
sage: g = e*f

but it is not valid in the sense that the keys do not correspond to the basis

sage: map(type, g.monomial_coefficients().keys())
[<type 'tuple'>]
sage: g.parent().basis().keys()
The cartesian product of ({1, 2}, {1, 2})

And hence you silently got

sage: g.to_vector()
(0, 0, 0, 0)

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

d0f900a	Trac 18411: remove two final occurrences
8819c3e	Trac 18411: hash for EnumeratedSetFromIterator
4c50a1c	Trac 18411: more explicit call in CombinatorialFreeModule
321067f	Trac 18411: deprecate CartesianProduct

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 9f29074 to 321067f

[videlec]

comment:26

All right, my solution is to not use cartesian_product in combinatorial free module and postponed this step to #19195. Needs review again.

Vincent

[videlec]

Description changed:

--- 
+++ 
@@ -8,5 +8,5 @@
 - implement a (very naive) `random_element` for `Set([1,2,3])`
 - implement a (naive) hash for `EnumeratedSetFromIterator`

-see also : #15425
+see also : #15425, #19195
 one that can be closed as duplicate: #14224, #19192

[videlec]

comment:27

Patchbot's green light ;-)

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

c2eeac8

Trac 18411: removed unused import

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 321067f to c2eeac8

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from c2eeac8 to e7a2ff2

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

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

d269e6f	Trac 18411: use Sets.__call__ to make parents
e7a2ff2	Trac 18411: bounded_number_of_tuples -> some_tuples

[videlec]

comment:30

I just implemented the changes requested By Nicolas Thiéry while we discussed on phone.

[videlec]

Reviewer: Nicolas Thiéry