Opened 11 years ago
Last modified 8 years ago
#5457 closed enhancement
Refactor symmetric functions and k-bounded subspace — at Version 58
Reported by: | nthiery | Owned by: | mhansen |
---|---|---|---|
Priority: | major | Milestone: | sage-5.4 |
Component: | combinatorics | Keywords: | symmetric functions, days38, sd40 |
Cc: | sage-combinat, saliola, bump, chrisjamesberg, zabrocki, SimonKing | Merged in: | |
Authors: | Mike Zabrocki, Anne Schilling, Jason Bandlow | Reviewers: | Dan Bump, Nicolas M. Thiéry |
Report Upstream: | N/A | Work issues: | |
Branch: | Commit: | ||
Dependencies: | #11563, #13109, #12969 | Stopgaps: |
Description (last modified by )
This patch restructures the implementation of symmetric functions in sage
The new implementation makes use of multiple realizations and the category framework. The new access to symmetric functions is via
sage: Sym = SymmetricFunctions(QQ)
Further new features that are implemented:
- The ring of symmetric functions is now endowed with a Hopf algebra structure. The coproduct and antipode are implemented (which were missing before).
- A tutorial on how to use symmetric functions in sage is included at the beginning of sf.py which is also accessible via
sage: SymmetricFunctions??
- Symmetric functions should now work a lot better with respect to specializing parameters like
q
andt
for Hall-Littlewood, Jack and Macdonald symmetric functions. Certain functionalities before this change were broken or not possible.
- Documentation was added to LLT polynomials (which had very sparse documentation previously).
- The
k
-bounded subspace of the ring of symmetric function was implemented. Thek
-Schur functions now live in thek
-bounded subspace rather than in the ring of symmetric functions as before.
This patch gained tremendously by the tutorial on symmetric functions written by Jason Bandlow, a draft on the k
-bounded subspace by Jason Bandlow, and code multiple realizations written by Franco Saliola.
See also http://groups.google.com/group/sage-devel/msg/a49f3288fca1b75c
Apply
Change History (58)
comment:1 Changed 11 years ago by
- Description modified (diff)
comment:2 Changed 11 years ago by
- Cc sage-combinat added
comment:3 Changed 8 years ago by
- Dependencies set to 13109
- Report Upstream set to N/A
comment:4 Changed 8 years ago by
- Dependencies 13109 deleted
comment:5 Changed 8 years ago by
- Description modified (diff)
- Keywords symmetric functions sd38 sd40 added
- Reviewers set to Dan Bump, Franco Saliola
- Summary changed from Refactor symmetric functions to Refactor symmetric functions and k-bounded subspace
comment:6 Changed 8 years ago by
- Cc saliola bump chrisjamesberg added
comment:7 Changed 8 years ago by
- Cc zabrocki added
comment:8 Changed 8 years ago by
- Status changed from new to needs_review
comment:9 Changed 8 years ago by
- Status changed from needs_review to needs_work
comment:10 Changed 8 years ago by
comment:11 Changed 8 years ago by
comment:12 Changed 8 years ago by
- Dependencies set to #11563
comment:13 Changed 8 years ago by
- Dependencies changed from #11563 to #11563, #13109
comment:14 follow-up: ↓ 15 Changed 8 years ago by
- Description modified (diff)
comment:15 in reply to: ↑ 14 ; follow-up: ↓ 16 Changed 8 years ago by
Hi Mike,
I completed the doctests for sfa.py and also rebased everything on top of 13109. Please put your changes to
- hall_littlewood.py
- jack.py
- llt.py
- macdonald.py
on top of the current patch trac_5457-symmetric_functions-mz.patch. Unfortunately we need to abandon the sage-combinat queue for the moment since it would be very cumbersome to keep it backward compatible with 13109. I will send you a separate e-mail on how to proceed.
Cheers,
Anne
comment:16 in reply to: ↑ 15 Changed 8 years ago by
Ok, patch is ready for review! It should apply and run cleanly on sage.5.2.rc0!
Anne
comment:17 Changed 8 years ago by
- Status changed from needs_work to needs_review
comment:18 Changed 8 years ago by
comment:19 Changed 8 years ago by
- Cc SimonKing added
comment:20 follow-up: ↓ 21 Changed 8 years ago by
Hi Dan!
Thank you very much for your comments on the failing doctests in
- devel/sage/sage/algebras/nil_coxeter_algebra.py
- devel/sage/sage/categories/realizations.py
They are fixed in the updated version of the patch. I do not get failures for
- devel/sage/sage/sandpiles/sandpile.py
on my machine.
lolita-4:sandpiles anne$ sage -t sandpile.py sage -t "devel/sage-sf/sage/sandpiles/sandpile.py"
[19.0 s]
All tests passed! Total time for all tests: 19.0 seconds
Anne
comment:21 in reply to: ↑ 20 Changed 8 years ago by
Replying to aschilling:
Hi Dan!
Thank you very much for your comments on the failing doctests in
- devel/sage/sage/algebras/nil_coxeter_algebra.py
- devel/sage/sage/categories/realizations.py
They are fixed in the updated version of the patch. I do not get failures for
- devel/sage/sage/sandpiles/sandpile.py
on my machine.
I also get a doctest failure in sandpile.py with unpatched sage-5.2.rc0 so this failure is not caused by the patch.
comment:22 Changed 8 years ago by
Applies cleanly to sage-5.2 and passes all tests.
comment:23 follow-up: ↓ 24 Changed 8 years ago by
- Description modified (diff)
comment:24 in reply to: ↑ 23 Changed 8 years ago by
The attached review patch trac_5457-review-as.patch incorporates most of the comments that Dan Bump raised in e-mail conversations.
Anne
comment:25 follow-up: ↓ 26 Changed 8 years ago by
- Description modified (diff)
- Status changed from needs_review to positive_review
This patch is a huge step forward for symmetric functions.
In addition to normal testing I spent quite a bit of time and privately sent comments (mainly on documentation) that have been taken into account in trac_5457-review-as.patch. I'm changing the status to positive review.
comment:26 in reply to: ↑ 25 Changed 8 years ago by
Replying to bump:
This patch is a huge step forward for symmetric functions.
In addition to normal testing I spent quite a bit of time and privately sent comments (mainly on documentation) that have been taken into account in trac_5457-review-as.patch. I'm changing the status to positive review.
Dear Dan, Thank you so much for your thorough and quick review of this huge patch! Mike and I just finished the review patch. Tests pass on both of our machines.
Anne
comment:27 follow-up: ↓ 28 Changed 8 years ago by
I have reviewed the latest version of the patch and it still has positive review.
comment:28 in reply to: ↑ 27 Changed 8 years ago by
Since http://trac.sagemath.org/sage_trac/ticket/12969 just got merged into sage-5.3.beta0, please also apply the attachment trac12969_rel_5457.patch on the ticket 12969 to this patch. Otherwise there will be doctest failures.
Thanks,
Anne
comment:29 Changed 8 years ago by
- Dependencies changed from #11563, #13109 to #11563, #13109, #12969
comment:30 Changed 8 years ago by
- Description modified (diff)
comment:31 Changed 8 years ago by
This needs to be rebased to sage-5.3.beta0 (not yet released):
patching file sage/categories/realizations.py Hunk #1 FAILED at 74 1 out of 1 hunks FAILED -- saving rejects to file sage/categories/realizations.py.rej patching file sage/combinat/sf/classical.py Hunk #3 FAILED at 88 1 out of 9 hunks FAILED -- saving rejects to file sage/combinat/sf/classical.py.rej patching file sage/combinat/sf/sfa.py Hunk #61 FAILED at 2589 1 out of 63 hunks FAILED -- saving rejects to file sage/combinat/sf/sfa.py.rej abort: patch failed to apply
comment:32 Changed 8 years ago by
- Status changed from positive_review to needs_work
comment:33 follow-up: ↓ 34 Changed 8 years ago by
- Description modified (diff)
comment:34 in reply to: ↑ 33 Changed 8 years ago by
I rebased and folded the three patches with respect to sage-5.3.beta0 from yesterday. It should apply cleanly now.
Mike and I are still going to fix one math bug that someone at FPSAC found.
Anne
comment:35 Changed 8 years ago by
Hi Dan,
Our new rebased patch is attached. You only need to apply trac_5457-symmetric_functions-mz.patch. Note that we did not fix apply_linear_morphism in /category/module_with_basis.py yet since Nicolas seems to have a review patch for this, but unfortunately his tests did not pass. Either he needs to fix his review patch or we will add your suggestion from the e-mail.
Anne
comment:36 Changed 8 years ago by
- Status changed from needs_work to needs_review
comment:37 follow-up: ↓ 38 Changed 8 years ago by
Note: I inserted in the queue the latest patch rebased for 5.3 beta0 from here. My review patch is also there, ready to be folded if you are happy with it.
comment:38 in reply to: ↑ 37 Changed 8 years ago by
Dear Dan and Nicolas,
Thank you so much for your reviews and work on this patch! I incorporated the changes that Dan suggested by e-mail and folded Nicolas' review patch. In addition, Mike had some minor improvements in the documentation of llt.py which are incorporated.
The new patch should apply cleanly on sage-5.3.beta0.
Anne
comment:39 Changed 8 years ago by
- Reviewers changed from Dan Bump, Franco Saliola to Dan Bump, Nicolas M. Thiery
comment:40 Changed 8 years ago by
All tests pass with sage-5.3.beta0. The changes discussed over the last few days have all been incorporated. I think this is ready to go.
comment:41 Changed 8 years ago by
- Status changed from needs_review to positive_review
comment:42 Changed 8 years ago by
- Description modified (diff)
comment:43 Changed 8 years ago by
- Status changed from positive_review to needs_work
This fails on arando (32-bit i386 Linux):
sage -t --long "devel/sage/sage/combinat/sf/llt.py" ********************************************************************** File "/var/lib/buildbot/build/sage/arando-1/arando_full/build/sage-5.3.beta1/devel/sage/sage/combinat/sf/llt.py", line 329: sage: cmp(L3Q, L3Z) Expected: -1 Got: 1 **********************************************************************
comment:44 follow-up: ↓ 45 Changed 8 years ago by
- Description modified (diff)
- Status changed from needs_work to needs_review
comment:45 in reply to: ↑ 44 Changed 8 years ago by
Hi Jeroen,
Hmm, this is hard for us to check since we are not running our code on that operating system. We attached a patch which will hopefully fix the problem
Apply
Thanks,
Anne
comment:46 Changed 8 years ago by
- Status changed from needs_review to positive_review
comment:47 Changed 8 years ago by
- Reviewers changed from Dan Bump, Nicolas M. Thiery to Dan Bump, Nicolas M. Thiéry
- Status changed from positive_review to needs_work
comment:48 Changed 8 years ago by
- Status changed from needs_work to needs_review
I think one of the original reviewers should review this additional patch.
comment:49 Changed 8 years ago by
- Status changed from needs_review to positive_review
I think one of the original reviewers should review this additional patch.
This adds a cmp method for SymmetricFunctions?, which was missing, and that was uncovered by the test in llt.py. Two SymmetricFunctions? instances are compared equal if and only if they have the same base ring, which is as it should be.
Unless I'm missing something the patch is obviously correct. I ran --testall --long in the sf directory and all tests passed.
comment:50 Changed 8 years ago by
There is no compelling reason to have a cmp function for symmetric functions (nor its bases). The order has no meaning, and equality testing should be taken care of by UniqueRepresentation?.
So altogether, I'd rather not add a cmp function, and instead would rather replace the failing test by:
sage: cmp(L3Q, L3Z) != 0
which is platform independent, and is all we care about. We could even just discard this test.
Now, I don't want to slow down the integration of this patch, so I am happy leaving this issue for a latter ticket, at the author's choice.
Cheers,
Nicolas
comment:51 Changed 8 years ago by
- Status changed from positive_review to needs_work
As per Nicolas' suggestion, we are deleting the cmp function from llt and moving and modifying the doctests elsewhere in the llt.py file.
comment:52 follow-up: ↓ 53 Changed 8 years ago by
- Description modified (diff)
- Status changed from needs_work to needs_review
This new patch deletes the function cmp from llt.py and inserts doctests into init in llt.py
comment:53 in reply to: ↑ 52 Changed 8 years ago by
sage -testall passes on both mine and Mike's machine (both MacOS) with trac_5457-symmetric_functions-mz.patch and trac_5457_docfix2-mz.patch applied.
Dan and/or Nicolas, please set a positive review if you are happy with the changes.
Thanks!
Anne
comment:54 Changed 8 years ago by
- Status changed from needs_review to positive_review
Nicolas argued that there is no real reason for Sym to have a __cmp__
method and there is no reason for the test to be deterministic. We don't
care if two rings or ring elements are > or < than each other, just whether they are distinct.
Therefore the test that failed can be rewritten and the __cmp___
method
is not needed.
This patch removes the __cmp__
method from llt polynomials instead
of adding one for sf. In fact, SymmetricFunctions
should not have had a
__cmp__
method because it already inherits one from UniqueRepresentation
.
(Tested!) On the other hand, there is no reason for the test as previously written
to be deterministic. So this is the correct approach.
comment:55 Changed 8 years ago by
- Status changed from positive_review to needs_work
comment:56 follow-up: ↓ 57 Changed 8 years ago by
- Status changed from needs_work to positive_review
comment:57 in reply to: ↑ 56 Changed 8 years ago by
Thank you Dan for the review!
Anne
comment:58 Changed 8 years ago by
- Description modified (diff)
Hi Mike,
I finished the doctests for the following files:
In particular, at the beginning of sf.py I incorporated the tutorial that Jason and Nicolas wrote (which was further down the queue) and updated it. I marked them there as coauthors in that file.
This leaves the doctests for
which I suppose you will do in the next couple of days? In particular, in the sfa.py the deprecation warnings need to be activated which I have not yet done.
Best,
Anne