BIBD from Transversal Designs

[ncohen]

This branch implements three constructions of BIBD from transversal designs. Short, efficient, AND useful for the general construction of BIBD with k=5.

As found in ​http://www.argilo.net/files/bibd.pdf

The constructions are explained in the documentation, too.

[git]

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

​054d2a2	trac #16227: Product construction of Transversal Designs
​a512ab1	trac #16227: Merged with updated #15310
​e62fae8	corrected doctests + new doctests
​4d6e964	trac #16227: Replace exception with booleans in the doctests
​a46446f	trac #16231: Equivalence between OA/TD/MOLS
​a9dce70	trac #16231: Merged with updated #16227
​8d8b928	more documentation to orthogonal_arrays.py
​8257178	remove MOLS construction for prime powers + doc
​d678326	trac #16272: Replacing availability by existence and forwarding the results between design constructors
​c881070	trac #16279: BIBD from Transversal Designs

[git]

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

​569c485

trac #16279: BIBD from Transversal Designs

[git]

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

​0812a73	trac #16281: Simplification
​61dc86b	16281: long comment for the construction of the projective plane
​51daa7f	trac #16281: correct a doctest
​e090f92	trac #16272: merge #16281
​9a221be	trac #16272: fix doctests
​5074eee	trac #16272: finer doctest to test the output of transversal_design
​d81f265	trac #16272: ultimate doctest
​47798d2	trac #16272: simplifying the structure of orthogonal_array
​d34b012	trac #16272: Merged with updated #16227
​60e8d35	trac #16279: Merged with updated #16272

[ncohen]

Still updating..

Nathann

[vdelecroix]

adding keyword...

[knsam]

Hi Nathann,

I would like to review this patch and I have the following remarks about your patch:

• The doc does not build and I have no idea how to fix these errors!
• Linguistic: A BIBD is a pair (X, B) where X is the set of points and B is the set of blocks. But, the description of the new BIBD as well as how the given BIBD is used in the construction does not make things clear: you say replace sth by a BIBD: what does this mean?

In view of these remarks, I am setting this to needs_work ;-).

[If not for the fact that does not build, I would have went ahead and made the other changes...]

-- Kannappan.

[ncohen]

That was because `(v,k,1)`-BIBD is okay for Sphinx, but `(v,k,1)-`BIBD is not. Which he clearly reported by saying "Something is wrong". I hate this thing.

Anyway, fixed.

I also avoided the word "replaced" using the same phrasing as in the two other constructions. I also replaced two occurrences of BIBD by BIBDs to make it clearer that we were talking of several BIBD.

I also merged #16091 in, as it is a dependency of this patch.

Nathann

[git]

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

​927d325	trac #16279: Merged with 6.2
​e1b29bf	trac #16279: Fixing the doc
​e722cb1	trac #16091: A missing obvious case for BIBD
​e941a54	trac #16279: Merged with updated #16091

[knsam]

Hi!

• My doc building stops at:

[tensor   ] build succeeded.

I have seen that there is a discussion on sage-devel. So, I will see what I can do to fix it. I will have this patch reviewed by evening tommorrow (for me!). Sorry for the delay, Nathann.

-- Kannappan.

---

New commits:

​927d325	trac #16279: Merged with 6.2
​e1b29bf	trac #16279: Fixing the doc
​e722cb1	trac #16091: A missing obvious case for BIBD
​e941a54	trac #16279: Merged with updated #16091

[git]

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

​c127a6d	trac #16272: Merged with updated #16231
​e2749b3	trac #16272: Merged with 6.3.beta0
​4115b72	trac #16279: Merged with updated #16272

[knsam]

Hi Nathann,

I just have one query: given that the method is called BIBD_From_TD, I was expecting the function to return a block design, but it returns a list instead. I also noted that this method is used in the BalancedIncompleteBlockDesign to construct a block design when existence = True. So, this is probably the desiderata, but I thought that this is worth pointing out...

You may set this to positive review if you would like to return a list instead of a BlockDesign...

This ticket reviewed using SMC!! :-)

--Kannappan.

[ncohen]

Helloooooooo !!!

I just have one query: given that the method is called BIBD_From_TD, I was expecting the function to return a block design, but it returns a list instead.

Hmmmm... Well, admittedly I do not like to wrap a list in a class if I don't need to (because doing so may cost non-trivial computations), but in this case the function is really only made to be called by the "main" method, i.e. the constructor of BIBD. It can be called recursively, and it would be nice if we don't convert all the stuff into a Block Design during those calls as we have no reason to ^^;

You may set this to positive review if you would like to return a list instead of a BlockDesign...

Thank you very much ! Now we are only one ticket away from having ALL bibd with k=5, and I have another code ready on my computer : Wilson's construction of MOLS with two truncated groups, which should make a big difference in our table of MOLS. But that's for a bit later, for many patches are still waiting to be merged and this is already sufficiently complicated as it is. Pretty soon, I hope !!!

Nathann