A real difference matrix has k columns

[ncohen]

Change Sage's standard to be coherent with OA which have k columns. This is much easier to work with on a computer, for it is much easier to add a row to a matrix than to add a column. And this is what we have to do most of the time.

Nathann

[git]

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

​5b8500b	trac #16864: designs/database.py: auto-generated doc index
​bfbad46	trac #16864: use Python string .format() + doc
​3ca48dc	trac #16817: OA for n=205,254,469,520,522,524,1262
​f9e6569	trac #16817: docfix
​c92e9bf	trac #16846: a difference_matrices module
​fdab06e	trac #16846: Remove obsolete functions
​f7afe6f	trac #16846: Broken doctests
​fd1bbc6	trac #16846: Merge with updated #16817
​c66c19b	trac #16846: review
​361c403	trac #16868: A real difference matrix has k columns

[vdelecroix]

I thought it was better standardization but in OA_from_quasi_difference_matrix you have

# Each line is expanded by [g+x for x in line for g in G] then relabeled
# with integers. Missing values are also handled.
new_M = []
for line in izip(*M):
    ...
...
new_M = zip(*new_M)

what's the point?

Vincent

[ncohen]

Replying to vdelecroix:

I thought it was better standardization but in OA_from_quasi_difference_matrix you have

# Each line is expanded by [g+x for x in line for g in G] then relabeled
# with integers. Missing values are also handled.
new_M = []
for line in izip(*M):
    ...
...
new_M = zip(*new_M)

what's the point?

Well, first if you only have one "zip*" in the whole library then the point is reached. I looked at that part of the code and I felt that the best way to implement it required two calls to zip, but for all others applications (functions in database.py) it is easier to work on this version of the matrix, for one constantly add rows, not columns.

So even if there is a "zip*" somewhere, the rest of the code is easier to write/read.

Nathann

[git]

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

​d6a2615

trac #16868: Mb was initialized twice

[vdelecroix]

A small commit where I removed a line.

Vincent

[ncohen]

I have non strong opposition against that :-P

Thanks !

Nathann

[vbraun]

Reviewer name

[vdelecroix]

Oh yeah!