add transformation matrix interface for BKZ

[user ctf]

Hi. sage-devel members.
I am kiona. I want to contribute devel for some math/crypto packages.

[background]

I created Coppersmith small root method packages: kionactf/coppersmith(github.com)
I am considering the package would be included in Sagemath tree.
But I am not welcomed if my codes affect Sagemath devel.
Actually, my codes depend on external packages(fplll, flatter).
It might violate Sagemath philosophy: should not reinvent wheel.

So I am starting to contribute Sagemath with a few related things.
My code uses some transformation matrix interface for BKZ,
so firstly I itry to add the feature on Sagemath code.

[main issue]
I want to add the feature that we can compute a transformation matrix on BKZ computation.
If I want to use the recent merge on fpylll(add transformation interface to bkz_reduction #267 · fplll/fpylll (github.com)) for adding the feature on Sagemath,
how could we start?
We need to wait next release on fpylll?

---
Best regards,
kiona

[Gareth Ma]

Regarding fplll and flatter, Sage already has an interface with fplll (via fpylll), and issue #37207 is adding an interface call to flatter (the new pari implementation).

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/9009f60a-bd2b-409a-8540-ac7c37d20c38n%40googlegroups.com.

[user ctf]

thanks for reply.

I am talking about transformation matrix interface.
LLL has transformation matrix interface:
LLL(delta=None, eta=None, algorithm='fpLLL:wrapper', fp=None, prec=0, early_red=False, use_givens=False, use_siegel=False, transformation=False, **kwds)
if set transformation=True, we have the output for transformation matrix.

However, BKZ does not have transformation matrix interface.
BKZ(delta=None, algorithm='fpLLL', fp=None, block_size=10, prune=0, use_givens=False, precision=0, proof=None, **kwds)

BKZ (sagemath.org)

Though we can extract transformation matrix manually after computing BKZ result, it would be annoying.
fplll have an interface for transformation matrix, so I think we had better use the interface (via fpylll).

2024年2月12日月曜日 10:50:58 UTC+9 Gareth Ma: