Wrap fmpq_reconstruct_fmpz_2..

[sumiya11]

..because it can be faster when the bound is precomputed.

using Nemo, BenchmarkTools, Primes
m = ZZRingElem(prod(nextprimes(BigInt(2)^30, 1_000)));
n, d = ZZRingElem(1), ZZRingElem(100)
N = D = isqrt((m >> 1) - 1);
a = mod(n * invmod(d, m), m);

@btime reconstruct($a,$m)
#  17.160 μs (17 allocations: 29.23 KiB)

@btime reconstruct2($a,$m,$N,$N)
#  4.656 μs (15 allocations: 25.58 KiB)

BTW, perhaps Nemo.reconstruct can be changed to do the same as Nemo.unsafe_reconstruct (and Nemo.unsafe_reconstruct be removed).

[fingolfin]

Why reconstruct2 and not reconstruct?

[sumiya11]

No particular reason other than the name of the wrapped function, fmpq_reconstruct_fmpz_2

[thofma]

Can you rename it to reconstruct? The method names are not translated 1:1 from flint and it it just another method for reconstruct. Once renamed, I'll merge and tag a new release.

[sumiya11]

Sure; done; thanks.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.59%. Comparing base (b04719f) to head (bb189c0).
Report is 4 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1760      +/-   ##
==========================================
+ Coverage   85.48%   85.59%   +0.11%     
==========================================
  Files          95       95              
  Lines       36923    36921       -2     
==========================================
+ Hits        31565    31604      +39     
+ Misses       5358     5317      -41     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.