Deprecation warnings on KATs generation

[Dioprz]

Running make docker-sage-generate-kat within #255 's branch (in the meantime, while it's merged into develop)

/home/cryptographic_estimators/tests/kat_tests/external_estimators/reference_implementations/estimator.py:910: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract asingle element from your array before performing this operation. (Deprecated NumPy 1.25.)
  f = lambda x: 2 * log2((binom(p, p // 2) * binom(k // 2 - p, p1 - p // 2)) * (
/home/cryptographic_estimators/tests/kat_tests/external_estimators/reference_implementations/estimator.py:939: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract asingle element from your array before performing this operation. (Deprecated NumPy 1.25.)
  f = lambda x: log2(L12) + 2 * log2(binom_sp(x, w2) + 1) - 2 * x
/home/cryptographic_estimators/tests/kat_tests/external_estimators/reference_implementations/estimator.py:784: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract asingle element from your array before performing this operation. (Deprecated NumPy 1.25.)
  f = lambda x: log2((binom(p, p // 2) * binom_sp((k + x) / 2 - p, p1 - p // 2))) * 2 - x
/home/cryptographic_estimators/tests/kat_tests/external_estimators/reference_implementations/estimator.py:810: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract asingle element from your array before performing this operation. (Deprecated NumPy 1.25.)
  f = lambda x: log2(int(L12)) + int(2) * log2(binom_sp(x, int(w2))) - int(2) * x

[FloydZ]

diff --git a/tests/kat_tests/external_estimators/reference_implementations/estimator.py b/tests/kat_tests/external_estimators/reference_implementations/estimator.py
index a93549a..c289c41 100644
--- a/tests/kat_tests/external_estimators/reference_implementations/estimator.py
+++ b/tests/kat_tests/external_estimators/reference_implementations/estimator.py
@@ -9,6 +9,13 @@ from warnings import filterwarnings
 filterwarnings("ignore", category=RuntimeWarning)
 
 
+def extract(e):
+    try:
+        return e[0]
+    except:
+        return e
+
+
 def binom(n, k):
     return comb(int(n), int(k))
 
@@ -781,7 +788,7 @@ def bjmm_depth_2_partially_disjoint_weight_complexity(n, k, w, mem=inf, hmap=1,
                     ######choose start value for l1 close to the logarithm of the number of representations######
                     #############################################################################################
                     try:
-                        f = lambda x: log2((binom(p, p // 2) * binom_sp((k + x) / 2 - p, p1 - p // 2))) * 2 - x
+                        f = lambda x: log2((binom(p, p // 2) * binom_sp((k + extract(x)) / 2 - p, p1 - p // 2))) * 2 - extract(x)
                         l1_val = int(fsolve(f, 0)[0])
                     except:
                         continue
@@ -807,7 +814,7 @@ def bjmm_depth_2_partially_disjoint_weight_complexity(n, k, w, mem=inf, hmap=1,
                         #######choose start value for l2 such that resultlist size is close to L12#######
                         #################################################################################
                         try:
-                            f = lambda x: log2(int(L12)) + int(2) * log2(binom_sp(x, int(w2))) - int(2) * x
+                            f = lambda x: log2(int(L12)) + int(2) * log2(binom_sp(extract(x), int(w2))) - int(2) * extract(x)
                             l2_val = int(fsolve(f, 0)[0])
                         except:
                             continue
@@ -908,7 +915,7 @@ def bjmm_depth_2_disjoint_weight_complexity(n, k, w, mem=inf, hmap=1, p_range=[0
                             ##################################################################################
                             try:
                                 f = lambda x: 2 * log2((binom(p, p // 2) * binom(k // 2 - p, p1 - p // 2)) * (
-                                        binom_sp(x, w1 // 2) * binom_sp(x - w1, w11 - w1 // 2)) + 1) - 2 * x
+                                        binom_sp(extract(x), w1 // 2) * binom_sp(extract(x) - w1, w11 - w1 // 2)) + 1) - 2 * extract(x)
                                 l1_val = int(
                                     fsolve(f, 2 * log2((binom(p, p // 2) * binom(k // 2 - p, p1 - p // 2))))[0])
                             except:
@@ -936,7 +943,7 @@ def bjmm_depth_2_disjoint_weight_complexity(n, k, w, mem=inf, hmap=1, p_range=[0
                                 #######choose start value for l2 such that resultlist size is equal to L12#######
                                 #################################################################################
                                 try:
-                                    f = lambda x: log2(L12) + 2 * log2(binom_sp(x, w2) + 1) - 2 * x
+                                    f = lambda x: log2(L12) + 2 * log2(binom_sp(extract(x), w2) + 1) - 2 * extract(x)
                                     l2_val = int(fsolve(f, 50)[0])
                                 except:
                                     continue

this patch is doing it. A little bit weird, but thats what we are doing in the main code. Should I push this onto #255 ?

[Dioprz]

Why main has code that the develop branch doesn't have? It was a regression from the develop branch or something like that?

And I don't think it's necessary to merge to #255; at the end is just a warning, so we could keep it like that until the next release.

[FloydZ]

Why main has code that the develop branch doesn't have?

what do you mean? What code is in master but not in develop?

[Dioprz]

What I understood was that there is code in the main branch that works as a patch for the warnings we are having in #255 (and that branch comes from develop, and develop from a prior version of main). So my question is... We made a regression, or there is code in main that we don't have in develop? Or I misunderstand something?