From c366c05616fa67b2634fd4f894a7b8039870132c Mon Sep 17 00:00:00 2001
From: Csaba Kiraly <csaba.kiraly@gmail.com>
Date: Tue, 14 Mar 2023 15:26:48 +0100
Subject: [PATCH] handle overlap for multiple validators per node correctly

Signed-off-by: Csaba Kiraly <csaba.kiraly@gmail.com>
---
 DAS/tools.py     |  6 ++++++
 DAS/validator.py | 13 ++++---------
 2 files changed, 10 insertions(+), 9 deletions(-)

diff --git a/DAS/tools.py b/DAS/tools.py
index cd26850..2b47b11 100644
--- a/DAS/tools.py
+++ b/DAS/tools.py
@@ -79,3 +79,9 @@ def sampleLine(line, limit):
                         r[i] = 1
                         limit -= 1
                 return r
+
+def unionOfSamples(population, sampleSize, times):
+    selected = set()
+    for t in range(times):
+        selected |= set(random.sample(population, sampleSize))
+    return selected
diff --git a/DAS/validator.py b/DAS/validator.py
index b0fbede..9abb125 100644
--- a/DAS/validator.py
+++ b/DAS/validator.py
@@ -4,7 +4,7 @@ import random
 import collections
 import logging
 from DAS.block import *
-from DAS.tools import shuffled, shuffledDict
+from DAS.tools import shuffled, shuffledDict, unionOfSamples
 from bitarray.util import zeros
 from collections import deque
 from itertools import chain
@@ -60,12 +60,7 @@ class Validator:
         elif self.shape.chi > self.shape.blockSize:
             self.logger.error("Chi has to be smaller than %d" % self.shape.blockSize, extra=self.format)
         else:
-            if self.amIproposer:
-                self.chi = 1 # not used
-            elif self.ID <= shape.numberNodes * shape.class1ratio:
-                self.chi = shape.chi * shape.vpn1
-            else:
-                self.chi = shape.chi * shape.vpn2 # TODO: union of random subsets vpn2 times
+            self.chi = shape.chi
             if amIproposer:
                 self.rowIDs = range(shape.blockSize)
                 self.columnIDs = range(shape.blockSize)
@@ -75,11 +70,11 @@ class Validator:
                 if rows:
                     self.rowIDs = rows
                 else:
-                    self.rowIDs = random.sample(range(self.shape.blockSize), self.chi)
+                    self.rowIDs = unionOfSamples(range(self.shape.blockSize), self.chi, self.shape.vpn1)
                 if columns:
                     self.columnIDs = columns
                 else:
-                    self.columnIDs = random.sample(range(self.shape.blockSize), self.chi)
+                    self.columnIDs = unionOfSamples(range(self.shape.blockSize), self.chi, self.shape.vpn1)
         self.rowNeighbors = collections.defaultdict(dict)
         self.columnNeighbors = collections.defaultdict(dict)