fix: Fix incorrect indexing for determining eval prob in CB

[jackgerrits]

This fixes an incorrect assumption of the probs array being ordered by action index, the list must instead be searched for the desired index.

[jackgerrits]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[codecov]

Codecov Report

Merging #922 into master will increase coverage by 0.06%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master     #922      +/-   ##
==========================================
+ Coverage   84.90%   84.97%   +0.06%     
==========================================
  Files         193      193              
  Lines        9039     9041       +2     
  Branches      570      573       +3     
==========================================
+ Hits         7675     7683       +8     
+ Misses       1364     1358       -6     

Impacted Files	Coverage Δ
...oft/ml/spark/vw/VowpalWabbitContextualBandit.scala	89.61% <50.00%> (-1.18%)	⬇️
...osoft/ml/spark/core/schema/DatasetExtensions.scala	90.90% <0.00%> (-9.10%)	⬇️
...om/microsoft/ml/spark/core/utils/ClusterUtil.scala	62.50% <0.00%> (-6.95%)	⬇️
.../microsoft/ml/spark/core/schema/Categoricals.scala	82.29% <0.00%> (-4.17%)	⬇️
...a/com/microsoft/ml/spark/io/http/HTTPClients.scala	57.14% <0.00%> (-3.58%)	⬇️
...m/microsoft/ml/spark/core/schema/SparkSchema.scala	81.25% <0.00%> (-2.09%)	⬇️
.../apache/spark/ml/feature/FastVectorAssembler.scala	67.12% <0.00%> (-1.37%)	⬇️
...rosoft/ml/spark/train/ComputeModelStatistics.scala	86.40% <0.00%> (+0.97%)	⬆️
...osoft/ml/spark/io/http/PartitionConsolidator.scala	95.55% <0.00%> (+26.66%)	⬆️
...oft/ml/spark/isolationforest/IsolationForest.scala	83.33% <0.00%> (+66.66%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e9d8802...cca7df3. Read the comment docs.