From c56859daf6a7fc05f1744a6d0845d7c16c0b288c Mon Sep 17 00:00:00 2001
From: Peter Hedenskog <peter@soulgalore.com>
Date: Thu, 7 Dec 2023 10:20:46 -0800
Subject: [PATCH] Make it easier to understand alternatives (#4023)

* Makes it easier to understand alternative settings

* make it easier to understand tests

* better sync
---
 lib/cli/cli.js                     |  4 ++--
 lib/plugins/compare/index.js       |  4 ++--
 lib/plugins/compare/pug/index.pug  |  6 ++++++
 lib/plugins/compare/statistical.py | 10 +++++-----
 4 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/lib/cli/cli.js b/lib/cli/cli.js
index 681e0a7388..e589d85989 100644
--- a/lib/cli/cli.js
+++ b/lib/cli/cli.js
@@ -1899,9 +1899,9 @@ export async function parseCommandLine() {
     })
     .option('compare.alternative', {
       choices: ['less', ' greater', 'two-sided'],
-      default: 'less',
+      default: 'greater',
       describe:
-        'Specifies the alternative hypothesis to be tested. Options are less for one-sided test where the first group is expected to be less than the second, greater for one-sided test with the first group expected to be greater, or two-sided for a two-sided test.',
+        'Specifies the alternative hypothesis to be tested. Default is greater than means current data is greater than the baseline. two-sided means we look for different both ways and less means current is less than baseline. ',
       group: 'compare'
     })
     .option('compare.wilcoxon.correction', {
diff --git a/lib/plugins/compare/index.js b/lib/plugins/compare/index.js
index b18e62be57..bf27fb9bd8 100644
--- a/lib/plugins/compare/index.js
+++ b/lib/plugins/compare/index.js
@@ -143,8 +143,8 @@ export default class ComparePlugin extends SitespeedioPlugin {
                     baselineMetric.getValues()
                   );
                   metricsInputData.metrics[group][metricName] = {
-                    sample1: baselineStats.data,
-                    sample2: currentStats.data
+                    baseline: baselineStats.data,
+                    current: currentStats.data
                   };
                 } else {
                   log.info(
diff --git a/lib/plugins/compare/pug/index.pug b/lib/plugins/compare/pug/index.pug
index cc9ccf9b2b..a94cee03bd 100644
--- a/lib/plugins/compare/pug/index.pug
+++ b/lib/plugins/compare/pug/index.pug
@@ -17,6 +17,12 @@ p
 h2 Settings
 p 
   | The test conducted in this comparison is the #{compare.meta.testOptions.testType} test. The alternative hypothesis used for this test is "#{compare.meta.testOptions.alternative}".
+  if compare.meta.testOptions.alternative === 'less'
+    | This means that we test if the current test is significant less that the baseline tests.
+  else if compare.meta.testOptions.alternative === 'greater'
+    | This means that we test if the current test is siginficant higher than the baseline tests.
+  else if compare.meta.testOptions.alternative === 'two-sided'
+    | This means that we test the baseline vs the current tests both ways, if there are any change in both directions.
   if compare.meta.testOptions.testType === 'mannwhitneyu'
     | For more information on the settings of the Mann-Whitney U test, please refer to the 
     a(href='https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.mannwhitneyu.html') official documentation.
diff --git a/lib/plugins/compare/statistical.py b/lib/plugins/compare/statistical.py
index 97a53fd4fd..29ca1ce6d3 100644
--- a/lib/plugins/compare/statistical.py
+++ b/lib/plugins/compare/statistical.py
@@ -6,15 +6,15 @@ def has_variability(sample):
     """Check if the sample has more than one unique value."""
     return len(set(sample)) > 1
 
-def perform_test(test_type, sample1, sample2, **kwargs):
+def perform_test(test_type, baseline, current, **kwargs):
     """Perform the statistical test based on the test type."""
-    if not has_variability(sample1) or not has_variability(sample2):
+    if not has_variability(baseline) or not has_variability(current):
         return None, "No variability"
 
     if test_type == 'wilcoxon':
-        return wilcoxon(sample1, sample2, **kwargs)
+        return wilcoxon(current, baseline, **kwargs)
     elif test_type == 'mannwhitneyu':
-        return mannwhitneyu(sample1, sample2, **kwargs)
+        return mannwhitneyu(current, baseline, **kwargs)
     else:
         raise ValueError("Invalid test type. Choose 'wilcoxon' or 'mannwhitneyu'.")
 
@@ -27,7 +27,7 @@ def perform_test(test_type, sample1, sample2, **kwargs):
 for group_name, metrics in input_data['metrics'].items():
     group_results = {}
     for metric_name, metric_data in metrics.items():
-        stat, p = perform_test(test_type, metric_data['sample1'], metric_data['sample2'], **options)
+        stat, p = perform_test(test_type, metric_data['baseline'], metric_data['current'], **options)
         if p == "No variability":
             group_results[metric_name] = {'statistic': "N/A", 'p-value': "N/A"}
         else: