From 6251260d14beb84bebc200a795e07adad393672b Mon Sep 17 00:00:00 2001
From: Vincent Whitchurch <vincent.whitchurch@datadoghq.com>
Date: Thu, 28 Nov 2024 12:28:27 +0100
Subject: [PATCH 1/2] discovery: Tweak Python name heuristics

Add a couple of small improvements to the Python name heuristics based
on real data seen during testing:

 - If the directory name is not useful ("/" or "bin"), for example,
   try falling back to the filename.

 - Ignore the -u flag while skipping arguments since some cases
   were seen with "python -u foo" which ended up not generating
   a name.
---
 .../corechecks/servicediscovery/usm/python.go | 13 ++++++--
 .../servicediscovery/usm/python_test.go       | 30 +++++++++++++++++--
 2 files changed, 38 insertions(+), 5 deletions(-)

diff --git a/pkg/collector/corechecks/servicediscovery/usm/python.go b/pkg/collector/corechecks/servicediscovery/usm/python.go
index 16d774c191e010..44b2c02309463a 100644
--- a/pkg/collector/corechecks/servicediscovery/usm/python.go
+++ b/pkg/collector/corechecks/servicediscovery/usm/python.go
@@ -78,14 +78,23 @@ func (p pythonDetector) detect(args []string) (ServiceMetadata, bool) {
 			if value, ok := p.deducePackageName(path.Clean(stripped), filename); ok {
 				return NewServiceMetadata(value), true
 			}
-			return NewServiceMetadata(p.findNearestTopLevel(stripped)), true
+
+			name := p.findNearestTopLevel(stripped)
+			// If we have generic/useless directory names, fallback to the filename.
+			if name == "." || name == "/" || name == "bin" || name == "sbin" {
+				name = p.findNearestTopLevel(filename)
+			}
+
+			return NewServiceMetadata(name), true
 		}
 
 		if hasFlagPrefix && a == "-m" {
 			moduleFlag = true
 		}
 
-		prevArgIsFlag = hasFlagPrefix
+		// The -u (unbuffered) option doesn't take an argument so we should
+		// consider the next arg even though this one is a flag.
+		prevArgIsFlag = hasFlagPrefix && a != "-u"
 	}
 
 	return ServiceMetadata{}, false
diff --git a/pkg/collector/corechecks/servicediscovery/usm/python_test.go b/pkg/collector/corechecks/servicediscovery/usm/python_test.go
index cfa80ae32413a7..d5f0c42acb3fee 100644
--- a/pkg/collector/corechecks/servicediscovery/usm/python_test.go
+++ b/pkg/collector/corechecks/servicediscovery/usm/python_test.go
@@ -16,8 +16,9 @@ import (
 )
 
 func TestPythonDetect(t *testing.T) {
-	//prepare the in mem fs
-	memFs := fstest.MapFS{
+	// Wrap the MapFS in a SubDirFS since we want to test absolute paths and the
+	// former doesn't allow them in calls to Open().
+	memFs := SubDirFS{FS: fstest.MapFS{
 		"modules/m1/first/nice/package/__init__.py": &fstest.MapFile{},
 		"modules/m1/first/nice/__init__.py":         &fstest.MapFile{},
 		"modules/m1/first/nice/something.py":        &fstest.MapFile{},
@@ -28,7 +29,9 @@ func TestPythonDetect(t *testing.T) {
 		"apps/app2/cmd/run.py":                      &fstest.MapFile{},
 		"apps/app2/setup.py":                        &fstest.MapFile{},
 		"example.py":                                &fstest.MapFile{},
-	}
+		"usr/bin/pytest":                            &fstest.MapFile{},
+		"bin/WALinuxAgent.egg":                      &fstest.MapFile{},
+	}}
 	tests := []struct {
 		name     string
 		cmd      string
@@ -79,6 +82,27 @@ func TestPythonDetect(t *testing.T) {
 			cmd:      "python example.py",
 			expected: "example",
 		},
+		{
+			name:     "root level script",
+			cmd:      "python /example.py",
+			expected: "example",
+		},
+		{
+			name: "root level script with ..",
+			// This results in a path of "." after findNearestTopLevel is called on the split path.
+			cmd:      "python /../example.py",
+			expected: "example",
+		},
+		{
+			name:     "script in bin",
+			cmd:      "python /usr/bin/pytest",
+			expected: "pytest",
+		},
+		{
+			name:     "script in bin with -u",
+			cmd:      "python3 -u bin/WALinuxAgent.egg",
+			expected: "WALinuxAgent",
+		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

From a8dddfe08f5977bc43f829741d9cec7aeb210214 Mon Sep 17 00:00:00 2001
From: Vincent Whitchurch <vincent.whitchurch@datadoghq.com>
Date: Thu, 28 Nov 2024 15:00:08 +0100
Subject: [PATCH 2/2] Disable Python test on Windows

The SubDirFs stuff doesn't work on Windows so just disable the tests
there.
---
 pkg/collector/corechecks/servicediscovery/usm/python_test.go | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/pkg/collector/corechecks/servicediscovery/usm/python_test.go b/pkg/collector/corechecks/servicediscovery/usm/python_test.go
index d5f0c42acb3fee..631c843e8a9d7f 100644
--- a/pkg/collector/corechecks/servicediscovery/usm/python_test.go
+++ b/pkg/collector/corechecks/servicediscovery/usm/python_test.go
@@ -3,6 +3,8 @@
 // This product includes software developed at Datadog (https://www.datadoghq.com/).
 // Copyright 2016-present Datadog, Inc.
 
+//go:build !windows
+
 package usm
 
 import (