From 5c31bec8c20626c0fb2558f08145ca1ab46e85be Mon Sep 17 00:00:00 2001
From: "J. Nick Koston" <nick@koston.org>
Date: Fri, 24 Nov 2023 00:29:08 +0100
Subject: [PATCH] Ensure names containing characters other than `a-z` `A-Z`
 `0-9` or `_` are unique (#5810)

---
 esphome/core/helpers.cpp         | 11 +++++++----
 esphome/helpers.py               |  5 ++++-
 tests/unit_tests/test_helpers.py |  6 +++---
 3 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/esphome/core/helpers.cpp b/esphome/core/helpers.cpp
index 714a1642f8..c95c0470de 100644
--- a/esphome/core/helpers.cpp
+++ b/esphome/core/helpers.cpp
@@ -278,10 +278,13 @@ std::string str_snake_case(const std::string &str) {
   return result;
 }
 std::string str_sanitize(const std::string &str) {
-  std::string out;
-  std::copy_if(str.begin(), str.end(), std::back_inserter(out), [](const char &c) {
-    return c == '-' || c == '_' || (c >= '0' && c <= '9') || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z');
-  });
+  std::string out = str;
+  std::replace_if(
+      out.begin(), out.end(),
+      [](const char &c) {
+        return !(c == '-' || c == '_' || (c >= '0' && c <= '9') || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'));
+      },
+      '_');
   return out;
 }
 std::string str_snprintf(const char *fmt, size_t len, ...) {
diff --git a/esphome/helpers.py b/esphome/helpers.py
index 4012b2067f..00416b591f 100644
--- a/esphome/helpers.py
+++ b/esphome/helpers.py
@@ -357,6 +357,9 @@ def snake_case(value):
     return value.replace(" ", "_").lower()
 
 
+_DISALLOWED_CHARS = re.compile(r"[^a-zA-Z0-9_]")
+
+
 def sanitize(value):
     """Same behaviour as `helpers.cpp` method `str_sanitize`."""
-    return re.sub("[^-_0-9a-zA-Z]", r"", value)
+    return _DISALLOWED_CHARS.sub("_", value)
diff --git a/tests/unit_tests/test_helpers.py b/tests/unit_tests/test_helpers.py
index 67fabd7af8..79d39901f0 100644
--- a/tests/unit_tests/test_helpers.py
+++ b/tests/unit_tests/test_helpers.py
@@ -258,9 +258,9 @@ def test_snake_case(text, expected):
     "text, expected",
     (
         ("foo_bar", "foo_bar"),
-        ('!"§$%&/()=?foo_bar', "foo_bar"),
-        ('foo_!"§$%&/()=?bar', "foo_bar"),
-        ('foo_bar!"§$%&/()=?', "foo_bar"),
+        ('!"§$%&/()=?foo_bar', "___________foo_bar"),
+        ('foo_!"§$%&/()=?bar', "foo____________bar"),
+        ('foo_bar!"§$%&/()=?', "foo_bar___________"),
     ),
 )
 def test_sanitize(text, expected):