improve tag normalization

Author: dubloom

ddtrace/internal/process_tags/__init__.py

@@ -16,9 +16,14 @@
 
 log = get_logger(__name__)
 
+_INVALID_CHARS_PATTERN = re.compile(r"[^a-z0-9/._-]")
+_CONSECUTIVE_UNDERSCORES_PATTERN = re.compile(r"_{2,}")
+
 
 def normalize_tag(value: str) -> str:
-    return re.sub(r"[^a-z0-9/._-]", "_", value.lower())
+    normalized = _INVALID_CHARS_PATTERN.sub("_", value.lower())
+    normalized = _CONSECUTIVE_UNDERSCORES_PATTERN.sub("_", normalized)
+    return normalized.strip("_")
 
 
 def generate_process_tags() -> Optional[str]:

tests/internal/test_process_tags.py

@@ -14,16 +14,22 @@
     "input_tag,expected",
     [
         ("HelloWorld", "helloworld"),
-        ("Hello@World!", "hello_world_"),
+        ("Hello@World!", "hello_world"),
         ("HeLLo123", "hello123"),
         ("hello world", "hello_world"),
         ("a/b.c_d-e", "a/b.c_d-e"),
-        ("héllø", "h_ll_"),
+        ("héllø", "h_ll"),
         ("", ""),
-        ("💡⚡️", "___"),
-        ("!foo@", "_foo_"),
+        ("💡⚡️", ""),
+        ("!foo@", "foo"),
         ("123_abc.DEF-ghi/jkl", "123_abc.def-ghi/jkl"),
         ("Env:Prod-Server#1", "env_prod-server_1"),
+        ("__hello__world__", "hello_world"),
+        ("___test___", "test"),
+        ("_leading", "leading"),
+        ("trailing_", "trailing"),
+        ("double__underscore", "double_underscore"),
+        ("test\x99\x8faaa", "test_aaa"),
     ],
 )
 def test_normalize_tag(input_tag, expected):