created util method to normalise http protocol in http path

Author: nikhilsuri-db

src/databricks/sql/common/url_utils.py

@@ -0,0 +1,32 @@
+"""
+URL utility functions for the Databricks SQL connector.
+"""
+
+
+def normalize_host_with_protocol(host: str) -> str:
+    """
+    Normalize a connection hostname by ensuring it has a protocol and removing trailing slashes.
+    
+    This is useful for handling cases where users may provide hostnames with or without protocols
+    (common with dbt-databricks users copying URLs from their browser).
+    
+    Args:
+        host: Connection hostname which may or may not include a protocol prefix (https:// or http://)
+              and may or may not have a trailing slash
+        
+    Returns:
+        Normalized hostname with protocol prefix and no trailing slash
+        
+    Examples:
+        normalize_host_with_protocol("myserver.com") -> "https://myserver.com"
+        normalize_host_with_protocol("https://myserver.com") -> "https://myserver.com"
+    """
+    # Remove trailing slash
+    host = host.rstrip("/")
+    
+    # Add protocol if not present
+    if not host.startswith("https://") and not host.startswith("http://"):
+        host = f"https://{host}"
+    
+    return host
+

tests/unit/test_url_utils.py

@@ -0,0 +1,41 @@
+"""Tests for URL utility functions."""
+import pytest
+from databricks.sql.common.url_utils import normalize_host_with_protocol
+
+
+class TestNormalizeHostWithProtocol:
+    """Tests for normalize_host_with_protocol function."""
+
+    @pytest.mark.parametrize("input_host,expected_output", [
+        # Hostname without protocol - should add https://
+        ("myserver.com", "https://myserver.com"),
+        ("workspace.databricks.com", "https://workspace.databricks.com"),
+        
+        # Hostname with https:// - should not duplicate
+        ("https://myserver.com", "https://myserver.com"),
+        ("https://workspace.databricks.com", "https://workspace.databricks.com"),
+        
+        # Hostname with http:// - should preserve
+        ("http://localhost", "http://localhost"),
+        ("http://myserver.com:8080", "http://myserver.com:8080"),
+        
+        # Hostname with port numbers
+        ("myserver.com:443", "https://myserver.com:443"),
+        ("https://myserver.com:443", "https://myserver.com:443"),
+        ("http://localhost:8080", "http://localhost:8080"),
+        
+        # Trailing slash - should be removed
+        ("myserver.com/", "https://myserver.com"),
+        ("https://myserver.com/", "https://myserver.com"),
+        ("http://localhost/", "http://localhost"),
+        
+        # Real Databricks patterns
+        ("adb-123456.azuredatabricks.net", "https://adb-123456.azuredatabricks.net"),
+        ("https://adb-123456.azuredatabricks.net", "https://adb-123456.azuredatabricks.net"),
+        ("https://adb-123456.azuredatabricks.net/", "https://adb-123456.azuredatabricks.net"),
+        ("workspace.databricks.com/", "https://workspace.databricks.com"),
+    ])
+    def test_normalize_host_with_protocol(self, input_host, expected_output):
+        """Test host normalization with various input formats."""
+        assert normalize_host_with_protocol(input_host) == expected_output
+