fix: path extraction regex

Author: hatamiarash7

src/functions/extract_path.cpp

@@ -36,11 +36,12 @@ namespace duckdb
             // Explanation:
             // ^                - Start of the string
             // (?:              - Non-capturing group for the protocol and domain part
-            //   (?:(?:ftp|https?|rsync):\/\/)? - Optional ftp://, http://, https://, or rsync://
-            //   (?:[^\/\s]+)    - Domain name (any characters except '/' or whitespace)
+            //   (?:(?:ftp|https?|rsync):\/\/)? - Optional protocol (ftp://, http://, https://, or rsync://)
+            //   (?:[^\/\s]+)   - Domain name or IP address (any characters except '/' or whitespace)
             // )
-            // (\/[^?#]*)       - Capturing group for the path (starts with '/', followed by any characters except '?' or '#')
-            std::regex path_regex (R"(^(?:(?:(?:ftp|https?|rsync):\/\/)?(?:[^\/\s]+))(\/[^?#]*))");
+            // (\/[^?#]*)?      - Optional capturing group for the path (starts with '/', followed by any characters except '?' or '#')
+            //                  - The '?' at the end makes the path component optional, allowing the regex to match URLs with or without a path
+            std::regex path_regex (R"(^(?:(?:(?:ftp|https?|rsync):\/\/)?(?:[^\/\s]+))(\/[^?#]*)?)");
             std::smatch path_match;
 
             // Use regex_search to find the path component in the input string

test/sql/extract_path.test

@@ -7,6 +7,11 @@ require netquack
 statement ok
 CREATE TABLE uri_list AS SELECT * FROM read_csv('test/data/examples.csv', header=false, columns={'uri': 'VARCHAR'});
 
+query I
+SELECT extract_path('http://example.com.ac');
+----
+/
+
 query I
 SELECT extract_path('http://example.com.ac/path');
 ----