didn't expect a remote URL

cpp/src/io/utilities/datasource.cpp

@@ -33,6 +33,7 @@
 #include <sys/mman.h>
 #include <unistd.h>
 
+#include <regex>
 #include <vector>
 
 namespace cudf {
@@ -389,6 +390,24 @@ class user_datasource_wrapper : public datasource {
   datasource* const source;  ///< A non-owning pointer to the user-implemented datasource
 };
 
+/**
+ * @brief Remote file source backed by KvikIO, which handles S3 filepaths seamlessly.
+ */
+class remote_file_source {
+ public:
+  /**
+   * @brief Is `url` referring to a remote file supported by KvikIO?
+   *
+   * For now, only S3 urls (urls starting with "s3://") are supported.
+   */
+  static bool is_supported_remote_url(std::string const& url)
+  {
+    // Regular expression to match "s3://"
+    std::regex pattern{R"(^s3://)", std::regex_constants::icase};
+    return std::regex_search(url, pattern);
+  }
+};
+
 }  // namespace
 
 std::unique_ptr<datasource> datasource::create(std::string const& filepath,
@@ -403,7 +422,8 @@ std::unique_ptr<datasource> datasource::create(std::string const& filepath,
 
     CUDF_FAIL("Invalid LIBCUDF_MMAP_ENABLED value: " + policy);
   }();
-
+  CUDF_EXPECTS(!remote_file_source::is_supported_remote_url(filepath),
+               "didn't expect a remote URL");
   if (use_memory_mapping) {
     return std::make_unique<memory_mapped_source>(filepath.c_str(), offset, max_size_estimate);
   } else {