feat: support https azure OS urls

rust/lance-io/src/object_store/providers.rs

@@ -235,7 +235,11 @@ impl Default for ObjectStoreRegistry {
             providers.insert("s3+ddb".into(), aws);
         }
         #[cfg(feature = "azure")]
-        providers.insert("az".into(), Arc::new(azure::AzureBlobStoreProvider));
+        {
+            let azure = Arc::new(azure::AzureBlobStoreProvider);
+            providers.insert("az".into(), azure.clone());
+            providers.insert("https".into(), azure);
+        }
         #[cfg(feature = "gcp")]
         providers.insert("gs".into(), Arc::new(gcp::GcsStoreProvider));
         #[cfg(feature = "oss")]

rust/lance-io/src/object_store/providers/azure.rs

@@ -10,6 +10,7 @@ use snafu::location;
 
 use object_store::{
     azure::{AzureConfigKey, MicrosoftAzureBuilder},
+    path::Path,
     RetryConfig,
 };
 use url::Url;
@@ -87,6 +88,20 @@ impl AzureBlobStoreProvider {
 
 #[async_trait::async_trait]
 impl ObjectStoreProvider for AzureBlobStoreProvider {
+    fn extract_path(&self, url: &Url) -> Result<Path> {
+        // Azure https paths in ObjectSore encode the container name as the first path segment.
+        // The actual object path starts from the second segment.
+        if url.scheme() == "https" {
+            url.path_segments()
+                .map(|s| Path::from_iter(s.skip(1)))
+                .ok_or_else(|| {
+                    Error::invalid_input(format!("Invalid Azure URL: {url}"), location!())
+                })
+        } else {
+            Ok(Path::from(url.path()))
+        }
+    }
+
     async fn new_store(&self, base_path: Url, params: &ObjectStoreParams) -> Result<ObjectStore> {
         let block_size = params.block_size.unwrap_or(DEFAULT_CLOUD_BLOCK_SIZE);
         let mut storage_options =
@@ -164,6 +179,16 @@ mod tests {
         assert_eq!(path, expected_path);
     }
 
+    #[test]
+    fn test_azure_store_https_path() {
+        let provider = AzureBlobStoreProvider;
+
+        let url = Url::parse("https://account.blob.core.windows.net/bucket/path/to/file").unwrap();
+        let path = provider.extract_path(&url).expect("Failed to extract path");
+        let expected_path = object_store::path::Path::from("path/to/file");
+        assert_eq!(path, expected_path);
+    }
+
     #[tokio::test]
     async fn test_use_opendal_flag() {
         let provider = AzureBlobStoreProvider;