feat: exposed file_size_in_bytes

crates/iceberg/src/arrow/reader.rs

@@ -1523,6 +1523,7 @@ message schema {
                 deletes: vec![],
                 sequence_number: 0,
                 equality_ids: vec![],
+                file_size_in_bytes: 0,
             })]
             .into_iter(),
         )) as FileScanTaskStream;

crates/iceberg/src/scan/context.rs

@@ -143,6 +143,7 @@ impl ManifestEntryContext {
             deletes,
             sequence_number: self.manifest_entry.sequence_number().unwrap_or(0),
             equality_ids: self.manifest_entry.data_file().equality_ids().to_vec(),
+            file_size_in_bytes: self.manifest_entry.data_file().file_size_in_bytes(),
         })
     }
 }

crates/iceberg/src/scan/mod.rs

@@ -1795,6 +1795,7 @@ pub mod tests {
             deletes: vec![],
             sequence_number: 0,
             equality_ids: vec![],
+            file_size_in_bytes: 0,
         };
         test_fn(task);
 
@@ -1812,6 +1813,7 @@ pub mod tests {
             deletes: vec![],
             sequence_number: 0,
             equality_ids: vec![],
+            file_size_in_bytes: 0,
         };
         test_fn(task);
     }

crates/iceberg/src/scan/task.rs

@@ -63,6 +63,9 @@ pub struct FileScanTask {
     pub sequence_number: i64,
     /// equality ids
     pub equality_ids: Vec<i32>,
+
+    /// The size of the file in bytes.
+    pub file_size_in_bytes: u64,
 }
 
 impl FileScanTask {
@@ -127,6 +130,7 @@ impl From<&DeleteFileContext> for FileScanTask {
             deletes: vec![],
             sequence_number: ctx.manifest_entry.sequence_number().unwrap_or(0),
             equality_ids: ctx.manifest_entry.data_file().equality_ids().to_vec(),
+            file_size_in_bytes: ctx.manifest_entry.data_file().file_size_in_bytes(),
         }
     }
 }

crates/iceberg/src/writer/file_writer/location_generator.rs

@@ -39,7 +39,8 @@ const DEFAULT_DATA_DIR: &str = "/data";
 /// `DefaultLocationGenerator` used to generate the data dir location of data file.
 /// The location is generated based on the table location and the data location in table properties.
 pub struct DefaultLocationGenerator {
-    dir_path: String,
+    /// The data dir path of the table.
+    pub dir_path: String,
 }
 
 impl DefaultLocationGenerator {

crates/integrations/datafusion/src/lib.rs

@@ -21,7 +21,7 @@ pub use catalog::*;
 mod error;
 pub use error::*;
 
-mod physical_plan;
+pub mod physical_plan;
 mod schema;
 mod table;
 pub use table::table_provider_factory::IcebergTableProviderFactory;

crates/integrations/datafusion/src/physical_plan/mod.rs

@@ -15,5 +15,5 @@
 // specific language governing permissions and limitations
 // under the License.
 
-pub(crate) mod expr_to_predicate;
+pub mod expr_to_predicate;
 pub(crate) mod scan;