Add a data storage optimizer

crates/modelardb_embedded/bindings/c/modelardb_embedded.h

@@ -199,7 +199,7 @@ int modelardb_embedded_vacuum(void* maybe_operations_ptr,
                               const char* table_name_ptr,
                               const uint64_t* retention_period_in_seconds_ptr);
 
-// Optimize the table by compacting its many small files into fewer larger files.
+// Optimize the table by merging its many small files into fewer larger files.
 int modelardb_embedded_optimize(void* maybe_operations_ptr,
                                 bool is_data_folder,
                                 const char* table_name_ptr,

crates/modelardb_embedded/bindings/python/modelardb/operations.py

@@ -685,7 +685,7 @@ def optimize(self, table_name: str, target_size_in_bytes: None | int = None):
 
         :param table_name: The name of the table to optimize.
         :type table_name: str
-        :param target_size_in_bytes: The target file size in bytes. Many small files are compacted
+        :param target_size_in_bytes: The target file size in bytes. Many small files are merged
          into fewer larger files of approximately this size. If `None`, the default target size of
          64 MiB is used.
         :type target_size_in_bytes: int, optional

crates/modelardb_embedded/bindings/python/tests/test_operations.py

@@ -512,10 +512,10 @@ def test_data_folder_optimize(self):
 
             data_folder.optimize(TIME_SERIES_TABLE_NAME)
 
-            # Vacuum to remove the compacted files.
+            # Vacuum to remove the stale files left by the merge.
             data_folder.vacuum(TIME_SERIES_TABLE_NAME, retention_period_in_seconds=0)
 
-            # The small files should be compacted into a single active file.
+            # The small files should be merged into a single active file.
             file_count = len(os.listdir(folder_path))
             self.assertEqual(file_count, 1)
 

crates/modelardb_embedded/src/capi.rs

@@ -1007,7 +1007,7 @@ unsafe fn vacuum(
 }
 
 /// Optimizes the table with the name in `table_name_ptr` in the [`DataFolder`] or [`Client`] in
-/// `maybe_operations_ptr` by compacting its many small files into fewer larger files of
+/// `maybe_operations_ptr` by merging its many small files into fewer larger files of
 /// approximately `target_size_in_bytes_ptr` bytes. Assumes `maybe_operations_ptr` points to a
 /// [`DataFolder`] or [`Client`]; `table_name_ptr` points to a valid C string; and
 /// `target_size_in_bytes_ptr` points to a valid `u64`, or is null to use the default target size.

crates/modelardb_embedded/src/operations/client.rs

@@ -316,7 +316,7 @@ impl Operations for Client {
         Ok(())
     }
 
-    /// Optimize the table with the name in `table_name` by compacting its many small files into
+    /// Optimize the table with the name in `table_name` by merging its many small files into
     /// fewer larger files of approximately `maybe_target_size_in_bytes` bytes. If a target size is
     /// not given, the default target size of 64 MiB is used. If the table does not exist, the table
     /// could not be optimized, or the target size is zero, [`ModelarDbEmbeddedError`] is returned.

crates/modelardb_embedded/src/operations/data_folder.rs

@@ -540,7 +540,7 @@ impl Operations for DataFolder {
         }
     }
 
-    /// Optimize the table with the name in `table_name` by compacting its many small files into
+    /// Optimize the table with the name in `table_name` by merging its many small files into
     /// fewer larger files of approximately `maybe_target_size_in_bytes` bytes. If a target size is
     /// not given, the default target size of 64 MiB is used. If the table does not exist, the table
     /// could not be optimized, or the target size is zero, [`ModelarDbEmbeddedError`] is returned.
@@ -2314,7 +2314,7 @@ mod tests {
 
         data_folder.optimize(NORMAL_TABLE_NAME, None).await.unwrap();
 
-        // The small files should be compacted into a single active file.
+        // The small files should be merged into a single active file.
         let delta_table = data_folder.delta_table(NORMAL_TABLE_NAME).await.unwrap();
         assert_eq!(delta_table.get_file_uris().unwrap().count(), 1);
     }
@@ -2343,7 +2343,7 @@ mod tests {
             .await
             .unwrap();
 
-        // The files in each of the two partitions should be compacted into a single active file.
+        // The files in each of the two partitions should be merged into a single active file.
         let delta_table = data_folder
             .delta_table(TIME_SERIES_TABLE_NAME)
             .await

crates/modelardb_embedded/src/operations/mod.rs

@@ -146,7 +146,7 @@ pub trait Operations: Sync + Send {
         maybe_retention_period_in_seconds: Option<u64>,
     ) -> Result<()>;
 
-    /// Optimize the table with the name in `table_name` by compacting its many small files into
+    /// Optimize the table with the name in `table_name` by merging its many small files into
     /// fewer larger files of approximately `maybe_target_size_in_bytes` bytes. If a target size is
     /// not given, the default target size of 64 MiB is used.
     async fn optimize(