Cleanup

apache · Jan 21, 2025 · 782cd85 · 782cd85
1 parent 6c9f5c5
commit 782cd85
Show file tree

Hide file tree

Showing 3 changed files with 37 additions and 50 deletions.
diff --git a/parquet/src/arrow/async_reader/mod.rs b/parquet/src/arrow/async_reader/mod.rs
@@ -154,7 +154,7 @@ impl<T: AsyncRead + AsyncSeek + Unpin + Send> AsyncFileReader for T {
             let mut buf = Vec::with_capacity(metadata_len);
             self.take(metadata_len as _).read_to_end(&mut buf).await?;
 
-            // todo: use file_decryption_properties
+            // todo: provide file_decryption_properties
             Ok(Arc::new(ParquetMetaDataReader::decode_metadata(
                 &buf,
                 footer.encrypted_footer(),
@@ -971,6 +971,7 @@ impl RowGroups for InMemoryRowGroup<'_> {
                     // filter out empty offset indexes (old versions specified Some(vec![]) when no present)
                     .filter(|index| !index.is_empty())
                     .map(|index| index[i].page_locations.clone());
+                // todo: provide crypto_context
                 let page_reader: Box<dyn PageReader> = Box::new(SerializedPageReader::new(
                     data.clone(),
                     self.metadata.column(i),

diff --git a/parquet/src/file/metadata/mod.rs b/parquet/src/file/metadata/mod.rs
@@ -551,8 +551,6 @@ pub struct RowGroupMetaData {
     ordinal: Option<i16>,
 }
 
-// todo:rok
-
 impl RowGroupMetaData {
     /// Returns builder for row group metadata.
     pub fn builder(schema_descr: SchemaDescPtr) -> RowGroupMetaDataBuilder {
@@ -646,49 +644,43 @@ impl RowGroupMetaData {
             .zip(schema_descr.columns())
             .enumerate()
         {
-            let cc;
             #[cfg(feature = "encryption")]
-            if let Some(ColumnCryptoMetaData::ENCRYPTIONWITHCOLUMNKEY(crypto_metadata)) =
-                c.crypto_metadata.clone()
-            {
-                if c.encrypted_column_metadata.is_none() {
-                    cc = ColumnChunkMetaData::from_thrift(d.clone(), c)?;
-                } else {
-                    let decryptor = decryptor.unwrap();
-                    let column_name = crypto_metadata.path_in_schema.join(".");
-                    let column_decryptor = decryptor
-                        .get_column_decryptor(column_name.as_bytes())
-                        .footer_decryptor()
-                        .unwrap();
-
-                    let aad_file_unique = decryptor.aad_file_unique();
-                    let aad_prefix: Vec<u8> = decryptor
-                        .decryption_properties()
-                        .aad_prefix()
-                        .unwrap_or_default();
-                    let column_aad = create_module_aad(
-                        [aad_prefix.as_slice(), aad_file_unique.as_slice()]
-                            .concat()
-                            .as_slice(),
-                        ModuleType::ColumnMetaData,
-                        rg.ordinal.unwrap() as usize,
-                        i as usize,
-                        None,
-                    )?;
-
-                    let buf = c.encrypted_column_metadata.clone().unwrap();
-                    let decrypted_cc_buf =
-                        column_decryptor.decrypt(buf.as_slice().as_ref(), column_aad.as_ref())?;
-
-                    let mut prot = TCompactSliceInputProtocol::new(decrypted_cc_buf.as_slice());
-                    c.meta_data = Some(ColumnMetaData::read_from_in_protocol(&mut prot)?);
-                    cc = ColumnChunkMetaData::from_thrift(d.clone(), c)?;
-                }
-            } else {
-                cc = ColumnChunkMetaData::from_thrift(d.clone(), c)?;
+            if c.encrypted_column_metadata.is_some() {
+                let decryptor = decryptor.unwrap();
+                let Some(ColumnCryptoMetaData::ENCRYPTIONWITHCOLUMNKEY(crypto_metadata)) =
+                    c.crypto_metadata.clone()
+                else {
+                    todo!()
+                };
+                let column_name = crypto_metadata.path_in_schema.join(".");
+                let column_decryptor = decryptor
+                    .get_column_decryptor(column_name.as_bytes())
+                    .footer_decryptor()
+                    .unwrap();
+
+                let aad_file_unique = decryptor.aad_file_unique();
+                let aad_prefix: Vec<u8> = decryptor
+                    .decryption_properties()
+                    .aad_prefix()
+                    .unwrap_or_default();
+                let column_aad = create_module_aad(
+                    [aad_prefix.as_slice(), aad_file_unique.as_slice()]
+                        .concat()
+                        .as_slice(),
+                    ModuleType::ColumnMetaData,
+                    rg.ordinal.unwrap() as usize,
+                    i,
+                    None,
+                )?;
+
+                let buf = c.encrypted_column_metadata.clone().unwrap();
+                let decrypted_cc_buf =
+                    column_decryptor.decrypt(buf.as_slice(), column_aad.as_ref())?;
+
+                let mut prot = TCompactSliceInputProtocol::new(decrypted_cc_buf.as_slice());
+                c.meta_data = Some(ColumnMetaData::read_from_in_protocol(&mut prot)?);
             }
-
-            columns.push(cc);
+            columns.push(ColumnChunkMetaData::from_thrift(d.clone(), c)?);
         }
         let sorting_columns = rg.sorting_columns;
         Ok(RowGroupMetaData {

diff --git a/parquet/src/file/metadata/reader.rs b/parquet/src/file/metadata/reader.rs
@@ -736,11 +736,6 @@ impl ParquetMetaDataReader {
             }; // todo decr: add support for GCMCTRV1
 
             // todo decr: get key_metadata
-
-            // remaining buffer contains encrypted FileMetaData
-
-            // todo decr: get aad_prefix
-            // todo decr: set both aad_prefix and aad_file_unique in file_decryptor
             let aad_file_unique = aes_gcm_algo.aad_file_unique.unwrap();
             let aad_footer = create_footer_aad(aad_file_unique.as_ref())?;
             let aad_prefix: Vec<u8> = aes_gcm_algo.aad_prefix.unwrap_or_default();
@@ -762,7 +757,6 @@ impl ParquetMetaDataReader {
         let schema = types::from_thrift(&t_file_metadata.schema)?;
         let schema_descr = Arc::new(SchemaDescriptor::new(schema));
 
-        // todo add file decryptor
         #[cfg(feature = "encryption")]
         if t_file_metadata.encryption_algorithm.is_some() {
             let algo = t_file_metadata.encryption_algorithm;