[GitHub] [arrow] sunchao commented on a change in pull request #8300: ARROW-10135: [Rust] [Parquet] Refactor file module to help adding sources

[GitBox <gi...@apache.org>]

sunchao commented on a change in pull request #8300:
URL: https://github.com/apache/arrow/pull/8300#discussion_r508895342



##########
File path: rust/parquet/src/file/footer.rs
##########
@@ -0,0 +1,266 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Contains file reader API and provides methods to access file metadata, row group

Review comment:
       I think this needs to be updated?

##########
File path: rust/parquet/src/util/io.rs
##########
@@ -17,11 +17,25 @@
 
 use std::{cell::RefCell, cmp, io::*};
 
-use crate::file::{reader::ParquetReader, writer::ParquetWriter};
+use crate::file::{reader::Length, writer::ParquetWriter};
 
 const DEFAULT_BUF_SIZE: usize = 8 * 1024;
 
 // ----------------------------------------------------------------------
+
+/// TryClone tries to clone the type and should maintain the `Seek` position of the given
+/// instance.
+pub trait TryClone: Sized {
+    /// Clones the type returning a new instance or an error if it's not possible
+    /// to clone it.
+    fn try_clone(&self) -> Result<Self>;
+}
+
+/// ParquetReader is the interface which needs to be fulfilled to be able to parse a
+/// parquet source.
+pub trait ParquetReader: Read + Seek + Length + TryClone {}

Review comment:
       (not related to this PR) I think `ParquetReader` is a little confusing - perhaps a better name can be `ParquetSource` or something. And similarly `ParquetSink` for `ParquetWriter`.

##########
File path: rust/parquet/src/file/footer.rs
##########
@@ -0,0 +1,266 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Contains file reader API and provides methods to access file metadata, row group
+//! readers to read individual column chunks, or access record iterator.
+
+use std::{
+    cmp::min,
+    io::{Cursor, Read, Seek, SeekFrom},
+    rc::Rc,
+};
+
+use byteorder::{ByteOrder, LittleEndian};
+use parquet_format::{ColumnOrder as TColumnOrder, FileMetaData as TFileMetaData};
+use thrift::protocol::TCompactInputProtocol;
+
+use crate::basic::ColumnOrder;
+
+use crate::errors::{ParquetError, Result};
+use crate::file::{
+    metadata::*, reader::ChunkReader, DEFAULT_FOOTER_READ_SIZE, FOOTER_SIZE,
+    PARQUET_MAGIC,
+};
+
+use crate::schema::types::{self, SchemaDescriptor};
+
+/// Layout of Parquet file
+/// +---------------------------+-----+---+
+/// |      Rest of file         |  B  | A |
+/// +---------------------------+-----+---+
+/// where A: parquet footer, B: parquet metadata.
+///
+/// The reader first reads DEFAULT_FOOTER_SIZE bytes from the end of the file.
+/// If it is not enough according to the length indicated in the footer, it reads more bytes.
+pub fn parse_metadata<R: ChunkReader>(chunk_reader: &R) -> Result<ParquetMetaData> {

Review comment:
       +1 - nice to see this separated into another file

##########
File path: rust/parquet/src/file/serialized_reader.rs
##########
@@ -0,0 +1,747 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Contains file reader implementations and provides methods to access file metadata, row group

Review comment:
       Needs update as well - specifically for serialized reader

##########
File path: rust/parquet/src/file/reader.rs
##########
@@ -18,35 +18,37 @@
 //! Contains file reader API and provides methods to access file metadata, row group
 //! readers to read individual column chunks, or access record iterator.
 
-use std::{
-    convert::TryFrom,
-    fs::File,
-    io::{BufReader, Cursor, Read, Seek, SeekFrom},
-    path::Path,
-    rc::Rc,
-};
+use std::{boxed::Box, io::Read, rc::Rc};
 
-use byteorder::{ByteOrder, LittleEndian};
-use parquet_format::{
-    ColumnOrder as TColumnOrder, FileMetaData as TFileMetaData, PageHeader, PageType,
-};
-use thrift::protocol::TCompactInputProtocol;
-
-use crate::basic::{ColumnOrder, Compression, Encoding, Type};
 use crate::column::page::PageIterator;
-use crate::column::{
-    page::{Page, PageReader},
-    reader::{ColumnReader, ColumnReaderImpl},
-};
-use crate::compression::{create_codec, Codec};
+use crate::column::{page::PageReader, reader::ColumnReader};
 use crate::errors::{ParquetError, Result};
-use crate::file::{metadata::*, statistics, FOOTER_SIZE, PARQUET_MAGIC};
+use crate::file::metadata::*;
+pub use crate::file::serialized_reader::{SerializedFileReader, SerializedPageReader};
 use crate::record::reader::RowIter;
-use crate::record::Row;
-use crate::schema::types::{
-    self, ColumnDescPtr, SchemaDescPtr, SchemaDescriptor, Type as SchemaType,
-};
-use crate::util::{io::FileSource, memory::ByteBufferPtr};
+use crate::schema::types::{ColumnDescPtr, SchemaDescPtr, Type as SchemaType};
+
+use crate::basic::Type;
+
+use crate::column::reader::ColumnReaderImpl;
+
+/// Length should return the total number of bytes in the input source.
+/// It's mainly used to read the metadata, which is at the end of the source.
+#[allow(clippy::len_without_is_empty)]
+pub trait Length {

Review comment:
       maybe io.rs is more suitable for this trait?

##########
File path: rust/parquet/src/util/cursor.rs
##########
@@ -0,0 +1,203 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+use std::cmp;
+use std::io::{self, Error, ErrorKind, Read, Seek, SeekFrom};
+use std::rc::Rc;
+
+/// This is object to use if your file is already in memory.
+/// The sliceable cursor is similar to std::io::Cursor, except that it makes it easy to create "cursor slices".
+/// To achieve this, it uses Rc instead of shared references. Indeed reference fields are painfull

Review comment:
       Yes maybe move this into a separate PR? given the current one is already big enough.

##########
File path: rust/parquet/src/file/reader.rs
##########
@@ -18,35 +18,37 @@
 //! Contains file reader API and provides methods to access file metadata, row group
 //! readers to read individual column chunks, or access record iterator.
 
-use std::{
-    convert::TryFrom,
-    fs::File,
-    io::{BufReader, Cursor, Read, Seek, SeekFrom},
-    path::Path,
-    rc::Rc,
-};
+use std::{boxed::Box, io::Read, rc::Rc};
 
-use byteorder::{ByteOrder, LittleEndian};
-use parquet_format::{
-    ColumnOrder as TColumnOrder, FileMetaData as TFileMetaData, PageHeader, PageType,
-};
-use thrift::protocol::TCompactInputProtocol;
-
-use crate::basic::{ColumnOrder, Compression, Encoding, Type};
 use crate::column::page::PageIterator;
-use crate::column::{
-    page::{Page, PageReader},
-    reader::{ColumnReader, ColumnReaderImpl},
-};
-use crate::compression::{create_codec, Codec};
+use crate::column::{page::PageReader, reader::ColumnReader};
 use crate::errors::{ParquetError, Result};
-use crate::file::{metadata::*, statistics, FOOTER_SIZE, PARQUET_MAGIC};
+use crate::file::metadata::*;
+pub use crate::file::serialized_reader::{SerializedFileReader, SerializedPageReader};
 use crate::record::reader::RowIter;
-use crate::record::Row;
-use crate::schema::types::{
-    self, ColumnDescPtr, SchemaDescPtr, SchemaDescriptor, Type as SchemaType,
-};
-use crate::util::{io::FileSource, memory::ByteBufferPtr};
+use crate::schema::types::{ColumnDescPtr, SchemaDescPtr, Type as SchemaType};
+
+use crate::basic::Type;
+
+use crate::column::reader::ColumnReaderImpl;
+
+/// Length should return the total number of bytes in the input source.
+/// It's mainly used to read the metadata, which is at the end of the source.
+#[allow(clippy::len_without_is_empty)]
+pub trait Length {
+    /// Returns the amount of bytes of the inner source.
+    fn len(&self) -> u64;
+}
+
+/// The ChunkReader trait generates readers of chunks of a source.
+/// For a file system reader, each chunk might contain a clone of File bounded on a given range.
+/// For an object store reader, each read can be mapped to a range request.
+pub trait ChunkReader: Length {

Review comment:
       nit: maybe `ChunkRead`?




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org