arrow_avro/reader/
header.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18//! Decoder for [`Header`]
19
20use crate::compression::{CompressionCodec, CODEC_METADATA_KEY};
21use crate::reader::vlq::VLQDecoder;
22use crate::schema::{Schema, SCHEMA_METADATA_KEY};
23use arrow_schema::ArrowError;
24
25#[derive(Debug)]
26enum HeaderDecoderState {
27    /// Decoding the [`MAGIC`] prefix
28    Magic,
29    /// Decoding a block count
30    BlockCount,
31    /// Decoding a block byte length
32    BlockLen,
33    /// Decoding a key length
34    KeyLen,
35    /// Decoding a key string
36    Key,
37    /// Decoding a value length
38    ValueLen,
39    /// Decoding a value payload
40    Value,
41    /// Decoding sync marker
42    Sync,
43    /// Finished decoding
44    Finished,
45}
46
47/// A decoded header for an [Object Container File](https://avro.apache.org/docs/1.11.1/specification/#object-container-files)
48#[derive(Debug, Clone)]
49pub struct Header {
50    meta_offsets: Vec<usize>,
51    meta_buf: Vec<u8>,
52    sync: [u8; 16],
53}
54
55impl Header {
56    /// Returns an iterator over the meta keys in this header
57    pub fn metadata(&self) -> impl Iterator<Item = (&[u8], &[u8])> {
58        let mut last = 0;
59        self.meta_offsets.chunks_exact(2).map(move |w| {
60            let start = last;
61            last = w[1];
62            (&self.meta_buf[start..w[0]], &self.meta_buf[w[0]..w[1]])
63        })
64    }
65
66    /// Returns the value for a given metadata key if present
67    pub fn get(&self, key: impl AsRef<[u8]>) -> Option<&[u8]> {
68        self.metadata()
69            .find_map(|(k, v)| (k == key.as_ref()).then_some(v))
70    }
71
72    /// Returns the sync token for this file
73    pub fn sync(&self) -> [u8; 16] {
74        self.sync
75    }
76
77    /// Returns the [`CompressionCodec`] if any
78    pub fn compression(&self) -> Result<Option<CompressionCodec>, ArrowError> {
79        let v = self.get(CODEC_METADATA_KEY);
80
81        match v {
82            None | Some(b"null") => Ok(None),
83            Some(b"deflate") => Ok(Some(CompressionCodec::Deflate)),
84            Some(b"snappy") => Ok(Some(CompressionCodec::Snappy)),
85            Some(b"zstandard") => Ok(Some(CompressionCodec::ZStandard)),
86            Some(v) => Err(ArrowError::ParseError(format!(
87                "Unrecognized compression codec \'{}\'",
88                String::from_utf8_lossy(v)
89            ))),
90        }
91    }
92
93    /// Returns the [`Schema`] if any
94    pub fn schema(&self) -> Result<Option<Schema<'_>>, ArrowError> {
95        self.get(SCHEMA_METADATA_KEY)
96            .map(|x| {
97                serde_json::from_slice(x).map_err(|e| {
98                    ArrowError::ParseError(format!("Failed to parse Avro schema JSON: {e}"))
99                })
100            })
101            .transpose()
102    }
103}
104
105/// A decoder for [`Header`]
106///
107/// The avro file format does not encode the length of the header, and so it
108/// is necessary to provide a push-based decoder that can be used with streams
109#[derive(Debug)]
110pub struct HeaderDecoder {
111    state: HeaderDecoderState,
112    vlq_decoder: VLQDecoder,
113
114    /// The end offsets of strings in `meta_buf`
115    meta_offsets: Vec<usize>,
116    /// The raw binary data of the metadata map
117    meta_buf: Vec<u8>,
118
119    /// The decoded sync marker
120    sync_marker: [u8; 16],
121
122    /// The number of remaining tuples in the current block
123    tuples_remaining: usize,
124    /// The number of bytes remaining in the current string/bytes payload
125    bytes_remaining: usize,
126}
127
128impl Default for HeaderDecoder {
129    fn default() -> Self {
130        Self {
131            state: HeaderDecoderState::Magic,
132            meta_offsets: vec![],
133            meta_buf: vec![],
134            sync_marker: [0; 16],
135            vlq_decoder: Default::default(),
136            tuples_remaining: 0,
137            bytes_remaining: MAGIC.len(),
138        }
139    }
140}
141
142const MAGIC: &[u8; 4] = b"Obj\x01";
143
144impl HeaderDecoder {
145    /// Parse [`Header`] from `buf`, returning the number of bytes read
146    ///
147    /// This method can be called multiple times with consecutive chunks of data, allowing
148    /// integration with chunked IO systems like [`BufRead::fill_buf`]
149    ///
150    /// All errors should be considered fatal, and decoding aborted
151    ///
152    /// Once the entire [`Header`] has been decoded this method will not read any further
153    /// input bytes, and the header can be obtained with [`Self::flush`]
154    ///
155    /// [`BufRead::fill_buf`]: std::io::BufRead::fill_buf
156    pub fn decode(&mut self, mut buf: &[u8]) -> Result<usize, ArrowError> {
157        let max_read = buf.len();
158        while !buf.is_empty() {
159            match self.state {
160                HeaderDecoderState::Magic => {
161                    let remaining = &MAGIC[MAGIC.len() - self.bytes_remaining..];
162                    let to_decode = buf.len().min(remaining.len());
163                    if !buf.starts_with(&remaining[..to_decode]) {
164                        return Err(ArrowError::ParseError("Incorrect avro magic".to_string()));
165                    }
166                    self.bytes_remaining -= to_decode;
167                    buf = &buf[to_decode..];
168                    if self.bytes_remaining == 0 {
169                        self.state = HeaderDecoderState::BlockCount;
170                    }
171                }
172                HeaderDecoderState::BlockCount => {
173                    if let Some(block_count) = self.vlq_decoder.long(&mut buf) {
174                        match block_count.try_into() {
175                            Ok(0) => {
176                                self.state = HeaderDecoderState::Sync;
177                                self.bytes_remaining = 16;
178                            }
179                            Ok(remaining) => {
180                                self.tuples_remaining = remaining;
181                                self.state = HeaderDecoderState::KeyLen;
182                            }
183                            Err(_) => {
184                                self.tuples_remaining = block_count.unsigned_abs() as _;
185                                self.state = HeaderDecoderState::BlockLen;
186                            }
187                        }
188                    }
189                }
190                HeaderDecoderState::BlockLen => {
191                    if self.vlq_decoder.long(&mut buf).is_some() {
192                        self.state = HeaderDecoderState::KeyLen
193                    }
194                }
195                HeaderDecoderState::Key => {
196                    let to_read = self.bytes_remaining.min(buf.len());
197                    self.meta_buf.extend_from_slice(&buf[..to_read]);
198                    self.bytes_remaining -= to_read;
199                    buf = &buf[to_read..];
200                    if self.bytes_remaining == 0 {
201                        self.meta_offsets.push(self.meta_buf.len());
202                        self.state = HeaderDecoderState::ValueLen;
203                    }
204                }
205                HeaderDecoderState::Value => {
206                    let to_read = self.bytes_remaining.min(buf.len());
207                    self.meta_buf.extend_from_slice(&buf[..to_read]);
208                    self.bytes_remaining -= to_read;
209                    buf = &buf[to_read..];
210                    if self.bytes_remaining == 0 {
211                        self.meta_offsets.push(self.meta_buf.len());
212
213                        self.tuples_remaining -= 1;
214                        match self.tuples_remaining {
215                            0 => self.state = HeaderDecoderState::BlockCount,
216                            _ => self.state = HeaderDecoderState::KeyLen,
217                        }
218                    }
219                }
220                HeaderDecoderState::KeyLen => {
221                    if let Some(len) = self.vlq_decoder.long(&mut buf) {
222                        self.bytes_remaining = len as _;
223                        self.state = HeaderDecoderState::Key;
224                    }
225                }
226                HeaderDecoderState::ValueLen => {
227                    if let Some(len) = self.vlq_decoder.long(&mut buf) {
228                        self.bytes_remaining = len as _;
229                        self.state = HeaderDecoderState::Value;
230                    }
231                }
232                HeaderDecoderState::Sync => {
233                    let to_decode = buf.len().min(self.bytes_remaining);
234                    let write = &mut self.sync_marker[16 - to_decode..];
235                    write[..to_decode].copy_from_slice(&buf[..to_decode]);
236                    self.bytes_remaining -= to_decode;
237                    buf = &buf[to_decode..];
238                    if self.bytes_remaining == 0 {
239                        self.state = HeaderDecoderState::Finished;
240                    }
241                }
242                HeaderDecoderState::Finished => return Ok(max_read - buf.len()),
243            }
244        }
245        Ok(max_read)
246    }
247
248    /// Flush this decoder returning the parsed [`Header`] if any
249    pub fn flush(&mut self) -> Option<Header> {
250        match self.state {
251            HeaderDecoderState::Finished => {
252                self.state = HeaderDecoderState::Magic;
253                Some(Header {
254                    meta_offsets: std::mem::take(&mut self.meta_offsets),
255                    meta_buf: std::mem::take(&mut self.meta_buf),
256                    sync: self.sync_marker,
257                })
258            }
259            _ => None,
260        }
261    }
262}
263
264#[cfg(test)]
265mod test {
266    use super::*;
267    use crate::codec::{AvroDataType, AvroField};
268    use crate::reader::read_header;
269    use crate::schema::SCHEMA_METADATA_KEY;
270    use crate::test_util::arrow_test_data;
271    use arrow_schema::{DataType, Field, Fields, TimeUnit};
272    use std::fs::File;
273    use std::io::{BufRead, BufReader};
274
275    #[test]
276    fn test_header_decode() {
277        let mut decoder = HeaderDecoder::default();
278        for m in MAGIC {
279            decoder.decode(std::slice::from_ref(m)).unwrap();
280        }
281
282        let mut decoder = HeaderDecoder::default();
283        assert_eq!(decoder.decode(MAGIC).unwrap(), 4);
284
285        let mut decoder = HeaderDecoder::default();
286        decoder.decode(b"Ob").unwrap();
287        let err = decoder.decode(b"s").unwrap_err().to_string();
288        assert_eq!(err, "Parser error: Incorrect avro magic");
289    }
290
291    fn decode_file(file: &str) -> Header {
292        let file = File::open(file).unwrap();
293        read_header(BufReader::with_capacity(100, file)).unwrap()
294    }
295
296    #[test]
297    fn test_header() {
298        let header = decode_file(&arrow_test_data("avro/alltypes_plain.avro"));
299        let schema_json = header.get(SCHEMA_METADATA_KEY).unwrap();
300        let expected = br#"{"type":"record","name":"topLevelRecord","fields":[{"name":"id","type":["int","null"]},{"name":"bool_col","type":["boolean","null"]},{"name":"tinyint_col","type":["int","null"]},{"name":"smallint_col","type":["int","null"]},{"name":"int_col","type":["int","null"]},{"name":"bigint_col","type":["long","null"]},{"name":"float_col","type":["float","null"]},{"name":"double_col","type":["double","null"]},{"name":"date_string_col","type":["bytes","null"]},{"name":"string_col","type":["bytes","null"]},{"name":"timestamp_col","type":[{"type":"long","logicalType":"timestamp-micros"},"null"]}]}"#;
301        assert_eq!(schema_json, expected);
302        let schema: Schema<'_> = serde_json::from_slice(schema_json).unwrap();
303        let field = AvroField::try_from(&schema).unwrap();
304
305        assert_eq!(
306            field.field(),
307            Field::new(
308                "topLevelRecord",
309                DataType::Struct(Fields::from(vec![
310                    Field::new("id", DataType::Int32, true),
311                    Field::new("bool_col", DataType::Boolean, true),
312                    Field::new("tinyint_col", DataType::Int32, true),
313                    Field::new("smallint_col", DataType::Int32, true),
314                    Field::new("int_col", DataType::Int32, true),
315                    Field::new("bigint_col", DataType::Int64, true),
316                    Field::new("float_col", DataType::Float32, true),
317                    Field::new("double_col", DataType::Float64, true),
318                    Field::new("date_string_col", DataType::Binary, true),
319                    Field::new("string_col", DataType::Binary, true),
320                    Field::new(
321                        "timestamp_col",
322                        DataType::Timestamp(TimeUnit::Microsecond, Some("+00:00".into())),
323                        true
324                    ),
325                ])),
326                false
327            )
328        );
329
330        assert_eq!(
331            u128::from_le_bytes(header.sync()),
332            226966037233754408753420635932530907102
333        );
334
335        let header = decode_file(&arrow_test_data("avro/fixed_length_decimal.avro"));
336
337        let meta: Vec<_> = header
338            .metadata()
339            .map(|(k, _)| std::str::from_utf8(k).unwrap())
340            .collect();
341
342        assert_eq!(
343            meta,
344            &["avro.schema", "org.apache.spark.version", "avro.codec"]
345        );
346
347        let schema_json = header.get(SCHEMA_METADATA_KEY).unwrap();
348        let expected = br#"{"type":"record","name":"topLevelRecord","fields":[{"name":"value","type":[{"type":"fixed","name":"fixed","namespace":"topLevelRecord.value","size":11,"logicalType":"decimal","precision":25,"scale":2},"null"]}]}"#;
349        assert_eq!(schema_json, expected);
350        let _schema: Schema<'_> = serde_json::from_slice(schema_json).unwrap();
351        assert_eq!(
352            u128::from_le_bytes(header.sync()),
353            325166208089902833952788552656412487328
354        );
355    }
356}
arrow_avro/reader/header.rs

arrow_avro/reader/
header.rs
