|
| 1 | +// Copyright 2023 Datafuse Labs. |
| 2 | +// |
| 3 | +// Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 | +// you may not use this file except in compliance with the License. |
| 5 | +// You may obtain a copy of the License at |
| 6 | +// |
| 7 | +// http://www.apache.org/licenses/LICENSE-2.0 |
| 8 | +// |
| 9 | +// Unless required by applicable law or agreed to in writing, software |
| 10 | +// distributed under the License is distributed on an "AS IS" BASIS, |
| 11 | +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 | +// See the License for the specific language governing permissions and |
| 13 | +// limitations under the License. |
| 14 | + |
| 15 | +use std::borrow::Cow; |
| 16 | +use std::collections::VecDeque; |
| 17 | + |
| 18 | +use byteorder::BigEndian; |
| 19 | +use byteorder::ReadBytesExt; |
| 20 | + |
| 21 | +use super::constants::*; |
| 22 | +use super::error::*; |
| 23 | +use super::jentry::JEntry; |
| 24 | +use super::number::Number; |
| 25 | +use super::parser::parse_value; |
| 26 | +use super::value::Object; |
| 27 | +use super::value::Value; |
| 28 | + |
| 29 | +/// The binary `JSONB` contains three parts, `Header`, `JEntry` and `RawData`. |
| 30 | +/// This structure can be nested. Each group of structures starts with a `Header`. |
| 31 | +/// The upper-level `Value` will store the `Header` length or offset of |
| 32 | +/// the lower-level `Value`. |
| 33 | +
|
| 34 | +/// `Header` stores the type of the `Value`, include `Array`, `Object` and `Scalar`, |
| 35 | +/// `Scalar` has only one `Value`, and a corresponding `JEntry`. |
| 36 | +/// `Array` and `Object` are nested type, they have multiple lower-level `Values`. |
| 37 | +/// So the `Header` also stores the number of lower-level `Values`. |
| 38 | +
|
| 39 | +/// `JEntry` stores the types of `Scalar Value`, including `Null`, `True`, `False`, |
| 40 | +/// `Number`, `String` and `Container`. They have three different decode methods. |
| 41 | +/// 1. `Null`, `True` and `False` can be obtained by `JEntry`, no extra work required. |
| 42 | +/// 2. `Number` and `String` has related `RawData`, `JEntry` store the length |
| 43 | +/// or offset of this data, the `Value` can be read out and then decoded. |
| 44 | +/// 3. `Container` is actually a nested `Array` or `Object` with the same structure, |
| 45 | +/// `JEntry` store the length or offset of the lower-level `Header`, |
| 46 | +/// from where the same decode process can begin. |
| 47 | +
|
| 48 | +/// `RawData` is the encoded `Value`. |
| 49 | +/// `Number` is a variable-length `Decimal`, store both int and float value. |
| 50 | +/// `String` is the original string, can be borrowed directly without extra decode. |
| 51 | +/// `Array` and `Object` is a lower-level encoded `JSONB` value. |
| 52 | +/// The upper-level doesn't care about the specific content. |
| 53 | +/// Decode can be executed recursively. |
| 54 | +
|
| 55 | +/// Decode `JSONB` Value from binary bytes. |
| 56 | +pub fn from_slice(buf: &[u8]) -> Result<Value<'_>, Error> { |
| 57 | + let mut decoder = Decoder::new(buf); |
| 58 | + match decoder.decode() { |
| 59 | + Ok(value) => Ok(value), |
| 60 | + // for compatible with the first version of `JSON` text, parse it again |
| 61 | + Err(_) => parse_value(buf), |
| 62 | + } |
| 63 | +} |
| 64 | + |
| 65 | +#[repr(transparent)] |
| 66 | +pub struct Decoder<'a> { |
| 67 | + buf: &'a [u8], |
| 68 | +} |
| 69 | + |
| 70 | +impl<'a> Decoder<'a> { |
| 71 | + pub fn new(buf: &'a [u8]) -> Decoder<'a> { |
| 72 | + Self { buf } |
| 73 | + } |
| 74 | + |
| 75 | + pub fn decode(&mut self) -> Result<Value<'a>, Error> { |
| 76 | + // Valid `JSONB` Value has at least one `Header` |
| 77 | + if self.buf.len() < 4 { |
| 78 | + return Err(Error::InvalidJsonb); |
| 79 | + } |
| 80 | + let value = self.decode_jsonb()?; |
| 81 | + Ok(value) |
| 82 | + } |
| 83 | + |
| 84 | + // Read value type from the `Header` |
| 85 | + // `Scalar` has one `JEntry` |
| 86 | + // `Array` and `Object` store the numbers of elements |
| 87 | + fn decode_jsonb(&mut self) -> Result<Value<'a>, Error> { |
| 88 | + let container_header = self.buf.read_u32::<BigEndian>()?; |
| 89 | + |
| 90 | + match container_header & CONTAINER_HEADER_TYPE_MASK { |
| 91 | + SCALAR_CONTAINER_TAG => { |
| 92 | + let encoded = self.buf.read_u32::<BigEndian>()?; |
| 93 | + let jentry = JEntry::decode_jentry(encoded); |
| 94 | + self.decode_scalar(jentry) |
| 95 | + } |
| 96 | + ARRAY_CONTAINER_TAG => self.decode_array(container_header), |
| 97 | + OBJECT_CONTAINER_TAG => self.decode_object(container_header), |
| 98 | + _ => Err(Error::InvalidJsonbHeader), |
| 99 | + } |
| 100 | + } |
| 101 | + |
| 102 | + // Decode `Value` based on the `JEntry` |
| 103 | + // `Null` and `Boolean` don't need to read extra data |
| 104 | + // `Number` and `String` `JEntry` stores the length or offset of the data, |
| 105 | + // read them and decode to the `Value` |
| 106 | + // `Array` and `Object` need to read nested data from the lower-level `Header` |
| 107 | + fn decode_scalar(&mut self, jentry: JEntry) -> Result<Value<'a>, Error> { |
| 108 | + match jentry.type_code { |
| 109 | + NULL_TAG => Ok(Value::Null), |
| 110 | + TRUE_TAG => Ok(Value::Bool(true)), |
| 111 | + FALSE_TAG => Ok(Value::Bool(false)), |
| 112 | + STRING_TAG => { |
| 113 | + let offset = jentry.length as usize; |
| 114 | + let s = std::str::from_utf8(&self.buf[..offset]).unwrap(); |
| 115 | + self.buf = &self.buf[offset..]; |
| 116 | + Ok(Value::String(Cow::Borrowed(s))) |
| 117 | + } |
| 118 | + NUMBER_TAG => { |
| 119 | + let offset = jentry.length as usize; |
| 120 | + let n = Number::decode(&self.buf[..offset]); |
| 121 | + self.buf = &self.buf[offset..]; |
| 122 | + Ok(Value::Number(n)) |
| 123 | + } |
| 124 | + CONTAINER_TAG => self.decode_jsonb(), |
| 125 | + _ => Err(Error::InvalidJsonbJEntry), |
| 126 | + } |
| 127 | + } |
| 128 | + |
| 129 | + // Decode the numbers of values from the `Header`, |
| 130 | + // then read all `JEntries`, finally decode the `Value` by `JEntry` |
| 131 | + fn decode_array(&mut self, container_header: u32) -> Result<Value<'a>, Error> { |
| 132 | + let length = (container_header & CONTAINER_HEADER_LEN_MASK) as usize; |
| 133 | + let jentries = self.decode_jentries(length)?; |
| 134 | + let mut values: Vec<Value> = Vec::with_capacity(length); |
| 135 | + // decode all values |
| 136 | + for jentry in jentries.into_iter() { |
| 137 | + let value = self.decode_scalar(jentry)?; |
| 138 | + values.push(value); |
| 139 | + } |
| 140 | + |
| 141 | + let value = Value::Array(values); |
| 142 | + Ok(value) |
| 143 | + } |
| 144 | + |
| 145 | + // The basic process is the same as that of `Array` |
| 146 | + // but first decode the keys and then decode the values |
| 147 | + fn decode_object(&mut self, container_header: u32) -> Result<Value<'a>, Error> { |
| 148 | + let length = (container_header & CONTAINER_HEADER_LEN_MASK) as usize; |
| 149 | + let mut jentries = self.decode_jentries(length * 2)?; |
| 150 | + |
| 151 | + let mut keys: VecDeque<Value> = VecDeque::with_capacity(length); |
| 152 | + // decode all keys first |
| 153 | + for _ in 0..length { |
| 154 | + let jentry = jentries.pop_front().unwrap(); |
| 155 | + let key = self.decode_scalar(jentry)?; |
| 156 | + keys.push_back(key); |
| 157 | + } |
| 158 | + |
| 159 | + let mut obj = Object::new(); |
| 160 | + // decode all values |
| 161 | + for _ in 0..length { |
| 162 | + let key = keys.pop_front().unwrap(); |
| 163 | + let k = key.as_str().unwrap(); |
| 164 | + let jentry = jentries.pop_front().unwrap(); |
| 165 | + let value = self.decode_scalar(jentry)?; |
| 166 | + obj.insert(k.to_string(), value); |
| 167 | + } |
| 168 | + |
| 169 | + let value = Value::Object(obj); |
| 170 | + Ok(value) |
| 171 | + } |
| 172 | + |
| 173 | + // Decode `JEntries` for `Array` and `Object` |
| 174 | + fn decode_jentries(&mut self, length: usize) -> Result<VecDeque<JEntry>, Error> { |
| 175 | + let mut jentries: VecDeque<JEntry> = VecDeque::with_capacity(length); |
| 176 | + for _ in 0..length { |
| 177 | + let encoded = self.buf.read_u32::<BigEndian>()?; |
| 178 | + let jentry = JEntry::decode_jentry(encoded); |
| 179 | + jentries.push_back(jentry); |
| 180 | + } |
| 181 | + Ok(jentries) |
| 182 | + } |
| 183 | +} |
0 commit comments