ason-cpp

Header-only C++17 support for ASON, a schema-driven data format for compact structured payloads.

中文文档

Why ASON

ASON writes the schema once and stores repeated rows as tuples:

[
  {"id": 1, "name": "Alice", "active": true},
  {"id": 2, "name": "Bob", "active": false}
]

[{id:int,name:str,active:bool}]:(1,Alice,true),(2,Bob,false)

That cuts repeated keys, reduces payload size, and keeps typed structure visible.

Highlights

• Header-only, just #include "ason.hpp"
• Current API uses encode / decode, not the older dump / load names
• Text and binary formats
• SIMD-aware parser and zero-copy-friendly decoding
• Support for std::optional, std::vector, std::unordered_map, and nested structs

Quick Start

#include "ason.hpp"

struct User {
  int64_t id = 0;
  std::string name;
  bool active = false;
};

ASON_FIELDS(User,
    (id, "id", "int"),
    (name, "name", "str"),
    (active, "active", "bool"))

Encode and decode a struct

User user{1, "Alice", true};

std::string text = ason::encode(user);
// {id,name,active}:(1,Alice,true)

std::string typed = ason::encode_typed(user);
// {id:int,name:str,active:bool}:(1,Alice,true)

User decoded = ason::decode<User>(text);

Encode and decode a vector

std::vector<User> users = {
    {1, "Alice", true},
    {2, "Bob", false},
};

auto text = ason::encode(users);
// [{id,name,active}]:(1,Alice,true),(2,Bob,false)

auto typed = ason::encode_typed(users);
auto decoded = ason::decode<std::vector<User>>(text);

Binary roundtrip

std::string bin = ason::encode_bin(user);
User decoded = ason::decode_bin<User>(bin);

Current API

Function	Purpose
ason::encode / ason::encode_typed	Encode to text
ason::decode<T>	Decode from text
ason::encode_pretty / ason::encode_pretty_typed	Pretty text output
ason::encode_bin	Encode to binary
ason::decode_bin<T>	Decode from binary

Build and Run

cmake -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build
./build/basic
./build/complex_example
./build/bench
ctest --test-dir build

Latest Benchmarks

Measured on this machine with:

./build/bench

Headline numbers:

• Flat 1,000-record dataset: ASON text serialize 67.79ms vs JSON 166.63ms, deserialize 133.95ms vs JSON 245.80ms
• Throughput summary: ASON text was 2.69x faster than JSON for serialize and 1.77x faster for deserialize
• Size summary for 1,000 flat records: JSON 121,675 B, ASON text 56,718 B (53% smaller), ASON binary 74,454 B (39% smaller)
• Binary decode was especially strong: 29.18ms vs JSON 245.80ms on flat 1,000-record data, or 8.42x faster

For 100 deep company objects, ASON text serialized 1.87e5 B vs JSON 4.32e5 B and decoded 2.38x faster.

Contributors

• Athan

License

MIT