rsonpath – SIMD-powered JSONPath 🚀

Experimental JSONPath engine for querying massive streamed datasets.

Features

The rsonpath crate provides a JSONPath parser and a query execution engine, which utilizes SIMD instructions to provide massive throughput improvements over conventional engines.

Supported selectors

The project is actively developed and currently supports only a subset of the JSONPath query language.

Installation

Easiest way to install is via cargo.

cargo install rsonpath

This might fail with the following error:

Target architecture is not supported by SIMD features of this crate. Disable the default `simd` feature.

This means the SIMD features of the engine are not implemented for the machine's CPU. You can still use rsonpath, but the speed will be much more limited. To install SIMD-less rsonpath, run:

cargo install rsonpath --no-default-features

Alternatively, you can download the source code and manually run just install (requires just) or cargo install --path ./crates/rsonpath.

Usage

To run a JSONPath query on a file execute:

rsonpath '$..a.b' ./file.json

If the file is omitted, the engine reads standard input.

For details, consult rsonpath --help.

Results

The results are presented as an array of indices at which a colon of a matching record was found. Alternatively, passing --result count returns only the number of matches.

Engine

By default, the main SIMD engine is used. On machines not supporting SIMD, the recursive implementation might be faster in some cases. To change the engine use --engine recursive.

Build & test

The dev workflow utilizes just. Use the included Justfile. It will automatically install Rust for you using the rustup tool if it detects there is no Cargo in your environment.

just build
just test

Benchmarks

Benchmarks for rsonpath are located in a separate repository, included as a git submodule in this main repository.

Easiest way to run all the benchmarks is just bench. For details, look at the README in the submodule.

Background

This project is the result of my thesis. You can read it for details on the theoretical background on the engine and details of its implementation.

Dependencies

Showing direct dependencies, for full graph see below.

cargo tree --package rsonpath --edges normal --depth 1

rsonpath v0.2.1 (/home/mgienieczko/rsonpath/crates/rsonpath)
├── clap v4.0.25
├── color-eyre v0.6.2
├── eyre v0.6.8
├── log v0.4.17
├── rsonpath-lib v0.2.1 (/home/mgienieczko/rsonpath/crates/rsonpath-lib)
└── simple_logger v4.0.0

cargo tree --package rsonpath-lib --edges normal --depth 1

rsonpath-lib v0.2.1 (/home/mgienieczko/rsonpath/crates/rsonpath-lib)
├── aligners v0.0.10
├── cfg-if v1.0.0
├── log v0.4.17
├── memchr v2.5.0
├── nom v7.1.3
├── replace_with v0.1.7
├── smallvec v1.10.0
├── thiserror v1.0.38
└── vector-map v1.0.1

Justification

• clap – standard crate to provide the CLI.

• color-eyre, eyre – more accessible error messages for the parser.

• log, simple-logger – diagnostic logs during compilation and execution.

• aligners – SIMD operations require correct input data alignment, putting those requirements at type level makes our code more robust.

• cfg-if – used to support SIMD and no-SIMD versions.

• memchr – rapid, SIMDified substring search for fast-forwarding to labels.

• nom – for parser implementation.

• replace_with – for safe handling of internal classifier state when switching classifiers.

• smallvec – crucial for small-stack performance.

• thiserror – idiomatic Error implementations.

• vector_map – used in the query compiler for measurably better performance.

Full dependency tree

cargo tree --package rsonpath --edges normal

rsonpath v0.2.1 (/home/mgienieczko/rsonpath/crates/rsonpath)
├── clap v4.0.25
│   ├── atty v0.2.14
│   │   └── libc v0.2.139
│   ├── bitflags v1.3.2
│   ├── clap_derive v4.0.21 (proc-macro)
│   │   ├── heck v0.4.0
│   │   ├── proc-macro-error v1.0.4
│   │   │   ├── proc-macro-error-attr v1.0.4 (proc-macro)
│   │   │   │   ├── proc-macro2 v1.0.47
│   │   │   │   │   └── unicode-ident v1.0.0
│   │   │   │   └── quote v1.0.18
│   │   │   │       └── proc-macro2 v1.0.47 (*)
│   │   │   ├── proc-macro2 v1.0.47 (*)
│   │   │   ├── quote v1.0.18 (*)
│   │   │   └── syn v1.0.75
│   │   │       ├── proc-macro2 v1.0.47 (*)
│   │   │       ├── quote v1.0.18 (*)
│   │   │       └── unicode-xid v0.2.2
│   │   ├── proc-macro2 v1.0.47 (*)
│   │   ├── quote v1.0.18 (*)
│   │   └── syn v1.0.75 (*)
│   ├── clap_lex v0.3.0
│   │   └── os_str_bytes v6.1.0
│   ├── once_cell v1.16.0
│   ├── strsim v0.10.0
│   ├── termcolor v1.1.3
│   └── terminal_size v0.2.2
│       └── rustix v0.35.13
│           ├── bitflags v1.3.2
│           ├── io-lifetimes v0.7.5
│           ├── libc v0.2.139
│           └── linux-raw-sys v0.0.46
├── color-eyre v0.6.2
│   ├── backtrace v0.3.65
│   │   ├── addr2line v0.17.0
│   │   │   └── gimli v0.26.1
│   │   ├── cfg-if v1.0.0
│   │   ├── libc v0.2.139
│   │   ├── miniz_oxide v0.5.1
│   │   │   └── adler v1.0.2
│   │   ├── object v0.28.3
│   │   │   └── memchr v2.5.0
│   │   └── rustc-demangle v0.1.21
│   ├── eyre v0.6.8
│   │   ├── indenter v0.3.3
│   │   └── once_cell v1.16.0
│   ├── indenter v0.3.3
│   ├── once_cell v1.16.0
│   └── owo-colors v3.3.0
├── eyre v0.6.8 (*)
├── log v0.4.17
│   └── cfg-if v1.0.0
├── rsonpath-lib v0.2.1 (/home/mgienieczko/rsonpath/crates/rsonpath-lib)
│   ├── aligners v0.0.10
│   │   ├── cfg-if v1.0.0
│   │   ├── lazy_static v1.4.0
│   │   └── page_size v0.4.2
│   │       └── libc v0.2.139
│   ├── cfg-if v1.0.0
│   ├── log v0.4.17 (*)
│   ├── memchr v2.5.0
│   ├── nom v7.1.3
│   │   ├── memchr v2.5.0
│   │   └── minimal-lexical v0.2.1
│   ├── replace_with v0.1.7
│   ├── smallvec v1.10.0
│   ├── thiserror v1.0.38
│   │   └── thiserror-impl v1.0.38 (proc-macro)
│   │       ├── proc-macro2 v1.0.47 (*)
│   │       ├── quote v1.0.18 (*)
│   │       └── syn v1.0.75 (*)
│   └── vector-map v1.0.1
│       ├── contracts v0.4.0 (proc-macro)
│       │   ├── proc-macro2 v1.0.47 (*)
│       │   ├── quote v1.0.18 (*)
│       │   └── syn v1.0.75 (*)
│       └── rand v0.7.3
│           ├── getrandom v0.1.16
│           │   ├── cfg-if v1.0.0
│           │   └── libc v0.2.139
│           ├── libc v0.2.139
│           ├── rand_chacha v0.2.2
│           │   ├── ppv-lite86 v0.2.16
│           │   └── rand_core v0.5.1
│           │       └── getrandom v0.1.16 (*)
│           └── rand_core v0.5.1 (*)
└── simple_logger v4.0.0
    ├── colored v2.0.0
    │   ├── atty v0.2.14 (*)
    │   └── lazy_static v1.4.0
    ├── log v0.4.17 (*)
    └── time v0.3.17
        ├── itoa v1.0.2
        ├── libc v0.2.139
        ├── num_threads v0.1.6
        ├── time-core v0.1.0
        └── time-macros v0.2.6 (proc-macro)
            └── time-core v0.1.0