Compare commits

...

3 Commits

Author SHA1 Message Date
Greg Shuflin
1e91bf4efd Add benchmarking code for json parser 2023-03-10 01:16:06 -08:00
Greg Shuflin
b86cbfd8ff Move json into testutil module 2023-03-10 01:12:46 -08:00
Greg Shuflin
d132d8dbdd Add criterion crate + basic code 2023-03-10 00:45:04 -08:00
5 changed files with 169 additions and 136 deletions

View File

@ -5,9 +5,21 @@ edition = "2021"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[features]
testutil = []
[dependencies] [dependencies]
arbitrary = "1.2.0" arbitrary = "1.2.0"
proptest = "1.0.0" proptest = "1.0.0"
[dev-dependencies] [dev-dependencies]
criterion = "0.4.0"
rstest = "0.16.0" rstest = "0.16.0"
# see https://github.com/rust-lang/cargo/issues/2911#issuecomment-749580481
parser-combinator = { path = ".", features = ["testutil"] }
[[bench]]
name = "json-benchmark"
harness = false

14
benches/json-benchmark.rs Normal file
View File

@ -0,0 +1,14 @@
use criterion::{black_box, criterion_group, criterion_main, Criterion};
use parser_combinator::testutil::json_object;
use parser_combinator::Parser;
pub fn criterion_benchmark(c: &mut Criterion) {
let test_json = include_str!("../tests/joplin-cfg.json");
c.bench_function("parse_json", |b| {
b.iter(|| json_object().parse(black_box(test_json)))
});
}
criterion_group!(benches, criterion_benchmark);
criterion_main!(benches);

View File

@ -4,4 +4,7 @@ mod parser;
pub mod primitives; pub mod primitives;
pub mod sequence; pub mod sequence;
#[cfg(feature = "testutil")]
pub mod testutil;
pub use parser::{ParseResult, Parser, ParserInput, Representation}; pub use parser::{ParseResult, Parser, ParserInput, Representation};

135
src/testutil/mod.rs Normal file
View File

@ -0,0 +1,135 @@
use crate::choice::choice;
use crate::combinators::repeated;
use crate::primitives::{any_char, literal, literal_char, one_of, pred};
use crate::sequence::seq;
use crate::Parser;
/*
* JSON BNF
* <JSON> ::= <value>
<value> ::= <object> | <array> | <boolean> | <string> | <number> | <null>
<array> ::= "[" [<value>] {"," <value>}* "]"
<object> ::= "{" [<property>] {"," <property>}* "}"
<property> ::= <string> ":" <value>
*/
#[derive(Debug, Clone, PartialEq)]
pub enum JsonValue {
Null,
Bool(bool),
Str(String),
Num(f64),
Array(Vec<JsonValue>),
Object(Vec<(String, JsonValue)>),
}
pub trait JsonParser<'a, T>: Parser<&'a str, T, &'a str> {}
impl<'a, T, P> JsonParser<'a, T> for P where P: Parser<&'a str, T, &'a str> {}
pub fn json_null<'a>() -> impl JsonParser<'a, JsonValue> {
literal("null").to(JsonValue::Null)
}
pub fn json_bool<'a>() -> impl JsonParser<'a, JsonValue> {
choice((
literal("true").to(JsonValue::Bool(true)),
literal("false").to(JsonValue::Bool(false)),
))
}
pub fn json_number<'a>() -> impl JsonParser<'a, JsonValue> {
fn digit<'a>() -> impl JsonParser<'a, &'a str> {
one_of("1234567890")
}
fn digits<'a>() -> impl JsonParser<'a, Vec<&'a str>> {
repeated(digit()).at_least(1)
}
let json_number_inner = choice((
seq((digits(), literal(".").ignore_then(digits()).optional())).map(
|(mut digits, maybe_decimal)| {
if let Some(decimal_digits) = maybe_decimal {
digits.push(".");
digits.extend(decimal_digits.into_iter());
}
digits.into_iter().collect::<String>()
},
),
literal(".").ignore_then(digits()).map(|decimal_digits| {
let mut d = vec!["."];
d.extend(decimal_digits.into_iter());
d.into_iter().collect::<String>()
}),
))
.map(|digits| digits.parse::<f64>().unwrap());
literal("-")
.optional()
.then(json_number_inner)
.map(|(maybe_sign, mut val)| {
if maybe_sign.is_some() {
val *= -1.0;
}
JsonValue::Num(val)
})
}
pub fn json_string_raw<'a>() -> impl JsonParser<'a, String> {
seq((
literal_char('"'),
repeated(pred(any_char, |ch| *ch != '"')),
literal_char('"'),
))
.map(|(_, s, _)| s.iter().cloned().collect::<String>())
}
pub fn json_string<'a>() -> impl JsonParser<'a, JsonValue> {
json_string_raw().map(JsonValue::Str)
}
fn whitespace<'a>() -> impl JsonParser<'a, ()> {
repeated(choice((
literal_char('\t'),
literal_char('\n'),
literal_char(' '),
)))
.to(())
}
pub fn json_array<'a>() -> impl JsonParser<'a, JsonValue> {
move |input| {
let val = json_value().surrounded_by(whitespace());
repeated(val)
.separated_by(literal(","), false)
.delimited(literal_char('['), literal_char(']'))
.map(JsonValue::Array)
.parse(input)
}
}
pub fn json_object<'a>() -> impl JsonParser<'a, JsonValue> {
move |input| {
let kv = json_string_raw()
.surrounded_by(whitespace())
.then_ignore(literal_char(':'))
.then(json_value().surrounded_by(whitespace()));
repeated(kv)
.separated_by(literal_char(','), false)
.delimited(literal_char('{'), literal_char('}'))
.map(JsonValue::Object)
.parse(input)
}
}
pub fn json_value<'a>() -> impl JsonParser<'a, JsonValue> {
choice((
json_null(),
json_bool(),
json_number(),
json_string(),
json_array(),
json_object(),
))
}

View File

@ -1,9 +1,8 @@
use parser_combinator::choice::choice; use parser_combinator::primitives::literal;
use parser_combinator::combinators::repeated; use parser_combinator::testutil::{
use parser_combinator::primitives::{any_char, literal, literal_char, one_of, pred}; json_array, json_bool, json_null, json_number, json_object, json_string, JsonValue,
use parser_combinator::sequence::seq; };
use parser_combinator::Parser; use parser_combinator::{Parser, Representation};
use parser_combinator::Representation;
use proptest::prelude::*; use proptest::prelude::*;
@ -32,136 +31,6 @@ fn test_parsing() {
assert_eq!(output.unwrap(), ("a", " yolo")); assert_eq!(output.unwrap(), ("a", " yolo"));
} }
/*
* JSON BNF
* <JSON> ::= <value>
<value> ::= <object> | <array> | <boolean> | <string> | <number> | <null>
<array> ::= "[" [<value>] {"," <value>}* "]"
<object> ::= "{" [<property>] {"," <property>}* "}"
<property> ::= <string> ":" <value>
*/
#[derive(Debug, Clone, PartialEq)]
enum JsonValue {
Null,
Bool(bool),
Str(String),
Num(f64),
Array(Vec<JsonValue>),
Object(Vec<(String, JsonValue)>),
}
trait JsonParser<'a, T>: Parser<&'a str, T, &'a str> {}
impl<'a, T, P> JsonParser<'a, T> for P where P: Parser<&'a str, T, &'a str> {}
fn json_null<'a>() -> impl JsonParser<'a, JsonValue> {
literal("null").to(JsonValue::Null)
}
fn json_bool<'a>() -> impl JsonParser<'a, JsonValue> {
choice((
literal("true").to(JsonValue::Bool(true)),
literal("false").to(JsonValue::Bool(false)),
))
}
fn json_number<'a>() -> impl JsonParser<'a, JsonValue> {
fn digit<'a>() -> impl JsonParser<'a, &'a str> {
one_of("1234567890")
}
fn digits<'a>() -> impl JsonParser<'a, Vec<&'a str>> {
repeated(digit()).at_least(1)
}
let json_number_inner = choice((
seq((digits(), literal(".").ignore_then(digits()).optional())).map(
|(mut digits, maybe_decimal)| {
if let Some(decimal_digits) = maybe_decimal {
digits.push(".");
digits.extend(decimal_digits.into_iter());
}
digits.into_iter().collect::<String>()
},
),
literal(".").ignore_then(digits()).map(|decimal_digits| {
let mut d = vec!["."];
d.extend(decimal_digits.into_iter());
d.into_iter().collect::<String>()
}),
))
.map(|digits| digits.parse::<f64>().unwrap());
literal("-")
.optional()
.then(json_number_inner)
.map(|(maybe_sign, mut val)| {
if maybe_sign.is_some() {
val *= -1.0;
}
JsonValue::Num(val)
})
}
fn json_string_raw<'a>() -> impl JsonParser<'a, String> {
seq((
literal_char('"'),
repeated(pred(any_char, |ch| *ch != '"')),
literal_char('"'),
))
.map(|(_, s, _)| s.iter().cloned().collect::<String>())
}
fn json_string<'a>() -> impl JsonParser<'a, JsonValue> {
json_string_raw().map(JsonValue::Str)
}
fn whitespace<'a>() -> impl JsonParser<'a, ()> {
repeated(choice((
literal_char('\t'),
literal_char('\n'),
literal_char(' '),
)))
.to(())
}
fn json_array<'a>() -> impl JsonParser<'a, JsonValue> {
move |input| {
let val = json_value().surrounded_by(whitespace());
repeated(val)
.separated_by(literal(","), false)
.delimited(literal_char('['), literal_char(']'))
.map(JsonValue::Array)
.parse(input)
}
}
fn json_object<'a>() -> impl JsonParser<'a, JsonValue> {
move |input| {
let kv = json_string_raw()
.surrounded_by(whitespace())
.then_ignore(literal_char(':'))
.then(json_value().surrounded_by(whitespace()));
repeated(kv)
.separated_by(literal_char(','), false)
.delimited(literal_char('{'), literal_char('}'))
.map(JsonValue::Object)
.parse(input)
}
}
fn json_value<'a>() -> impl JsonParser<'a, JsonValue> {
choice((
json_null(),
json_bool(),
json_number(),
json_string(),
json_array(),
json_object(),
))
}
#[test] #[test]
fn parse_json_primitives() { fn parse_json_primitives() {
assert_eq!( assert_eq!(