metrify/src/parse.rs

307 lines
8.7 KiB
Rust
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

use crate::units::{NonMetric, NonMetricQuantity};
enum Expect {
Number,
Unit,
}
#[derive(Debug, PartialEq)]
pub enum ParseError {
NotValidNumber(String),
UnexpectedUnit(String),
UnknownUnit(String),
ExpectedUnit,
}
pub fn parse(input: &str) -> Result<Vec<NonMetricQuantity>, ParseError> {
let mut quantities = Vec::new();
let mut state = Expect::Number;
let mut amount = None;
for token in tokenize(input) {
match (&state, token) {
(Expect::Number, Token::Number(number)) => {
let number = parse_number(number)?;
amount = Some(number);
state = Expect::Unit;
}
(Expect::Number, Token::Unit(unit)) => {
return Err(ParseError::UnexpectedUnit(unit));
}
(Expect::Unit, Token::Number(_)) => {
unreachable!("token stream can't contain two numbers in a row");
}
(Expect::Unit, Token::Unit(unit)) => {
let unit = parse_unit(unit)?;
let quantity = NonMetricQuantity {
amount: amount.take().expect("must have read a number to be in this state"),
unit: unit,
};
quantities.push(quantity);
state = Expect::Number;
}
}
}
match state {
Expect::Number => {}
Expect::Unit => {
return Err(ParseError::ExpectedUnit);
}
}
Ok(quantities)
}
fn parse_number(input: String) -> Result<f64, ParseError> {
let no_whitespace: String = input.chars().filter(|c| !c.is_whitespace()).collect();
no_whitespace.parse().or_else(|_| Err(ParseError::NotValidNumber(input)))
}
fn parse_unit(input: String) -> Result<NonMetric, ParseError> {
match input.as_str() {
// Length
"inch" => Ok(NonMetric::Inch),
"inches" => Ok(NonMetric::Inch),
"in" => Ok(NonMetric::Inch),
"\"" => Ok(NonMetric::Inch),
"" => Ok(NonMetric::Inch),
"foot" => Ok(NonMetric::Foot),
"feet" => Ok(NonMetric::Foot),
"ft" => Ok(NonMetric::Foot),
"'" => Ok(NonMetric::Foot),
"" => Ok(NonMetric::Foot),
"yard" => Ok(NonMetric::Yard),
"yards" => Ok(NonMetric::Yard),
"yd" => Ok(NonMetric::Yard),
"mile" => Ok(NonMetric::Mile),
"miles" => Ok(NonMetric::Mile),
"mi" => Ok(NonMetric::Mile),
"m" => Ok(NonMetric::Mile),
// Weight
"ounce" => Ok(NonMetric::Ounce),
"ounces" => Ok(NonMetric::Ounce),
"oz" => Ok(NonMetric::Ounce),
"pound" => Ok(NonMetric::Pound),
"pounds" => Ok(NonMetric::Pound),
"lb" => Ok(NonMetric::Pound),
"lbs" => Ok(NonMetric::Pound),
"#" => Ok(NonMetric::Pound),
"stone" => Ok(NonMetric::Stone),
"stones" => Ok(NonMetric::Stone),
"st" => Ok(NonMetric::Stone),
// Temperature
"°F" => Ok(NonMetric::Fahrenheit),
"F" => Ok(NonMetric::Fahrenheit),
_ => Err(ParseError::UnknownUnit(input)),
}
}
#[derive(Debug, PartialEq)]
enum Token {
Number(String),
Unit(String),
}
enum TokState {
Neutral,
Number,
Unit,
}
fn tokenize(input: &str) -> Vec<Token> {
let mut tokens = Vec::new();
let mut token = String::new();
let mut state = TokState::Neutral;
for c in input.chars() {
match state {
TokState::Neutral => {
if c.is_ascii_digit() || c == '-' {
token.push(c);
state = TokState::Number;
} else if !c.is_whitespace() {
token.push(c);
state = TokState::Unit;
}
}
TokState::Number => {
if c.is_ascii_digit() ||
c.is_whitespace() ||
c == '.' {
token.push(c);
} else {
tokens.push(Token::Number(token.trim().to_string()));
state = TokState::Unit;
token = String::new();
token.push(c);
}
}
TokState::Unit => {
if c.is_ascii_digit() || c == '-' {
tokens.push(Token::Unit(token));
state = TokState::Number;
token = String::new();
token.push(c);
} else if !c.is_whitespace() {
token.push(c);
} else {
tokens.push(Token::Unit(token));
state = TokState::Neutral;
token = String::new();
}
}
}
}
match state {
TokState::Neutral => { assert!(token.len() == 0); }
TokState::Number => { tokens.push(Token::Number(token.trim().to_string())); }
TokState::Unit => { tokens.push(Token::Unit(token)); }
}
tokens
}
#[cfg(test)]
mod test {
use super::*;
#[test]
fn parsing() {
assert_eq!(parse(""), Ok(vec![]));
assert_eq!(parse("5 ft"), Ok(vec![
NonMetricQuantity { amount: 5.0, unit: NonMetric::Foot },
]));
assert_eq!(parse("5 ft 8 in"), Ok(vec![
NonMetricQuantity { amount: 5.0, unit: NonMetric::Foot },
NonMetricQuantity { amount: 8.0, unit: NonMetric::Inch },
]));
assert_eq!(parse("20 000 lbs"), Ok(vec![
NonMetricQuantity { amount: 20_000.0, unit: NonMetric::Pound },
]));
assert_eq!(parse("12.0."), Err(ParseError::NotValidNumber("12.0.".to_string())));
assert_eq!(parse("ft"), Err(ParseError::UnexpectedUnit("ft".to_string())));
assert_eq!(parse("5 tf"), Err(ParseError::UnknownUnit("tf".to_string())));
assert_eq!(parse("12"), Err(ParseError::ExpectedUnit));
}
#[test]
fn numbers() {
assert_eq!(parse_number("".to_string()), Err(ParseError::NotValidNumber("".to_string())));
assert_eq!(parse_number("1".to_string()), Ok(1.0));
assert_eq!(parse_number("1.0".to_string()), Ok(1.0));
assert_eq!(parse_number("0.1".to_string()), Ok(0.1));
assert_eq!(parse_number("0.1.".to_string()), Err(ParseError::NotValidNumber("0.1.".to_string())));
assert_eq!(parse_number("-10".to_string()), Ok(-10.0));
assert_eq!(parse_number("10\t00\u{1680}000".to_string()), Ok(10_00_000.0));
}
#[test]
fn units() {
// Length
assert_eq!(parse_unit("inch".to_string()), Ok(NonMetric::Inch));
assert_eq!(parse_unit("inches".to_string()), Ok(NonMetric::Inch));
assert_eq!(parse_unit("in".to_string()), Ok(NonMetric::Inch));
assert_eq!(parse_unit("\"".to_string()), Ok(NonMetric::Inch));
assert_eq!(parse_unit("".to_string()), Ok(NonMetric::Inch));
assert_eq!(parse_unit("foot".to_string()), Ok(NonMetric::Foot));
assert_eq!(parse_unit("feet".to_string()), Ok(NonMetric::Foot));
assert_eq!(parse_unit("ft".to_string()), Ok(NonMetric::Foot));
assert_eq!(parse_unit("'".to_string()), Ok(NonMetric::Foot));
assert_eq!(parse_unit("".to_string()), Ok(NonMetric::Foot));
assert_eq!(parse_unit("yard".to_string()), Ok(NonMetric::Yard));
assert_eq!(parse_unit("yards".to_string()), Ok(NonMetric::Yard));
assert_eq!(parse_unit("yd".to_string()), Ok(NonMetric::Yard));
assert_eq!(parse_unit("mile".to_string()), Ok(NonMetric::Mile));
assert_eq!(parse_unit("miles".to_string()), Ok(NonMetric::Mile));
assert_eq!(parse_unit("mi".to_string()), Ok(NonMetric::Mile));
assert_eq!(parse_unit("m".to_string()), Ok(NonMetric::Mile));
// Weight
assert_eq!(parse_unit("ounce".to_string()), Ok(NonMetric::Ounce));
assert_eq!(parse_unit("ounces".to_string()), Ok(NonMetric::Ounce));
assert_eq!(parse_unit("oz".to_string()), Ok(NonMetric::Ounce));
assert_eq!(parse_unit("pound".to_string()), Ok(NonMetric::Pound));
assert_eq!(parse_unit("pounds".to_string()), Ok(NonMetric::Pound));
assert_eq!(parse_unit("lb".to_string()), Ok(NonMetric::Pound));
assert_eq!(parse_unit("lbs".to_string()), Ok(NonMetric::Pound));
assert_eq!(parse_unit("#".to_string()), Ok(NonMetric::Pound));
assert_eq!(parse_unit("stone".to_string()), Ok(NonMetric::Stone));
assert_eq!(parse_unit("stones".to_string()), Ok(NonMetric::Stone));
assert_eq!(parse_unit("st".to_string()), Ok(NonMetric::Stone));
// Temperature
assert_eq!(parse_unit("°F".to_string()), Ok(NonMetric::Fahrenheit));
assert_eq!(parse_unit("F".to_string()), Ok(NonMetric::Fahrenheit));
// Unknown unit
assert_eq!(parse_unit("hutenosa".to_string()), Err(ParseError::UnknownUnit("hutenosa".to_string())));
}
#[test]
fn tokens() {
assert_eq!(tokenize(""), vec![]);
assert_eq!(tokenize("10"), vec![Token::Number("10".to_string())]);
assert_eq!(tokenize(" 10 "), vec![Token::Number("10".to_string())]);
assert_eq!(tokenize("10 000"), vec![Token::Number("10 000".to_string())]);
assert_eq!(tokenize("10\t000"), vec![Token::Number("10\t000".to_string())]);
assert_eq!(tokenize("10\u{1680}000"), vec![Token::Number("10\u{1680}000".to_string())]);
assert_eq!(tokenize("10.0.1"), vec![Token::Number("10.0.1".to_string())]);
assert_eq!(tokenize("ft"), vec![Token::Unit("ft".to_string())]);
assert_eq!(
tokenize("10 ft"),
vec![
Token::Number("10".to_string()),
Token::Unit("ft".to_string()),
]
);
assert_eq!(
tokenize("ft in"),
vec![
Token::Unit("ft".to_string()),
Token::Unit("in".to_string()),
]
);
assert_eq!(
tokenize("5 ft 7 in"),
vec![
Token::Number("5".to_string()),
Token::Unit("ft".to_string()),
Token::Number("7".to_string()),
Token::Unit("in".to_string()),
]
);
assert_eq!(
tokenize("5\"7'"),
vec![
Token::Number("5".to_string()),
Token::Unit("\"".to_string()),
Token::Number("7".to_string()),
Token::Unit("'".to_string()),
]
);
assert_eq!(
tokenize(" 2.2lbs "),
vec![
Token::Number("2.2".to_string()),
Token::Unit("lbs".to_string()),
]
);
}
}