path: root/doc/nom_recipes.md

# Nom Recipes

These are short recipes for accomplishing common tasks with nom.

* [Whitespace](#whitespace)
  + [Wrapper combinators that eat whitespace before and after a parser](#wrapper-combinators-that-eat-whitespace-before-and-after-a-parser)
* [Comments](#comments)
  + [`// C++/EOL-style comments`](#-ceol-style-comments)
  + [`/* C-style comments */`](#-c-style-comments-)
* [Identifiers](#identifiers)
  + [`Rust-Style Identifiers`](#rust-style-identifiers)
* [Literal Values](#literal-values)
  + [Escaped Strings](#escaped-strings)
  + [Integers](#integers)
    - [Hexadecimal](#hexadecimal)
    - [Octal](#octal)
    - [Binary](#binary)
    - [Decimal](#decimal)
  + [Floating Point Numbers](#floating-point-numbers)

## Whitespace



### Wrapper combinators that eat whitespace before and after a parser

```rust
use nom::{
  IResult,
  error::ParseError,
  combinator::value,
  sequence::delimited,
  character::complete::multispace0,
};

/// A combinator that takes a parser `inner` and produces a parser that also consumes both leading and 
/// trailing whitespace, returning the output of `inner`.
fn ws<'a, F: 'a, O, E: ParseError<&'a str>>(inner: F) -> impl FnMut(&'a str) -> IResult<&'a str, O, E>
  where
  F: Fn(&'a str) -> IResult<&'a str, O, E>,
{
  delimited(
    multispace0,
    inner,
    multispace0
  )
}
```

To eat only trailing whitespace, replace `delimited(...)` with `terminated(&inner, multispace0)`.
Likewise, the eat only leading whitespace, replace `delimited(...)` with `preceded(multispace0,
&inner)`. You can use your own parser instead of `multispace0` if you want to skip a different set
of lexemes.

## Comments

### `// C++/EOL-style comments`

This version uses `%` to start a comment, does not consume the newline character, and returns an
output of `()`.

```rust
use nom::{
  IResult,
  error::ParseError,
  combinator::value,
  sequence::pair,
  bytes::complete::is_not,
  character::complete::char,
};

pub fn peol_comment<'a, E: ParseError<&'a str>>(i: &'a str) -> IResult<&'a str, (), E>
{
  value(
    (), // Output is thrown away.
    pair(char('%'), is_not("\n\r"))
  )(i)
}
```

### `/* C-style comments */`

Inline comments surrounded with sentinel tags `(*` and `*)`. This version returns an output of `()`
and does not handle nested comments.

```rust
use nom::{
  IResult,
  error::ParseError,
  combinator::value,
  sequence::tuple,
  bytes::complete::{tag, take_until},
};

pub fn pinline_comment<'a, E: ParseError<&'a str>>(i: &'a str) -> IResult<&'a str, (), E> {
  value(
    (), // Output is thrown away.
    tuple((
      tag("(*"),
      take_until("*)"),
      tag("*)")
    ))
  )(i)
}
```

## Identifiers

### `Rust-Style Identifiers`

Parsing identifiers that may start with a letter (or underscore) and may contain underscores,
letters and numbers may be parsed like this:

```rust
use nom::{
  IResult,
  branch::alt,
  multi::many0,
  combinator::recognize,
  sequence::pair,
  character::complete::{alpha1, alphanumeric1},
  bytes::complete::tag,
};

pub fn identifier(input: &str) -> IResult<&str, &str> {
  recognize(
    pair(
      alt((alpha1, tag("_"))),
      many0(alt((alphanumeric1, tag("_"))))
    )
  )(input)
}
```

Let's say we apply this to the identifier `hello_world123abc`. The first `alt` parser would
recognize `h`. The `pair` combinator ensures that `ello_world123abc` will be piped to the next
`alphanumeric0` parser, which recognizes every remaining character. However, the `pair` combinator
returns a tuple of the results of its sub-parsers. The `recognize` parser produces a `&str` of the
input text that was parsed, which in this case is the entire `&str` `hello_world123abc`.

## Literal Values

### Escaped Strings

This is [one of the examples](https://github.com/Geal/nom/blob/master/examples/string.rs) in the
examples directory.

### Integers

The following recipes all return string slices rather than integer values. How to obtain an
integer value instead is demonstrated for hexadecimal integers. The others are similar.

The parsers allow the grouping character `_`, which allows one to group the digits by byte, for
example: `0xA4_3F_11_28`. If you prefer to exclude the `_` character, the lambda to convert from a
string slice to an integer value is slightly simpler. You can also strip the `_` from the string
slice that is returned, which is demonstrated in the second hexdecimal number parser.

If you wish to limit the number of digits in a valid integer literal, replace `many1` with
`many_m_n` in the recipes.

#### Hexadecimal

The parser outputs the string slice of the digits without the leading `0x`/`0X`.

```rust
use nom::{
  IResult,
  branch::alt,
  multi::{many0, many1},
  combinator::recognize,
  sequence::{preceded, terminated},
  character::complete::{char, one_of},
  bytes::complete::tag,
};

fn hexadecimal(input: &str) -> IResult<&str, &str> { // <'a, E: ParseError<&'a str>>
  preceded(
    alt((tag("0x"), tag("0X"))),
    recognize(
      many1(
        terminated(one_of("0123456789abcdefABCDEF"), many0(char('_')))
      )
    )
  )(input)
}
```

If you want it to return the integer value instead, use map:

```rust
use nom::{
  IResult,
  branch::alt,
  multi::{many0, many1},
  combinator::{map_res, recognize},
  sequence::{preceded, terminated},
  character::complete::{char, one_of},
  bytes::complete::tag,
};

fn hexadecimal_value(input: &str) -> IResult<&str, i64> {
  map_res(
    preceded(
      alt((tag("0x"), tag("0X"))),
      recognize(
        many1(
          terminated(one_of("0123456789abcdefABCDEF"), many0(char('_')))
        )
      )
    ),
    |out: &str| i64::from_str_radix(&str::replace(&out, "_", ""), 16)
  )(input)
}
```

#### Octal

```rust
use nom::{
  IResult,
  branch::alt,
  multi::{many0, many1},
  combinator::recognize,
  sequence::{preceded, terminated},
  character::complete::{char, one_of},
  bytes::complete::tag,
};

fn octal(input: &str) -> IResult<&str, &str> {
  preceded(
    alt((tag("0o"), tag("0O"))),
    recognize(
      many1(
        terminated(one_of("01234567"), many0(char('_')))
      )
    )
  )(input)
}
```

#### Binary

```rust
use nom::{
  IResult,
  branch::alt,
  multi::{many0, many1},
  combinator::recognize,
  sequence::{preceded, terminated},
  character::complete::{char, one_of},
  bytes::complete::tag,
};

fn binary(input: &str) -> IResult<&str, &str> {
  preceded(
    alt((tag("0b"), tag("0B"))),
    recognize(
      many1(
        terminated(one_of("01"), many0(char('_')))
      )
    )
  )(input)
}
```

#### Decimal

```rust
use nom::{
  IResult,
  multi::{many0, many1},
  combinator::recognize,
  sequence::terminated,
  character::complete::{char, one_of},
};

fn decimal(input: &str) -> IResult<&str, &str> {
  recognize(
    many1(
      terminated(one_of("0123456789"), many0(char('_')))
    )
  )(input)
}
```

### Floating Point Numbers

The following is adapted from [the Python parser by Valentin Lorentz (ProgVal)](https://github.com/ProgVal/rust-python-parser/blob/master/src/numbers.rs).

```rust
use nom::{
  IResult,
  branch::alt,
  multi::{many0, many1},
  combinator::{opt, recognize},
  sequence::{preceded, terminated, tuple},
  character::complete::{char, one_of},
};

fn float(input: &str) -> IResult<&str, &str> {
  alt((
    // Case one: .42
    recognize(
      tuple((
        char('.'),
        decimal,
        opt(tuple((
          one_of("eE"),
          opt(one_of("+-")),
          decimal
        )))
      ))
    )
    , // Case two: 42e42 and 42.42e42
    recognize(
      tuple((
        decimal,
        opt(preceded(
          char('.'),
          decimal,
        )),
        one_of("eE"),
        opt(one_of("+-")),
        decimal
      ))
    )
    , // Case three: 42. and 42.42
    recognize(
      tuple((
        decimal,
        char('.'),
        opt(decimal)
      ))
    )
  ))(input)
}

fn decimal(input: &str) -> IResult<&str, &str> {
  recognize(
    many1(
      terminated(one_of("0123456789"), many0(char('_')))
    )
  )(input)
}
```

# implementing FromStr

The [FromStr trait](https://doc.rust-lang.org/std/str/trait.FromStr.html) provides
a common interface to parse from a string.

```rust
use nom::{
  IResult, Finish, error::Error,
  bytes::complete::{tag, take_while},
};
use std::str::FromStr;

// will recognize the name in "Hello, name!"
fn parse_name(input: &str) -> IResult<&str, &str> {
  let (i, _) = tag("Hello, ")(input)?;
  let (i, name) = take_while(|c:char| c.is_alphabetic())(i)?;
  let (i, _) = tag("!")(i)?;

  Ok((i, name))
}

// with FromStr, the result cannot be a reference to the input, it must be owned
#[derive(Debug)]
pub struct Name(pub String);

impl FromStr for Name {
  // the error must be owned as well
  type Err = Error<String>;

  fn from_str(s: &str) -> Result<Self, Self::Err> {
      match parse_name(s).finish() {
          Ok((_remaining, name)) => Ok(Name(name.to_string())),
          Err(Error { input, code }) => Err(Error {
              input: input.to_string(),
              code,
          })
      }
  }
}

fn main() {
  // parsed: Ok(Name("nom"))
  println!("parsed: {:?}", "Hello, nom!".parse::<Name>());

  // parsed: Err(Error { input: "123!", code: Tag })
  println!("parsed: {:?}", "Hello, 123!".parse::<Name>());
}
```