lolistealer/src/tags.rs

use std::{
    cmp::PartialEq,
    fmt,
    mem,
    collections::HashSet,
    str,
};

/// Tag matching rules
#[derive(Clone,Debug,PartialEq,Eq,Hash)]
pub enum Rule
{
    /// This tag /must/ be present.
    Required,
    /// At least one tag marked `optional` must be present.
    Optional,
    /// This tag /must not/ be present.
    Rejected,
}

/// Represents a tag
#[derive(Clone,Debug,PartialEq,Eq,Hash)]
pub struct Tag
{
    rule: Rule,
    strings: Vec<String>,
    normal_idx: usize,
    repr: RepresentationMode,
}

/// How should tags be interpreted?
#[derive(Debug,Clone,PartialEq,Eq,Hash)]
pub enum RepresentationMode
{
    /// Ignores spaces, case, underscores, things in parenthesis, and special symbols.
    Lenient,
    /// Ignores spaces, case, and underscores.
    Normal,
    /// Ignores nothing
    Strict,
}

impl Default for RepresentationMode
{
    fn default() -> Self
    {
	Self::Strict
    }
}

/// Remove consecutive whitespace, and normalise them to `' '`
fn remove_whitespace<T: IntoIterator<Item=char>>(input: T) -> String
{
    let mut last = false;
    input.into_iter()
	.filter_map(|ch| {
	    if ch.is_whitespace() {
		if last {
		    None
		} else {
		    last = true;
		    Some(' ')
		}
	    }
	    else {	
		last = false;
		Some(ch)
	    }
	})
	.collect()
}

/// How big are the max output of `fuzz` likely to be?
const FUZZ_SIZE_HINT: usize = 7;
/// Creates representations
fn fuzz<S: Into<String>>(input: S, output: &mut Vec<String>, mode: &RepresentationMode) -> usize
{
    let input = input.into();

    match mode {
	RepresentationMode::Strict => {
	    output.push(input);
	    0
	},
	RepresentationMode::Normal => {
	    output.push(input.to_lowercase());
	    {
		let normal = remove_whitespace(input.chars());
		
		output.push(normal.replace(" ", "_"));
		output.push(normal.chars().filter(|ch| !ch.is_whitespace()).collect());
		output.push(normal);
	    }
	    output.push(input);
	    output.len()-1
	},
	RepresentationMode::Lenient => {
	    const REMOVE: &[char] = &[
		':',
		'.',
		',',
		' ',
	    ];
	    output.push(input.chars().filter(|ch| !REMOVE.contains(ch)).collect());
	    fuzz(input, output, mode)
	},
    }
}

impl Tag
{

    /// Create a new tag representation
    pub fn new<S: Into<String>>(tag: S, rule: Rule, mode: RepresentationMode) -> Self
    {
	let mut fz = Vec::with_capacity(FUZZ_SIZE_HINT);
	let normal_idx = fuzz(tag, &mut fz, &mode);
	Self {
	    rule,
	    strings: fz,
	    normal_idx,
	    repr: mode,
	}
    }
}

/// Returned when an empty tag is tried to be parsed
#[derive(Debug)]
pub struct ParseError;

impl std::error::Error for ParseError{}

impl fmt::Display for ParseError
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	write!(f, "cannot have empty tag rule")
    }
}

impl str::FromStr for Tag
{
    type Err = ParseError;
    fn from_str(s: &str) -> std::result::Result<Self, Self::Err>
    {
	let mut chars = s.chars();
	let mut next = String::new();
	let rule = match chars.next() {
	    Some('+') => Rule::Optional,
	    Some('-') => Rule::Rejected,
	    Some(x) => {
		next.push(x);
		Rule::Required
	    },
	    _ => return Err(ParseError),
	};
	let s = chars.as_str();
	if s.len() == 0 {
	    return Err(ParseError);
	}
	
	Ok(Self::new(if next.len() > 0 {
	    next.push_str(s);
	    next
	} else {s.into()}, rule, RepresentationMode::default()))
    }
}

impl<T> PartialEq<T> for Tag
where T: AsRef<str>
{
    fn eq(&self, other: &T) -> bool
    {
	let mut fz = Vec::with_capacity(FUZZ_SIZE_HINT);
	fuzz(other.as_ref(), &mut fz, &self.repr);
	for (i,j) in self.strings.iter().zip(fz.into_iter())
	{
	    if i == j.as_str() {
		return true;
	    }
	}
	false
    }
}

impl From<Tag> for String
{
    fn from(mut tag: Tag) -> Self
    {
	mem::replace(&mut tag.strings[tag.normal_idx], Default::default())
    }
}

impl fmt::Display for Tag
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	write!(f, "{}", &self.strings[self.normal_idx])
    }
}

#[derive(Debug)]
pub enum Error<'a>
{
    Required(&'a Tag),
    NoOptional,
    Rejected(&'a Tag),
}

#[derive(Debug)]
pub enum OwnedError
{
    Required(Tag),
    NoOptional,
    Rejected(Tag),
}

pub type Result<'a> = std::result::Result<Vec<&'a Tag>, Error<'a>>;

/// Search a set of strings for a matching ruleset for `tags`. Returns the matched tags on success
pub fn search<'a, T, U, V>(matches: T, tags: U) -> Result<'a>
where T: IntoIterator<Item=V>,
      U: IntoIterator<Item=&'a Tag>,
      V: AsRef<str>,
{
    let tags: Vec<&'a Tag> = tags.into_iter().collect();
    let mut output = Vec::with_capacity(tags.len());
    
    let mut matched = HashSet::new();
    for matches in matches.into_iter()
    {
	for tag in tags.iter() {
	    if *tag == &matches {
		matched.insert(tag);
		break;
	    }
	}
    }

    let mut needs_opt = false;
    let mut has_opt = false;
    for tag in tags.iter()
    {
	if matched.contains(tag) {
	    match tag.rule {
		Rule::Optional => {has_opt = true; needs_opt = true;},
		Rule::Rejected => return Err(Error::Rejected(tag)),
		_ => (),
	    }

	    output.push(tag.clone());
	} else if tag.rule == Rule::Optional {
	    needs_opt = true;
	} else if tag.rule == Rule::Required {
	    return Err(Error::Required(tag));
	}
    }

    if needs_opt && !has_opt {
	return Err(Error::NoOptional);
    }
    
    Ok(output)
}

/// Parse a string of many tag rules, ignoring empty or invalid entries
#[inline]
pub fn parse<T>(string: T) -> Vec<Tag>
where T: AsRef<str>
{
    let string = string.as_ref();
    parse_iter(string.split(" "))
}

/// Parse a collection of tag rules, ignoring empty or invalid entries
pub fn parse_iter<T,I>(from: I) -> Vec<Tag>
where I: IntoIterator<Item=T>,
      T: AsRef<str>
{
    from.into_iter()
	.filter_map(|s| {
	    if let Ok(tag) = s.as_ref().parse() {
		Some(tag)
	    } else {
		None
	    }
	}).collect()
}

impl fmt::Display for Error<'_>
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	match self {
	    Self::NoOptional => write!(f, "there were no optional tags, at least one is needed"),
	    Self::Rejected(tag) => write!(f, "a rejected tag was present: {}", tag),
	    Self::Required(tag) => write!(f, "a required tag was not present: {}", tag),
	}
    }
}


impl fmt::Display for OwnedError
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	match self {
	    Self::NoOptional => write!(f, "there were no optional tags, at least one is needed"),
	    Self::Rejected(tag) => write!(f, "a rejected tag was present: {}", tag),
	    Self::Required(tag) => write!(f, "a required tag was not present: {}", tag),
	}
    }
}

impl std::error::Error for Error<'_>{}
impl std::error::Error for OwnedError{}

impl From<Error<'_>> for OwnedError
{
    fn from(er: Error<'_>) -> Self
    {
	match er {
	    Error::NoOptional => Self::NoOptional,
	    Error::Rejected(t) => Self::Rejected(t.to_owned()),
	    Error::Required(t) => Self::Required(t.to_owned()),
	}
    }
}