tre-regex

Safe API bindings to the TRE regex engine.

Documentation is available at docs.rs.

Requires Rust 1.85.1 or newer.

use tre_regex::{Regex, RegcompFlags, RegexecFlags};

fn main() -> tre_regex::Result<()> {
    let regex = Regex::new(
        "^([[:alpha:]]+) ([[:alpha:]]+)$",
        RegcompFlags::EXTENDED | RegcompFlags::ICASE,
    )?;

    assert!(regex.is_match("hello world", RegexecFlags::NONE)?);
    let captures = regex.captures("hello world", 3, RegexecFlags::NONE)?;
    assert_eq!(captures[1], Some("hello"));
    Ok(())
}

Approximate matching

Approximate matching allows insertions, deletions, and substitutions with configurable costs and limits. It requires the approx feature, which is enabled by default.

use tre_regex::{Regex, RegApproxParams, RegcompFlags, RegexecFlags};

fn main() -> tre_regex::Result<()> {
    let regex = Regex::new("^(hello).*(world)$", RegcompFlags::EXTENDED)?;
    let params = RegApproxParams::new()
        .cost_ins(1)
        .cost_del(1)
        .cost_subst(1)
        .max_cost(2)
        .max_ins(2)
        .max_del(2)
        .max_subst(2)
        .max_err(2);

    let result = regex.regaexec("hullo warld", &params, 3, RegexecFlags::NONE)?;
    assert_eq!(result.cost(), 2);
    assert_eq!(result.get_matches()[0], Some("hullo warld"));
    assert_eq!(result.get_matches()[1], Some("hullo"));
    assert_eq!(result.get_matches()[2], Some("warld"));
    Ok(())
}

Wide-character matching

Wide-character support requires the default wchar feature and a dependency on the widestring crate.

use tre_regex::{Regex, RegcompFlags, RegexecFlags};
use widestring::widestr;

fn main() -> tre_regex::Result<()> {
    let regex = Regex::new_wide(
        widestr!("^(hello).*(world)$"),
        RegcompFlags::EXTENDED | RegcompFlags::ICASE,
    )?;
    let captures = regex.regwexec(widestr!("hello wide world"), 3, RegexecFlags::NONE)?;

    assert_eq!(captures[1], Some(widestr!("hello")));
    assert_eq!(captures[2], Some(widestr!("world")));
    Ok(())
}

Features

• wchar: enable wide-character regex support. Enabled by default.
• approx: enable approximate matching support. Enabled by default.
• vendored: use the vendored copy of TRE with tre-regex-sys; otherwise use the system TRE. Enabled by default.