//! Text from outside (model output, tool arguments, server errors) made safe to show: control
//! characters, zero-width and bidirectional marks, line separators and the BOM become `\uXXXX`.
//! Used by `bxctl` for the terminal and by `gatewayd` for Mattermost, so both show the same.

// The control bytes, the zero-width and Unicode line/paragraph separators, and the BOM. A code
// point in any of these ranges must never reach the terminal raw.
const ESCAPE_RANGES: &[(u32, u32)] = &[
    (0x0000, 0x001f),
    (0x007f, 0x009f),
    (0x200b, 0x200f),
    (0x2028, 0x202e),
    (0x2060, 0x2069),
    (0xfeff, 0xfeff),
];

pub fn escape_json_text(s: &str) -> String {
    escape(s, false)
}

pub fn escape_model_text(s: &str) -> String {
    escape(s, true)
}

fn escape(s: &str, model_text: bool) -> String {
    let mut out = String::with_capacity(s.len());
    for c in s.chars() {
        let cp = u32::from(c);
        if should_escape(c, model_text) {
            out.push_str(&format!("\\u{:04x}", cp));
        } else {
            out.push(c);
        }
    }
    out
}

// A control point is escaped, except that model text keeps \n and \t.
fn should_escape(c: char, model_text: bool) -> bool {
    let cp = u32::from(c);
    if ESCAPE_RANGES.iter().any(|(lo, hi)| cp >= *lo && cp <= *hi) {
        !(model_text && (cp == 0x09 || cp == 0x0a))
    } else {
        false
    }
}