Passport MRZ

use passport_mrz.qweruse mrz_parse, mrz_find from passport_mrz      // Quillo 0.6: only these

serde_json = { version = "1", features = ["preserve_order"] }

// Hotel check-in: the guest's passport or ID card scanned, its MRZ read from the OCR text, every// check digit verified, and the guest's name and dates filled in.////   quillo run passport_mrz.qwe      then open http://localhost:8080/checkin and http://localhost:8080/carduse "passport_mrz.qwer"// What an OCR app read from the photo of ICAO's specimen passport, page noise included.let OCR = "PASSPORT  PASS  PASAPORTE\nUtopia\nEriksson, Anna Maria\nP<UTOERIKSSON<<ANNA<MARIA<<<<<<<<<<<<<<<<<<<\nL898902C36UTO7408122F1204159ZE184226B<<<<<10\n"route GET /checkin {    let mrz = mrz_find(OCR) ?? reply 422 "no MRZ in the picture: take it again"    let p = mrz_parse(mrz)    if !p.valid { reply 422 p.errors }    { guest: `${p.given_names} ${p.surname}`, document: p.document_number, nationality: p.nationality, birth_date: p.birth_date, expired: p.expired, checks: p.checks }}// ICAO's specimen ID card (TD1), and the same with one digit misread.route GET /card = {    card: mrz_parse("I<UTOD231458907<<<<<<<<<<<<<<<\n7408122F1204159UTO<<<<<<<<<<<6\nERIKSSON<<ANNA<MARIA<<<<<<<<<<"),    misread: mrz_parse("I<UTOD231458907<<<<<<<<<<<<<<<\n7408122F1204169UTO<<<<<<<<<<<6\nERIKSSON<<ANNA<MARIA<<<<<<<<<<").errors,}// A test passport for the kiosk, read back as the scanner would.route GET /make {    let mrz = mrz_make({ issuing_state: "UTO", surname: "Ñúñez-García", given_names: "José Luis", document_number: "X4RT2219", birth_date: "1990-02-28", expiry_date: "2034-06-30", sex: "M" })?    { mrz, read_back: mrz_parse(mrz).valid }}

//! Machine-readable zones of passports, ID cards and visas (ICAO 9303: TD1, TD2, TD3, MRV-A and MRV-B) read into names, document number, nationality, birth and expiry dates and sex, every check digit verified; MRZ lines found in OCR text; and MRZs made for test documents.//! version: 1.0.0//!//! [dependencies]//! serde_json = { version = "1", features = ["preserve_order"] }use serde_json::{json, Map, Value};fn value(c: char) -> Option<u32> {    match c {        '<' => Some(0),        '0'..='9' => c.to_digit(10),        'A'..='Z' => Some(c as u32 - 'A' as u32 + 10),        _ => None,    }}/// The ICAO check digit of `field` (weights 7, 3, 1; `<` counts 0, letters 10 to 35).fn check(field: &str) -> Option<u32> {    let mut t = 0;    for (i, c) in field.chars().enumerate() {        t += value(c)? * [7, 3, 1][i % 3];    }    Some(t % 10)}fn this_year() -> i32 {    let days = std::time::SystemTime::now().duration_since(std::time::UNIX_EPOCH).map_or(0, |d| d.as_secs() / 86400) as i64;    (1970.0 + days as f64 / 365.2425) as i32}fn today() -> String {    let secs = std::time::SystemTime::now().duration_since(std::time::UNIX_EPOCH).map_or(0, |d| d.as_secs()) as i64;    let z = secs.div_euclid(86400) + 719468;    let era = z.div_euclid(146097);    let doe = z - era * 146097;    let yoe = (doe - doe / 1460 + doe / 36524 - doe / 146096) / 365;    let doy = doe - (365 * yoe + yoe / 4 - yoe / 100);    let mp = (5 * doy + 2) / 153;    let d = doy - (153 * mp + 2) / 5 + 1;    let m = if mp < 10 { mp + 3 } else { mp - 9 };    format!("{:04}-{:02}-{:02}", yoe + era * 400 + if m <= 2 { 1 } else { 0 }, m, d)}/// YYMMDD as an ISO date; birth dates in the future are taken a century back, expiry dates are this century.fn date(s: &str, birth: bool) -> Option<String> {    if s.len() != 6 || !s.chars().all(|c| c.is_ascii_digit()) {        return None;    }    let (yy, m, d): (i32, u32, u32) = (s[..2].parse().ok()?, s[2..4].parse().ok()?, s[4..6].parse().ok()?);    if !(1..=12).contains(&m) || !(1..=31).contains(&d) {        return None;    }    let now = this_year();    let y = if birth { if 2000 + yy > now { 1900 + yy } else { 2000 + yy } } else if 2000 + yy > now + 50 { 1900 + yy } else { 2000 + yy };    Some(format!("{:04}-{:02}-{:02}", y, m, d))}fn names(field: &str) -> (String, String) {    let f = field.trim_end_matches('<');    let (sur, given) = f.split_once("<<").unwrap_or((f, ""));    let tidy = |s: &str| s.split('<').filter(|p| !p.is_empty()).collect::<Vec<_>>().join(" ");    (tidy(sur), tidy(given))}fn text(field: &str) -> String {    field.trim_end_matches('<').replace('<', " ").trim().to_string()}struct Checks(Map<String, Value>, Vec<String>);impl Checks {    fn add(&mut self, name: &str, field: &str, digit: char) {        let ok = match (check(field), digit) {            (Some(c), d) => d.to_digit(10) == Some(c) || (d == '<' && c == 0 && field.chars().all(|x| x == '<')),            (None, _) => false,        };        if !ok {            self.1.push(format!("the {} check digit does not match", name.replace('_', " ")));        }        self.0.insert(name.into(), json!(ok));    }}/// An MRZ read: the two or three lines as printed (line breaks between them; spaces and lower case from OCR are tidied)./// Gives `{ valid, format, document_type, issuing_state, surname, given_names, document_number, nationality, birth_date, sex,/// expiry_date, optional, expired, checks: { document_number, birth_date, expiry_date, optional, composite }, errors }`./// `valid` is true when every check digit matches; dates are ISO.pub fn mrz_parse(mrz: &str) -> Value {    let lines: Vec<String> = mrz.lines().map(|l| l.chars().filter(|c| !c.is_whitespace()).map(|c| c.to_ascii_uppercase()).collect::<String>()).filter(|l| !l.is_empty()).collect();    let shape: Vec<usize> = lines.iter().map(|l| l.len()).collect();    let mut c = Checks(Map::new(), vec![]);    let mut out = Map::new();    let visa = lines.first().is_some_and(|l| l.starts_with('V'));    let format = match (shape.as_slice(), visa) {        ([30, 30, 30], _) => "TD1",        ([36, 36], false) => "TD2",        ([44, 44], false) => "TD3",        ([44, 44], true) => "MRV-A",        ([36, 36], true) => "MRV-B",        _ => return json!({ "valid": false, "errors": [format!("an MRZ is 3 lines of 30, or 2 of 36 or 44 characters; these are {:?}", shape)] }),    };    if let Some(bad) = lines.iter().flat_map(|l| l.chars()).find(|ch| value(*ch).is_none()) {        return json!({ "valid": false, "format": format, "errors": [format!("`{}` is not an MRZ character (A-Z, 0-9 and <)", bad)] });    }    let (l1, l2) = (&lines[0], &lines[1]);    let (doc, doc_check, nat, dob, dob_check, sex, exp, exp_check, optional, name_field);    match format {        "TD1" => {            let l3 = &lines[2];            doc = &l1[5..14]; doc_check = l1.chars().nth(14).unwrap();            dob = &l2[0..6]; dob_check = l2.chars().nth(6).unwrap();            sex = l2.chars().nth(7).unwrap();            exp = &l2[8..14]; exp_check = l2.chars().nth(14).unwrap();            nat = &l2[15..18];            optional = format!("{}{}", &l1[15..30], &l2[18..29]);            name_field = l3.as_str();            // A long document number continues in the optional field, ended by its check digit.            c.add("document_number", doc, doc_check);            c.add("composite", &format!("{}{}{}{}", &l1[5..30], &l2[0..7], &l2[8..15], &l2[18..29]), l2.chars().nth(29).unwrap());        }        "TD2" | "MRV-B" => {            doc = &l2[0..9]; doc_check = l2.chars().nth(9).unwrap();            nat = &l2[10..13];            dob = &l2[13..19]; dob_check = l2.chars().nth(19).unwrap();            sex = l2.chars().nth(20).unwrap();            exp = &l2[21..27]; exp_check = l2.chars().nth(27).unwrap();            name_field = &l1[5..36];            c.add("document_number", doc, doc_check);            if format == "TD2" {                optional = l2[28..35].to_string();                c.add("composite", &format!("{}{}{}", &l2[0..10], &l2[13..20], &l2[21..35]), l2.chars().nth(35).unwrap());            } else {                optional = l2[28..36].to_string();            }        }        _ => {            doc = &l2[0..9]; doc_check = l2.chars().nth(9).unwrap();            nat = &l2[10..13];            dob = &l2[13..19]; dob_check = l2.chars().nth(19).unwrap();            sex = l2.chars().nth(20).unwrap();            exp = &l2[21..27]; exp_check = l2.chars().nth(27).unwrap();            name_field = &l1[5..44];            c.add("document_number", doc, doc_check);            if format == "TD3" {                optional = l2[28..42].to_string();                c.add("optional", &l2[28..42], l2.chars().nth(42).unwrap());                c.add("composite", &format!("{}{}{}", &l2[0..10], &l2[13..20], &l2[21..43]), l2.chars().nth(43).unwrap());            } else {                optional = l2[28..44].to_string();            }        }    }    c.add("birth_date", dob, dob_check);    c.add("expiry_date", exp, exp_check);    let (surname, given) = names(name_field);    let birth = date(dob, true);    let expiry = date(exp, false);    if birth.is_none() && dob.chars().all(|x| x.is_ascii_digit()) {        c.1.push(format!("{} is not a birth date", dob));    }    out.insert("valid".into(), json!(c.1.is_empty()));    out.insert("format".into(), json!(format));    out.insert("document_type".into(), json!(text(&l1[..2])));    out.insert("issuing_state".into(), json!(text(&l1[2..5])));    out.insert("surname".into(), json!(surname));    out.insert("given_names".into(), json!(given));    out.insert("document_number".into(), json!(text(doc)));    out.insert("nationality".into(), json!(text(nat)));    out.insert("birth_date".into(), json!(birth));    out.insert("sex".into(), json!(match sex { 'M' => "M", 'F' => "F", _ => "X" }));    out.insert("expiry_date".into(), json!(expiry));    out.insert("optional".into(), json!(text(&optional)));    out.insert("expired".into(), json!(expiry.as_ref().map(|e| *e < today())));    out.insert("checks".into(), Value::Object(c.0));    out.insert("errors".into(), json!(c.1));    Value::Object(out)}/// The MRZ lines found in OCR text of a document's photo (a run of lines of 30, 36 or 44 MRZ characters; `«` and spaces OCR adds/// are handled), joined by line breaks, or null when there is none.pub fn mrz_find(text: &str) -> Option<String> {    let lines: Vec<String> = text.lines().map(|l| l.replace('«', "<").chars().filter(|c| !c.is_whitespace()).map(|c| c.to_ascii_uppercase()).collect::<String>()).collect();    let is_mrz = |l: &str, n: usize| l.len() == n && l.contains('<') && l.chars().all(|c| value(c).is_some());    for i in 0..lines.len() {        for (n, count) in [(30usize, 3usize), (44, 2), (36, 2)] {            if i + count <= lines.len() && (0..count).all(|k| is_mrz(&lines[i + k], n)) {                return Some(lines[i..i + count].join("\n"));            }        }    }    None}/// A passport's MRZ (TD3) made from `fields`: `document_type` ("P"), `issuing_state` and `nationality` (3 letters), `surname`,/// `given_names`, `document_number`, `birth_date` and `expiry_date` (ISO dates), `sex` ("M", "F" or "X") and `optional`. Accents/// and punctuation are written as ICAO says (é as E, a hyphen as <). For test documents and kiosks' tests.pub fn mrz_make(fields: std::collections::HashMap<String, serde_json::Value>) -> Result<String, String> {    let f = |k: &str| fields.get(k).and_then(|v| v.as_str()).unwrap_or("").to_string();    let translit = |s: &str| -> String {        s.to_uppercase().chars().map(|c| match c {            'À' | 'Á' | 'Â' | 'Ã' | 'Ä' | 'Å' => 'A', 'Ç' => 'C', 'È' | 'É' | 'Ê' | 'Ë' => 'E', 'Ì' | 'Í' | 'Î' | 'Ï' => 'I', 'Ñ' => 'N',            'Ò' | 'Ó' | 'Ô' | 'Õ' | 'Ö' | 'Ø' => 'O', 'Ù' | 'Ú' | 'Û' | 'Ü' => 'U', 'Ý' => 'Y', c if c.is_ascii_alphanumeric() => c, _ => '<',        }).collect()    };    let pad = |s: String, n: usize| -> String { let mut s: String = s.chars().take(n).collect(); while s.len() < n { s.push('<'); } s };    let ymd = |iso: &str| -> Result<String, String> {        if iso.len() != 10 { return Err(format!("`{}` is not an ISO date", iso)); }        Ok(format!("{}{}{}", &iso[2..4], &iso[5..7], &iso[8..10]))    };    let state = translit(&f("issuing_state"));    let nat = if f("nationality").is_empty() { state.clone() } else { translit(&f("nationality")) };    if state.len() != 3 || nat.len() != 3 {        return Err("issuing_state and nationality are 3-letter codes".into());    }    let dtype = if f("document_type").is_empty() { "P".to_string() } else { translit(&f("document_type")) };    let name = format!("{}<<{}", translit(&f("surname")).trim_matches('<'), translit(&f("given_names")).trim_matches('<'));    let l1 = pad(format!("{}{}{}", pad(dtype, 2), state, name), 44);    let doc = pad(translit(&f("document_number")), 9);    let dob = ymd(&f("birth_date"))?;    let exp = ymd(&f("expiry_date"))?;    let sex = match f("sex").as_str() { "M" => 'M', "F" => 'F', _ => '<' };    let opt = pad(translit(&f("optional")), 14);    let cd = |s: &str| check(s).map(|d| char::from_digit(d, 10).unwrap()).ok_or_else(|| "not MRZ characters".to_string());    let opt_check = if opt.chars().all(|c| c == '<') { '<' } else { cd(&opt)? };    let mut l2 = format!("{}{}{}{}{}{}{}{}{}{}", doc, cd(&doc)?, nat, dob, cd(&dob)?, sex, exp, cd(&exp)?, opt, opt_check);    let composite = format!("{}{}{}", &l2[0..10], &l2[13..20], &l2[21..43]);    l2.push(cd(&composite)?);    Ok(format!("{}\n{}", l1, l2))}