@rudra-js/attested 0.7.4 → 0.7.6

package/dist/verify.js

@@ -1,43 +1,47 @@
 import { numeralsIn, supportedValues } from './numerals.js';
 import { BANNED_PHRASES, indexPhrasing, normalisePhrasing, spansIn, } from './phrases.js';
-const readings = new WeakMap();
-function unchanged(held, values) {
-    if (held.length !== values.length)
+const cache = new WeakMap();
+function sameValues(a, b) {
+    if (a.length !== b.length)
         return false;
-    for (let i = 0; i < held.length; i += 1) {
-        if (held[i] !== values[i])
+    for (let i = 0; i < a.length; i += 1) {
+        if (a[i] !== b[i])
             return false;
     }
     return true;
 }
-function valuesOf(facts) {
+function supportedOf(facts) {
     const values = facts.values;
     if (!Array.isArray(values))
         return supportedValues(values);
-    const read = readings.get(values);
-    if (read !== undefined && unchanged(read.held, values))
-        return read.supported;
+    const cached = cache.get(values);
+    if (cached !== undefined && sameValues(cached.copy, values))
+        return cached.supported;
     const supported = supportedValues(values);
-    readings.set(values, { held: [...values], supported });
+    cache.set(values, { copy: [...values], supported });
     return supported;
 }
-function phrasesOf(facts) {
-    const phrases = new Set(BANNED_PHRASES);
-    for (const phrase of facts.bannedPhrases ?? []) {
+function normaliseAll(phrases) {
+    const out = [];
+    for (const phrase of phrases ?? []) {
         const normalised = normalisePhrasing(phrase);
         if (normalised.length > 0)
-            phrases.add(normalised);
+            out.push(normalised);
     }
-    return [...phrases];
+    return out;
 }
-function allowedOf(facts) {
-    const allowed = [];
-    for (const phrase of facts.allowedPhrases ?? []) {
-        const normalised = normalisePhrasing(phrase);
-        if (normalised.length > 0)
-            allowed.push(normalised);
-    }
-    return allowed;
+function bannedOf(facts) {
+    const banned = new Set(BANNED_PHRASES);
+    for (const phrase of normaliseAll(facts.bannedPhrases))
+        banned.add(phrase);
+    return [...banned];
+}
+function readFacts(facts) {
+    return {
+        supported: supportedOf(facts),
+        banned: bannedOf(facts),
+        allowed: normaliseAll(facts.allowedPhrases),
+    };
 }
 function reasonFor(numeral) {
     if (numeral.kind === 'other-numeral') {
@@ -53,12 +57,9 @@ function checkQuantities(text, supported) {
     const findings = [];
     const reported = new Set();
     for (const numeral of numerals) {
-        let backed = false;
-        for (const form of numeral.forms) {
-            if (supported.has(form))
-                backed = true;
-        }
-        if (backed || reported.has(numeral.token))
+        if (numeral.forms.some((form) => supported.has(form)))
+            continue;
+        if (reported.has(numeral.token))
             continue;
         reported.add(numeral.token);
         findings.push({ layer: 'quantity', token: numeral.token, reason: reasonFor(numeral) });
@@ -70,41 +71,43 @@ function checkQuantities(text, supported) {
         findings,
     };
 }
-function standsBehind(allowed, hit) {
-    for (const span of allowed) {
+function isInside(spans, hit) {
+    for (const span of spans) {
         if (span.start <= hit.start && hit.end <= span.end)
             return true;
     }
     return false;
 }
-function checkWording(text, phrases, allowed) {
-    // Read twice, since Ν reads as N but ν as v; both readings are the same length, so spans line up.
-    const texts = [normalisePhrasing(text)];
-    const byLook = normalisePhrasing(text, true);
-    if (byLook !== texts[0])
-        texts.push(byLook);
-    const scans = [];
-    const spans = [];
-    for (const normalised of texts) {
-        const indexed = indexPhrasing(normalised);
-        scans.push(indexed);
+function forgivenInEitherReading(scans, allowed) {
+    const forgiven = [];
+    for (const scan of scans) {
         for (const phrase of allowed) {
-            for (const span of spansIn(indexed, phrase)) {
-                if (!normalised.slice(span.start, span.end + 1).includes('\n'))
-                    spans.push(span);
+            for (const span of spansIn(scan, phrase)) {
+                if (!scan.text.slice(span.start, span.end + 1).includes('\n'))
+                    forgiven.push(span);
             }
         }
     }
-    const findings = [];
-    for (const phrase of phrases) {
-        let caught = false;
-        for (const indexed of scans) {
-            for (const hit of spansIn(indexed, phrase)) {
-                if (!standsBehind(spans, hit))
-                    caught = true;
-            }
+    return forgiven;
+}
+function isCaught(scans, phrase, forgiven) {
+    for (const scan of scans) {
+        for (const hit of spansIn(scan, phrase)) {
+            if (!isInside(forgiven, hit))
+                return true;
         }
-        if (!caught)
+    }
+    return false;
+}
+function checkWording(text, prepared) {
+    const { banned, allowed } = prepared;
+    const plain = normalisePhrasing(text);
+    const byLook = normalisePhrasing(text, true);
+    const scans = byLook === plain ? [indexPhrasing(plain)] : [indexPhrasing(plain), indexPhrasing(byLook)];
+    const forgiven = forgivenInEitherReading(scans, allowed);
+    const findings = [];
+    for (const phrase of banned) {
+        if (!isCaught(scans, phrase, forgiven))
             continue;
         findings.push({
             layer: 'wording',
@@ -115,34 +118,32 @@ function checkWording(text, phrases, allowed) {
     return {
         supported: findings.length === 0,
         strength: 'best-effort',
-        checked: phrases.length,
+        checked: banned.length,
         findings,
     };
 }
-function check(text, supported, phrases, allowed) {
-    const quantity = checkQuantities(text, supported);
-    const wording = checkWording(text, phrases, allowed);
+function check(text, prepared) {
+    const quantity = checkQuantities(text, prepared.supported);
+    const wording = checkWording(text, prepared);
     return { supported: quantity.supported && wording.supported, quantity, wording };
 }
 export function verify(text, facts) {
-    return check(text, valuesOf(facts), phrasesOf(facts), allowedOf(facts));
+    return check(text, readFacts(facts));
 }
 export function verifyFields(fields, facts) {
-    const supported = valuesOf(facts);
-    const phrases = phrasesOf(facts);
-    const allowed = allowedOf(facts);
+    const prepared = readFacts(facts);
     const results = [];
-    const written = [];
+    const texts = [];
     let allSupported = true;
     for (const field of Object.keys(fields)) {
         const text = fields[field] ?? '';
-        const result = check(text, supported, phrases, allowed);
+        const result = check(text, prepared);
         if (!result.supported)
             allSupported = false;
         results.push({ field, result });
-        written.push(text);
+        texts.push(text);
     }
-    const acrossFields = checkWording(written.join(' '), phrases, allowed);
+    const acrossFields = checkWording(texts.join(' '), prepared);
     return { supported: allSupported && acrossFields.supported, fields: results, acrossFields };
 }
 //# sourceMappingURL=verify.js.map

package/dist/verify.js.map

@@ -1 +1 @@
-{"version":3,"file":"verify.js","sourceRoot":"","sources":["../src/verify.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,eAAe,EAAgB,MAAM,eAAe,CAAC;AAC1E,OAAO,EACL,cAAc,EACd,aAAa,EACb,iBAAiB,EACjB,OAAO,GAGR,MAAM,cAAc,CAAC;AAgDtB,MAAM,QAAQ,GAAG,IAAI,OAAO,EAAgE,CAAC;AAE7F,SAAS,SAAS,CAAC,IAAwB,EAAE,MAA0B;IACrE,IAAI,IAAI,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IAChD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACxC,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,MAAM,CAAC,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;IAC1C,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,QAAQ,CAAC,KAAY;IAC5B,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;IAE5B,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;QAAE,OAAO,eAAe,CAAC,MAAM,CAAC,CAAC;IAE3D,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAClC,IAAI,IAAI,KAAK,SAAS,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC;QAAE,OAAO,IAAI,CAAC,SAAS,CAAC;IAE9E,MAAM,SAAS,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;IAC1C,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,GAAG,MAAM,CAAC,EAAE,SAAS,EAAE,CAAC,CAAC;IACvD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,SAAS,CAAC,KAAY;IAC7B,MAAM,OAAO,GAAG,IAAI,GAAG,CAAS,cAAc,CAAC,CAAC;IAEhD,KAAK,MAAM,MAAM,IAAI,KAAK,CAAC,aAAa,IAAI,EAAE,EAAE,CAAC;QAC/C,MAAM,UAAU,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAC;QAC7C,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IACrD,CAAC;IACD,OAAO,CAAC,GAAG,OAAO,CAAC,CAAC;AACtB,CAAC;AAED,SAAS,SAAS,CAAC,KAAY;IAC7B,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,KAAK,MAAM,MAAM,IAAI,KAAK,CAAC,cAAc,IAAI,EAAE,EAAE,CAAC;QAChD,MAAM,UAAU,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAC;QAC7C,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IACtD,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,SAAS,CAAC,OAAgB;IACjC,IAAI,OAAO,CAAC,IAAI,KAAK,eAAe,EAAE,CAAC;QACrC,OAAO,cAAc,OAAO,CAAC,KAAK,yEAAyE,CAAC;IAC9G,CAAC;IACD,IAAI,OAAO,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;QACjC,OAAO,cAAc,OAAO,CAAC,KAAK,uFAAuF,CAAC;IAC5H,CAAC;IACD,OAAO,cAAc,OAAO,CAAC,KAAK,2CAA2C,CAAC;AAChF,CAAC;AAED,SAAS,eAAe,CAAC,IAAY,EAAE,SAAsB;IAC3D,MAAM,QAAQ,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IAEnC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;QAC/B,IAAI,MAAM,GAAG,KAAK,CAAC;QACnB,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;YACjC,IAAI,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC;gBAAE,MAAM,GAAG,IAAI,CAAC;QACzC,CAAC;QACD,IAAI,MAAM,IAAI,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC;YAAE,SAAS;QAEpD,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;QAC5B,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACzF,CAAC;IAED,OAAO;QACL,SAAS,EAAE,QAAQ,CAAC,MAAM,KAAK,CAAC;QAChC,QAAQ,EAAE,OAAO;QACjB,OAAO,EAAE,QAAQ,CAAC,MAAM;QACxB,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,SAAS,YAAY,CAAC,OAAe,EAAE,GAAS;IAC9C,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,CAAC;QAC3B,IAAI,IAAI,CAAC,KAAK,IAAI,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,IAAI,CAAC;IAClE,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,YAAY,CAAC,IAAY,EAAE,OAAiB,EAAE,OAAiB;IACtE,kGAAkG;IAClG,MAAM,KAAK,GAAG,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC;IACxC,MAAM,MAAM,GAAG,iBAAiB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IAC7C,IAAI,MAAM,KAAK,KAAK,CAAC,CAAC,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAE5C,MAAM,KAAK,GAAc,EAAE,CAAC;IAC5B,MAAM,KAAK,GAAW,EAAE,CAAC;IACzB,KAAK,MAAM,UAAU,IAAI,KAAK,EAAE,CAAC;QAC/B,MAAM,OAAO,GAAG,aAAa,CAAC,UAAU,CAAC,CAAC;QAC1C,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACpB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,CAAC;gBAC5C,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC;oBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACnF,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,MAAM,GAAG,KAAK,CAAC;QACnB,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;YAC5B,KAAK,MAAM,GAAG,IAAI,OAAO,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,CAAC;gBAC3C,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,GAAG,CAAC;oBAAE,MAAM,GAAG,IAAI,CAAC;YAC/C,CAAC;QACH,CAAC;QACD,IAAI,CAAC,MAAM;YAAE,SAAS;QAEtB,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,SAAS;YAChB,KAAK,EAAE,MAAM;YACb,MAAM,EAAE,aAAa,MAAM,uEAAuE;SACnG,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,SAAS,EAAE,QAAQ,CAAC,MAAM,KAAK,CAAC;QAChC,QAAQ,EAAE,aAAa;QACvB,OAAO,EAAE,OAAO,CAAC,MAAM;QACvB,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,SAAS,KAAK,CACZ,IAAY,EACZ,SAAsB,EACtB,OAAiB,EACjB,OAAiB;IAEjB,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;IAClD,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;IACrD,OAAO,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,IAAI,OAAO,CAAC,SAAS,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;AACnF,CAAC;AAED,MAAM,UAAU,MAAM,CAAC,IAAY,EAAE,KAAY;IAC/C,OAAO,KAAK,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC,EAAE,SAAS,CAAC,KAAK,CAAC,EAAE,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC;AAC1E,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,MAA8B,EAAE,KAAY;IACvE,MAAM,SAAS,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IAClC,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IACjC,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAEjC,MAAM,OAAO,GAAkB,EAAE,CAAC;IAClC,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,YAAY,GAAG,IAAI,CAAC;IACxB,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QACxC,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QACjC,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,CAAC,SAAS;YAAE,YAAY,GAAG,KAAK,CAAC;QAC5C,OAAO,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC;QAChC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACrB,CAAC;IAED,MAAM,YAAY,GAAG,YAAY,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;IAEvE,OAAO,EAAE,SAAS,EAAE,YAAY,IAAI,YAAY,CAAC,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,YAAY,EAAE,CAAC;AAC9F,CAAC"}
+{"version":3,"file":"verify.js","sourceRoot":"","sources":["../src/verify.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,eAAe,EAAgB,MAAM,eAAe,CAAC;AAC1E,OAAO,EACL,cAAc,EACd,aAAa,EACb,iBAAiB,EACjB,OAAO,GAGR,MAAM,cAAc,CAAC;AAwCtB,MAAM,KAAK,GAAG,IAAI,OAAO,EAAgE,CAAC;AAE1F,SAAS,UAAU,CAAC,CAAqB,EAAE,CAAqB;IAC9D,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IACxC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACrC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;IAClC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,WAAW,CAAC,KAAY;IAC/B,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;IAC5B,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;QAAE,OAAO,eAAe,CAAC,MAAM,CAAC,CAAC;IAE3D,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IACjC,IAAI,MAAM,KAAK,SAAS,IAAI,UAAU,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC;QAAE,OAAO,MAAM,CAAC,SAAS,CAAC;IAErF,MAAM,SAAS,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;IAC1C,KAAK,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,GAAG,MAAM,CAAC,EAAE,SAAS,EAAE,CAAC,CAAC;IACpD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,YAAY,CAAC,OAAsC;IAC1D,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,IAAI,EAAE,EAAE,CAAC;QACnC,MAAM,UAAU,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAC;QAC7C,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;YAAE,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAClD,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,QAAQ,CAAC,KAAY;IAC5B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,cAAc,CAAC,CAAC;IACvC,KAAK,MAAM,MAAM,IAAI,YAAY,CAAC,KAAK,CAAC,aAAa,CAAC;QAAE,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAC3E,OAAO,CAAC,GAAG,MAAM,CAAC,CAAC;AACrB,CAAC;AAQD,SAAS,SAAS,CAAC,KAAY;IAC7B,OAAO;QACL,SAAS,EAAE,WAAW,CAAC,KAAK,CAAC;QAC7B,MAAM,EAAE,QAAQ,CAAC,KAAK,CAAC;QACvB,OAAO,EAAE,YAAY,CAAC,KAAK,CAAC,cAAc,CAAC;KAC5C,CAAC;AACJ,CAAC;AAED,SAAS,SAAS,CAAC,OAAgB;IACjC,IAAI,OAAO,CAAC,IAAI,KAAK,eAAe,EAAE,CAAC;QACrC,OAAO,cAAc,OAAO,CAAC,KAAK,yEAAyE,CAAC;IAC9G,CAAC;IACD,IAAI,OAAO,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;QACjC,OAAO,cAAc,OAAO,CAAC,KAAK,uFAAuF,CAAC;IAC5H,CAAC;IACD,OAAO,cAAc,OAAO,CAAC,KAAK,2CAA2C,CAAC;AAChF,CAAC;AAED,SAAS,eAAe,CAAC,IAAY,EAAE,SAAsB;IAC3D,MAAM,QAAQ,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IAEnC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;QAC/B,IAAI,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAAE,SAAS;QAChE,IAAI,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC;YAAE,SAAS;QAE1C,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;QAC5B,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACzF,CAAC;IAED,OAAO;QACL,SAAS,EAAE,QAAQ,CAAC,MAAM,KAAK,CAAC;QAChC,QAAQ,EAAE,OAAO;QACjB,OAAO,EAAE,QAAQ,CAAC,MAAM;QACxB,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,SAAS,QAAQ,CAAC,KAAa,EAAE,GAAS;IACxC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,KAAK,IAAI,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,IAAI,CAAC;IAClE,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,uBAAuB,CAAC,KAAgB,EAAE,OAA0B;IAC3E,MAAM,QAAQ,GAAW,EAAE,CAAC;IAC5B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,CAAC;gBACzC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC;oBAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACrF,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,QAAQ,CAAC,KAAgB,EAAE,MAAc,EAAE,QAAgB;IAClE,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,KAAK,MAAM,GAAG,IAAI,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC;QAC5C,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,YAAY,CAAC,IAAY,EAAE,QAAuB;IACzD,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,GAAG,QAAQ,CAAC;IACrC,MAAM,KAAK,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,iBAAiB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IAC7C,MAAM,KAAK,GACT,MAAM,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,EAAE,aAAa,CAAC,MAAM,CAAC,CAAC,CAAC;IAE5F,MAAM,QAAQ,GAAG,uBAAuB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IAEzD,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,KAAK,MAAM,MAAM,IAAI,MAAM,EAAE,CAAC;QAC5B,IAAI,CAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,CAAC;YAAE,SAAS;QAEjD,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,SAAS;YAChB,KAAK,EAAE,MAAM;YACb,MAAM,EAAE,aAAa,MAAM,uEAAuE;SACnG,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,SAAS,EAAE,QAAQ,CAAC,MAAM,KAAK,CAAC;QAChC,QAAQ,EAAE,aAAa;QACvB,OAAO,EAAE,MAAM,CAAC,MAAM;QACtB,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,SAAS,KAAK,CAAC,IAAY,EAAE,QAAuB;IAClD,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,QAAQ,CAAC,SAAS,CAAC,CAAC;IAC3D,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC7C,OAAO,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,IAAI,OAAO,CAAC,SAAS,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;AACnF,CAAC;AAED,MAAM,UAAU,MAAM,CAAC,IAAY,EAAE,KAAY;IAC/C,OAAO,KAAK,CAAC,IAAI,EAAE,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC;AACvC,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,MAA8B,EAAE,KAAY;IACvE,MAAM,QAAQ,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAElC,MAAM,OAAO,GAAkB,EAAE,CAAC;IAClC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,YAAY,GAAG,IAAI,CAAC;IACxB,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QACxC,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QACjC,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;QACrC,IAAI,CAAC,MAAM,CAAC,SAAS;YAAE,YAAY,GAAG,KAAK,CAAC;QAC5C,OAAO,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC;QAChC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnB,CAAC;IAED,MAAM,YAAY,GAAG,YAAY,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC;IAC7D,OAAO,EAAE,SAAS,EAAE,YAAY,IAAI,YAAY,CAAC,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,YAAY,EAAE,CAAC;AAC9F,CAAC"}

package/package.json

@@ -1,14 +1,17 @@
 {
   "name": "@rudra-js/attested",
-  "version": "0.7.4",
-  "description": "Checks model-written product copy against the facts a shop stands behind",
+  "version": "0.7.6",
+  "description": "From rudra-js: checks model-written product copy against the facts a shop stands behind",
   "keywords": [
+    "rudra-js",
+    "product-recommendations",
     "llm",
     "hallucination",
     "grounding",
     "verification",
     "ecommerce",
-    "guardrails"
+    "guardrails",
+    "typescript"
   ],
   "license": "MIT",
   "author": "Clive Dsouza",

package/src/index.ts

@@ -10,4 +10,4 @@ export {
   type VerifyResult,
 } from './verify.js';
 
-export { BANNED_PHRASES } from './phrases.js';
+export { BANNED_PHRASES, foldLookalikes } from './phrases.js';

package/src/numerals.ts

@@ -1,21 +1,17 @@
 import { stripInvisible, stripMarks } from './hidden.js';
 
 const AMBIGUOUS = '.,';
-
 const DECIMAL_ONLY = '\u066b';
-
 const GROUPING_ONLY = '\u066c\u0020\u00a0\u2007\u2008\u2009\u202f\u0027\u2019';
-
 const SEPARATORS = AMBIGUOUS + DECIMAL_ONLY + GROUPING_ONLY;
 
 const RUN = `\\p{Nd}+(?:[${SEPARATORS}]\\p{Nd}+)*`;
-
 const MAGNITUDE = '(?:[十百千万萬億亿兆]|[kKMB](?!\\p{L}))';
-
 const OTHER_NUMERAL = '[\\p{No}\\p{Nl}]+';
 
 const SCAN = new RegExp(`${RUN}${MAGNITUDE}|${RUN}|${OTHER_NUMERAL}`, 'gu');
-const LOOSE = new RegExp(`[${GROUPING_ONLY}]`, 'u');
+const GROUPING_MARK = new RegExp(`[${GROUPING_ONLY}]`, 'u');
+const SEPARATOR = new RegExp(`[${SEPARATORS}]`, 'gu');
 
 const IS_DIGIT = /^\p{Nd}$/u;
 const STARTS_DIGIT = /^\p{Nd}/u;
@@ -23,25 +19,24 @@ const ENDS_DIGIT = /\p{Nd}$/u;
 
 export interface Numeral {
   token: string;
-
   forms: string[];
-
   kind: 'digits' | 'magnitude' | 'other-numeral';
 }
 
-const zeros = new Map<number, number>();
+const runStartOf = new Map<number, number>();
 
 function digitValue(char: string): number {
   const code = char.codePointAt(0) ?? 0;
 
-  let zero = zeros.get(code);
-  if (zero === undefined) {
-    zero = code;
-    while (zero > 0 && IS_DIGIT.test(String.fromCodePoint(zero - 1))) zero -= 1;
-    zeros.set(code, zero);
+  let runStart = runStartOf.get(code);
+  if (runStart === undefined) {
+    runStart = code;
+    while (runStart > 0 && IS_DIGIT.test(String.fromCodePoint(runStart - 1))) runStart -= 1;
+    runStartOf.set(code, runStart);
   }
 
-  return (code - zero) % 10;
+  const offsetInRun = code - runStart;
+  return offsetInRun % 10;
 }
 
 function toAscii(token: string): string {
@@ -66,119 +61,78 @@ function canonical(whole: string, fraction: string): string {
 
 function isGrouping(groups: string[]): boolean {
   if (groups.length === 1) return true;
-  if ((groups[groups.length - 1] ?? '').length !== 3) return false;
-
-  let three = true;
-  let two = true;
-  for (let index = 1; index < groups.length - 1; index += 1) {
-    const size = (groups[index] ?? '').length;
-    if (size !== 3) three = false;
-    if (size !== 2) two = false;
-  }
+  if ((groups.at(-1) ?? '').length !== 3) return false;
 
   const lead = (groups[0] ?? '').length;
-  if (three) return lead >= 1 && lead <= 3;
-
-  if (two) return lead >= 1 && lead <= 2;
+  const middle = groups.slice(1, -1);
+  if (middle.every((group) => group.length === 3)) return lead >= 1 && lead <= 3;
+  if (middle.every((group) => group.length === 2)) return lead >= 1 && lead <= 2;
   return false;
 }
 
-function allGrouping(separators: string[]): boolean {
-  for (const separator of separators) {
-    if (separator !== separators[0]) return false;
-    if (DECIMAL_ONLY.includes(separator)) return false;
+function sameGroupingMark(marks: string[]): boolean {
+  for (const mark of marks) {
+    if (mark !== marks[0] || mark === DECIMAL_ONLY) return false;
   }
   return true;
 }
 
 function formsOf(token: string): string[] {
   const ascii = toAscii(token);
-
-  const groups: string[] = [];
-  const separators: string[] = [];
-  let current = '';
-  for (const char of ascii) {
-    if (SEPARATORS.includes(char)) {
-      groups.push(current);
-      separators.push(char);
-      current = '';
-    } else {
-      current += char;
-    }
-  }
-  groups.push(current);
-
-  if (separators.length === 0) return [canonical(groups[0] ?? '', '')];
+  const groups = ascii.split(SEPARATOR);
+  const marks = ascii.match(SEPARATOR) ?? [];
+  if (marks.length === 0) return [canonical(ascii, '')];
 
   const forms: string[] = [];
-  if (allGrouping(separators) && isGrouping(groups)) forms.push(canonical(groups.join(''), ''));
+  const grouped = sameGroupingMark(marks) && isGrouping(groups);
+  if (grouped) forms.push(canonical(groups.join(''), ''));
 
   const whole = groups.slice(0, -1);
-  const leading = separators.slice(0, -1);
-  const point = separators[separators.length - 1] ?? '';
-  const fraction = groups[groups.length - 1] ?? '';
-
-  const groupingWins = forms.length > 0 && leading.length === 0 && fraction.length === 3;
-
-  if (
-    !groupingWins &&
-    !GROUPING_ONLY.includes(point) &&
-    allGrouping(leading) &&
-    leading[0] !== point &&
-    isGrouping(whole)
-  ) {
-    forms.push(canonical(whole.join(''), fraction));
-  }
+  const fraction = groups.at(-1) ?? '';
+  const groupingMarks = marks.slice(0, -1);
+  const decimalMark = marks.at(-1) ?? '';
+
+  if (grouped && groupingMarks.length === 0 && fraction.length === 3) return forms;
+  if (GROUPING_ONLY.includes(decimalMark)) return forms;
+  if (!sameGroupingMark(groupingMarks)) return forms;
+  if (groupingMarks[0] === decimalMark) return forms;
+  if (!isGrouping(whole)) return forms;
 
+  forms.push(canonical(whole.join(''), fraction));
   return forms;
 }
 
-function collect(found: Numeral[], token: string): void {
-  if (!STARTS_DIGIT.test(token)) {
-    found.push({ token, forms: [], kind: 'other-numeral' });
-    return;
-  }
-  if (!ENDS_DIGIT.test(token)) {
-    found.push({ token, forms: [], kind: 'magnitude' });
-    return;
-  }
+function numeralsOf(token: string): Numeral[] {
+  if (!STARTS_DIGIT.test(token)) return [{ token, forms: [], kind: 'other-numeral' }];
+  if (!ENDS_DIGIT.test(token)) return [{ token, forms: [], kind: 'magnitude' }];
 
   const forms = formsOf(token);
-  if (forms.length > 0) {
-    found.push({ token, forms, kind: 'digits' });
-    return;
-  }
+  if (forms.length > 0) return [{ token, forms, kind: 'digits' }];
+  if (!GROUPING_MARK.test(token)) return [{ token, forms: [toAscii(token)], kind: 'digits' }];
 
-  if (LOOSE.test(token)) {
-    for (const part of token.split(LOOSE)) {
-      if (part.length > 0) collect(found, part);
-    }
-    return;
-  }
-
-  found.push({ token, forms: [toAscii(token)], kind: 'digits' });
+  return token
+    .split(GROUPING_MARK)
+    .filter((part) => part.length > 0)
+    .flatMap((part) => numeralsOf(part));
 }
 
 export function numeralsIn(text: string): Numeral[] {
-  const found: Numeral[] = [];
-  for (const match of stripMarks(stripInvisible(text)).matchAll(SCAN)) collect(found, match[0]);
-  return found;
+  const visible = stripMarks(stripInvisible(text));
+  return [...visible.matchAll(SCAN)].flatMap((match) => numeralsOf(match[0]));
 }
 
 const BARE_EXPONENT = /^[+-]?\d+(?:\.\d+)?[eE][+-]?\d+$/;
-
 const DIGIT_LIMIT = 1000;
 
 function positional(written: string): string {
-  const marker = written.search(/[eE]/);
-  if (marker < 0) return written;
-
-  const power = Number(written.slice(marker + 1));
+  const exponentAt = written.search(/[eE]/);
+  if (exponentAt < 0) return written;
 
-  const body = written.slice(0, marker).replace(/^[+-]/, '');
-  const point = body.indexOf('.');
-  const digits = point < 0 ? body : body.slice(0, point) + body.slice(point + 1);
-  const place = (point < 0 ? body.length : point) + power;
+  const power = Number(written.slice(exponentAt + 1));
+  const mantissa = written.slice(0, exponentAt).replace(/^[+-]/, '');
+  const dot = mantissa.indexOf('.');
+  const digits = mantissa.replace('.', '');
+  const place = (dot < 0 ? mantissa.length : dot) + power;
 
   if (place <= 0) {
     if (digits.length - place > DIGIT_LIMIT) return '';
@@ -191,7 +145,7 @@ function positional(written: string): string {
   return `${digits.slice(0, place)}.${digits.slice(place)}`;
 }
 
-function digitsOf(value: string | number | bigint): string {
+function textOf(value: string | number | bigint): string {
   if (typeof value === 'number') return positional(String(value));
   if (typeof value !== 'string') return String(value);
 
@@ -202,7 +156,7 @@ function digitsOf(value: string | number | bigint): string {
 export function supportedValues(values: readonly (string | number | bigint)[]): Set<string> {
   const supported = new Set<string>();
   for (const value of values) {
-    for (const numeral of numeralsIn(digitsOf(value))) {
+    for (const numeral of numeralsIn(textOf(value))) {
       for (const form of numeral.forms) supported.add(form);
     }
   }

package/src/phrases.ts

@@ -158,7 +158,6 @@ const CONFUSABLES: Record<string, string> = {
   ѵ: 'v',
 };
 
-// Capitals that lowercasing reads wrong: Ν becomes ν, which reads as v.
 const CAPITALS: Record<string, string> = {
   Β: 'b',
   Ζ: 'z',
@@ -175,23 +174,28 @@ const CAPITALS: Record<string, string> = {
   Ԍ: 'g',
 };
 
+const BEFORE_NFKC: Record<string, string> = {
+  ϲ: 'c',
+  Ϲ: 'c',
+};
+
 const LINE_BREAK = /[\n\v\f\r\u0085\u2028\u2029]/;
 
-export function normalisePhrasing(text: string, byLook = false): string {
-  // NFKC would turn the lunate sigmas into ς and Σ, which the fold can't read as c.
-  const compatible = stripInvisible(text).replace(/[ϲϹ]/g, 'c').normalize('NFKC');
+function foldChars(text: string, table: Record<string, string>): string {
+  let folded = '';
+  for (const char of text) folded += table[char] ?? char;
+  return folded;
+}
 
-  let cased = compatible;
-  if (byLook) {
-    cased = '';
-    for (const char of compatible) cased += CAPITALS[char] ?? char;
-  }
+export function foldLookalikes(text: string, byLook = false): string {
+  const compatible = foldChars(stripInvisible(text), BEFORE_NFKC).normalize('NFKC');
+  const cased = byLook ? foldChars(compatible, CAPITALS) : compatible;
   const lowered = stripMarks(cased.toLowerCase().normalize('NFC'));
+  return foldChars(lowered, CONFUSABLES);
+}
 
-  let folded = '';
-  for (const char of lowered) folded += CONFUSABLES[char] ?? char;
-
-  return folded
+export function normalisePhrasing(text: string, byLook = false): string {
+  return foldLookalikes(text, byLook)
     .replace(/[\u2018\u2019]/g, "'")
     .replace(/[-_\u2010-\u2015]+/g, ' ')
     .replace(/\s+/g, (run) => (LINE_BREAK.test(run) ? '\n' : ' '))
@@ -206,37 +210,34 @@ function isGlued(char: string | undefined): boolean {
 
 export interface Span {
   start: number;
-
   end: number;
 }
 
 export interface Indexed {
   text: string;
-
-  haystack: string;
-
-  spots: number[];
+  squeezed: string;
+  offsets: number[];
 }
 
 export function indexPhrasing(text: string): Indexed {
-  const tight: string[] = [];
-  const spots: number[] = [];
+  const kept: string[] = [];
+  const offsets: number[] = [];
   for (let index = 0; index < text.length; index += 1) {
     const char = text[index] ?? '';
 
     if (char === ' ' || char === '\n') continue;
-    tight.push(char);
-    spots.push(index);
+    kept.push(char);
+    offsets.push(index);
   }
 
-  return { text, haystack: tight.join(''), spots };
+  return { text, squeezed: kept.join(''), offsets };
 }
 
 export function spansIn(indexed: Indexed, phrase: string): Span[] {
-  const { text, haystack, spots } = indexed;
+  const { text, squeezed, offsets } = indexed;
 
   const spans: Span[] = [];
-  const needle = phrase.replaceAll(' ', '').replaceAll('\n', '');
+  const needle = indexPhrasing(phrase).squeezed;
   if (needle.length === 0) return spans;
 
   const guardStart = ASCII_WORD.test(needle[0] ?? '');
@@ -244,18 +245,18 @@ export function spansIn(indexed: Indexed, phrase: string): Span[] {
 
   let from = 0;
   for (;;) {
-    const at = haystack.indexOf(needle, from);
+    const at = squeezed.indexOf(needle, from);
     if (at === -1) return spans;
 
-    const start = spots[at] ?? 0;
-    const end = spots[at + needle.length - 1] ?? 0;
+    const start = offsets[at] ?? 0;
+    const end = offsets[at + needle.length - 1] ?? 0;
     const next = text[end + 1];
 
     const plural = next === 's' && !isGlued(text[end + 2]);
 
-    const before = guardStart && isGlued(text[start - 1]);
-    const after = guardEnd && isGlued(next) && !plural;
-    if (!before && !after) spans.push({ start, end });
+    const gluedBefore = guardStart && isGlued(text[start - 1]);
+    const gluedAfter = guardEnd && isGlued(next) && !plural;
+    if (!gluedBefore && !gluedAfter) spans.push({ start, end });
 
     from = at + 1;
   }