@trazum/core 1.8.0

package/dist/rules.js

@@ -0,0 +1,279 @@
+import { EMPHASIS_PREFIXES, FILLER, HEDGES, INTENSIFIERS, POLITENESS, SELF_CHECK, SHOUTED_WORDS, VERBOSE_PHRASES, } from './phrases.js';
+import { jaccard, normalizeForCompare } from './similarity.js';
+import { EXAMPLE_FIELD_LINE } from './structure.js';
+/** Escapes a literal so it can go inside a regular expression. */
+function escapeRe(literal) {
+    return literal.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+}
+/**
+ * Builds a regex with accent-aware word boundaries.
+ * JavaScript's `\b` only knows ASCII, so "de" would match inside "ánde".
+ * Lookarounds over \p{L} avoid that.
+ */
+function phraseRegex(phrase, flags = 'giu') {
+    const body = escapeRe(phrase).replace(/\s+/g, '\\s+');
+    return new RegExp(`(?<![\\p{L}\\p{N}_])${body}(?![\\p{L}\\p{N}_])`, flags);
+}
+/** Cleans up what deleting a phrase leaves behind: spaces and orphan punctuation. */
+function tidyAfterRemoval(text) {
+    return (text
+        // Repeated spaces inside a line (leading indentation untouched).
+        .replace(/([^\n\S])[^\S\n]+/g, '$1')
+        // Comma or semicolon left with a space in front of it.
+        .replace(/[^\S\n]+([,;:.!?])/g, '$1')
+        // Commas duplicated by the removal.
+        .replace(/,(\s*,)+/g, ',')
+        // Punctuation left at the start of a line.
+        .replace(/^[^\S\n]*[,;:]\s*/gm, '')
+        // Spanish opening exclamation/question mark left with no content.
+        .replace(/[¡¿]+(?=[^\S\n]*(?:[!?.,;:]|$))/gm, '')
+        // Duplicated sentence-final punctuation when two fragments join ("answer.!").
+        .replace(/([.!?])[^\S\n]*[.!?]+/g, '$1')
+        // A line left holding nothing but punctuation.
+        .replace(/^[^\S\n]*[.,;:!?¡¿]+[^\S\n]*$/gm, '')
+        // A single leading space is removal residue, not indentation.
+        .replace(/^[^\S\n](?=\S)/gm, '')
+        // Trailing spaces.
+        .replace(/[^\S\n]+$/gm, '')
+        // Restore the capital when the removal left a sentence starting lowercase.
+        // Sentence starts: beginning of the text, after strong punctuation, and
+        // the start of a new paragraph.
+        .replace(/(^|[.!?]\s+|\n\n[^\S\n]*)(\p{Ll})/gu, (_m, pre, ch) => pre + ch.toUpperCase()));
+}
+/**
+ * Deletion regex: on top of the phrase, it swallows the commas that delimited
+ * it. Without this, removing an aside such as "if you don't mind," leaves
+ * "and, classify it".
+ */
+function dropRegex(phrase) {
+    const body = escapeRe(phrase).replace(/\s+/g, '\\s+');
+    // The boundary check goes AFTER the optional comma: in front of it, the "y"
+    // of "y, si no te importa," would fail the match and the comma would stay.
+    return new RegExp(`(?:,[^\\S\\n]*)?(?<![\\p{L}\\p{N}_])${body}[^\\S\\n]*,?(?![\\p{L}\\p{N}_])`, 'giu');
+}
+/** A rule that deletes a set of phrases. */
+function dropRule(id, level, phrases) {
+    // Longest first, so "thank you very much" wins over "thanks".
+    const sorted = [...phrases].sort((a, b) => b.length - a.length);
+    const regexes = sorted.map((p) => dropRegex(p));
+    return {
+        id,
+        level,
+        apply(text) {
+            let hits = 0;
+            let out = text;
+            for (const re of regexes) {
+                out = out.replace(re, () => {
+                    hits++;
+                    return '';
+                });
+            }
+            return { text: hits > 0 ? tidyAfterRemoval(out) : out, hits };
+        },
+    };
+}
+/** Keeps the original leading capital when substituting. */
+function matchCase(original, replacement) {
+    if (!replacement)
+        return replacement;
+    const firstChar = original[0];
+    if (firstChar && firstChar === firstChar.toUpperCase() && firstChar !== firstChar.toLowerCase()) {
+        return replacement[0].toUpperCase() + replacement.slice(1);
+    }
+    return replacement;
+}
+const whitespaceRule = {
+    id: 'whitespace',
+    level: 'safe',
+    apply(text) {
+        let hits = 0;
+        let out = text;
+        // The lookbehind is load-bearing, not decoration. Without it the engine
+        // restarts this match at every position inside a whitespace run, and when
+        // the run does not end the line it fails from each one — quadratic, and
+        // 17 seconds on a 100 KB line of spaces that the HTTP API happily accepts.
+        // Anchoring to the start of a run means each run is tried exactly once.
+        out = out.replace(/(?<![^\S\n])[^\S\n]+$/gm, () => {
+            hits++;
+            return '';
+        });
+        // Collapse repeated spaces only after the first non-space character.
+        out = out.replace(/(\S)[^\S\n]{2,}/g, (_m, ch) => {
+            hits++;
+            return `${ch} `;
+        });
+        out = out.replace(/\n{3,}/g, () => {
+            hits++;
+            return '\n\n';
+        });
+        out = out.replace(/\t/g, () => {
+            hits++;
+            return ' ';
+        });
+        return { text: out, hits };
+    },
+};
+const decorationRule = {
+    id: 'decoration',
+    level: 'safe',
+    apply(text) {
+        let hits = 0;
+        let out = text;
+        // A whole line of 8 or more identical decorative characters.
+        out = out.replace(/^[^\S\n]*([=\-_*~#])\1{7,}[^\S\n]*$/gm, () => {
+            hits++;
+            return '';
+        });
+        // Repeated exclamation or question marks, opening ones included.
+        out = out.replace(/([!?¡¿])\1{1,}/g, (_m, ch) => {
+            hits++;
+            return ch;
+        });
+        // Clean up the blank lines the deletion just left behind.
+        if (hits > 0)
+            out = out.replace(/\n{3,}/g, '\n\n');
+        return { text: out, hits };
+    },
+};
+const verbosePhrasesRule = {
+    id: 'verbose-phrases',
+    level: 'safe',
+    apply(text) {
+        let hits = 0;
+        let out = text;
+        // Longest first so the most specific substitution wins.
+        const entries = [...VERBOSE_PHRASES].sort((a, b) => b[0].length - a[0].length);
+        for (const [from, to] of entries) {
+            const re = phraseRegex(from);
+            out = out.replace(re, (match) => {
+                hits++;
+                return matchCase(match, to);
+            });
+        }
+        return { text: hits > 0 ? tidyAfterRemoval(out) : out, hits };
+    },
+};
+const politenessRule = dropRule('politeness', 'safe', POLITENESS);
+const fillerRule = dropRule('filler', 'safe', FILLER);
+const intensifiersRule = dropRule('intensifiers', 'aggressive', INTENSIFIERS);
+const hedgesRule = dropRule('hedges', 'aggressive', HEDGES);
+const selfCheckRule = dropRule('self-check', 'aggressive', SELF_CHECK);
+const emphasisRule = {
+    id: 'emphasis',
+    level: 'aggressive',
+    apply(text) {
+        let hits = 0;
+        let out = text;
+        for (const prefix of EMPHASIS_PREFIXES) {
+            const re = new RegExp(`^([^\\S\\n]*)${escapeRe(prefix)}\\s*:\\s*`, 'gm');
+            out = out.replace(re, (_m, indent) => {
+                hits++;
+                return indent;
+            });
+        }
+        for (const word of SHOUTED_WORDS) {
+            const re = phraseRegex(word, 'gu'); // case-sensitive: only the shouted form
+            out = out.replace(re, (match) => {
+                hits++;
+                return match.toLowerCase();
+            });
+        }
+        // Dropping "IMPORTANT:" leaves the sentence starting lowercase: recapitalise.
+        return { text: hits > 0 ? tidyAfterRemoval(out) : out, hits };
+    },
+};
+const duplicateLinesRule = {
+    id: 'duplicate-lines',
+    level: 'safe',
+    apply(text) {
+        const lines = text.split('\n');
+        const seen = new Set();
+        const kept = [];
+        let hits = 0;
+        for (const line of lines) {
+            const normalized = normalizeForCompare(line);
+            // A labelled example field is data, not repetition: two examples sharing
+            // an output line show that two inputs map to the same answer. Dropping
+            // the second leaves that example with no output.
+            if (normalized.length >= 25 && !EXAMPLE_FIELD_LINE.test(line)) {
+                if (seen.has(normalized)) {
+                    hits++;
+                    continue;
+                }
+                seen.add(normalized);
+            }
+            kept.push(line);
+        }
+        let out = kept.join('\n');
+        if (hits > 0)
+            out = out.replace(/\n{3,}/g, '\n\n');
+        return { text: out, hits };
+    },
+};
+const duplicateBlocksRule = {
+    id: 'duplicate-blocks',
+    level: 'safe',
+    apply(text) {
+        const blocks = text.split(/\n{2,}/);
+        const seen = new Set();
+        const kept = [];
+        let hits = 0;
+        for (const block of blocks) {
+            const normalized = normalizeForCompare(block);
+            if (normalized.length >= 40) {
+                if (seen.has(normalized)) {
+                    hits++;
+                    continue;
+                }
+                seen.add(normalized);
+            }
+            kept.push(block);
+        }
+        return { text: kept.join('\n\n'), hits };
+    },
+};
+const nearDuplicateBlocksRule = {
+    id: 'near-duplicate-blocks',
+    level: 'aggressive',
+    apply(text) {
+        const blocks = text.split(/\n{2,}/);
+        const keptNormalized = [];
+        const kept = [];
+        let hits = 0;
+        for (const block of blocks) {
+            const normalized = normalizeForCompare(block);
+            if (normalized.length >= 60) {
+                const isDuplicate = keptNormalized.some((prev) => jaccard(prev, normalized) >= 0.92);
+                if (isDuplicate) {
+                    hits++;
+                    continue;
+                }
+                keptNormalized.push(normalized);
+            }
+            kept.push(block);
+        }
+        return { text: kept.join('\n\n'), hits };
+    },
+};
+/** Every rule, in the order they must run. */
+export const RULES = [
+    // Whole-block deletions first, so the rest works over less text.
+    duplicateBlocksRule,
+    nearDuplicateBlocksRule,
+    duplicateLinesRule,
+    // Then the phrase-level rules.
+    verbosePhrasesRule,
+    politenessRule,
+    fillerRule,
+    hedgesRule,
+    intensifiersRule,
+    selfCheckRule,
+    emphasisRule,
+    // And finally typographic cleanup, which picks up what the others left.
+    decorationRule,
+    whitespaceRule,
+];
+export function getRule(id) {
+    return RULES.find((r) => r.id === id);
+}
+//# sourceMappingURL=rules.js.map

package/dist/rules.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"rules.js","sourceRoot":"","sources":["../src/rules.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,iBAAiB,EACjB,MAAM,EACN,MAAM,EACN,YAAY,EACZ,UAAU,EACV,UAAU,EACV,aAAa,EACb,eAAe,GAChB,MAAM,cAAc,CAAC;AACtB,OAAO,EAAE,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAC;AAC/D,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AAIpD,kEAAkE;AAClE,SAAS,QAAQ,CAAC,OAAe;IAC/B,OAAO,OAAO,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;AACxD,CAAC;AAED;;;;GAIG;AACH,SAAS,WAAW,CAAC,MAAc,EAAE,KAAK,GAAG,KAAK;IAChD,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACtD,OAAO,IAAI,MAAM,CAAC,uBAAuB,IAAI,qBAAqB,EAAE,KAAK,CAAC,CAAC;AAC7E,CAAC;AAED,qFAAqF;AACrF,SAAS,gBAAgB,CAAC,IAAY;IACpC,OAAO,CACL,IAAI;QACF,iEAAiE;SAChE,OAAO,CAAC,oBAAoB,EAAE,IAAI,CAAC;QACpC,uDAAuD;SACtD,OAAO,CAAC,qBAAqB,EAAE,IAAI,CAAC;QACrC,oCAAoC;SACnC,OAAO,CAAC,WAAW,EAAE,GAAG,CAAC;QAC1B,2CAA2C;SAC1C,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC;QACnC,kEAAkE;SACjE,OAAO,CAAC,mCAAmC,EAAE,EAAE,CAAC;QACjD,8EAA8E;SAC7E,OAAO,CAAC,wBAAwB,EAAE,IAAI,CAAC;QACxC,+CAA+C;SAC9C,OAAO,CAAC,iCAAiC,EAAE,EAAE,CAAC;QAC/C,8DAA8D;SAC7D,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC;QAChC,mBAAmB;SAClB,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC;QAC3B,2EAA2E;QAC3E,wEAAwE;QACxE,gCAAgC;SAC/B,OAAO,CACN,qCAAqC,EACrC,CAAC,EAAE,EAAE,GAAW,EAAE,EAAU,EAAE,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,WAAW,EAAE,CACxD,CACJ,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,SAAS,SAAS,CAAC,MAAc;IAC/B,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACtD,4EAA4E;IAC5E,2EAA2E;IAC3E,OAAO,IAAI,MAAM,CACf,uCAAuC,IAAI,iCAAiC,EAC5E,KAAK,CACN,CAAC;AACJ,CAAC;AAED,4CAA4C;AAC5C,SAAS,QAAQ,CAAC,EAAU,EAAE,KAAoB,EAAE,OAA0B;IAC5E,8DAA8D;IAC9D,MAAM,MAAM,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC;IAChE,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;IAEhD,OAAO;QACL,EAAE;QACF,KAAK;QACL,KAAK,CAAC,IAAI;YACR,IAAI,IAAI,GAAG,CAAC,CAAC;YACb,IAAI,GAAG,GAAG,IAAI,CAAC;YACf,KAAK,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;gBACzB,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,EAAE,EAAE,GAAG,EAAE;oBACzB,IAAI,EAAE,CAAC;oBACP,OAAO,EAAE,CAAC;gBACZ,CAAC,CAAC,CAAC;YACL,CAAC;YACD,OAAO,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;QAChE,CAAC;KACF,CAAC;AACJ,CAAC;AAED,4DAA4D;AAC5D,SAAS,SAAS,CAAC,QAAgB,EAAE,WAAmB;IACtD,IAAI,CAAC,WAAW;QAAE,OAAO,WAAW,CAAC;IACrC,MAAM,SAAS,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAC9B,IAAI,SAAS,IAAI,SAAS,KAAK,SAAS,CAAC,WAAW,EAAE,IAAI,SAAS,KAAK,SAAS,CAAC,WAAW,EAAE,EAAE,CAAC;QAChG,OAAO,WAAW,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAC9D,CAAC;IACD,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,MAAM,cAAc,GAAS;IAC3B,EAAE,EAAE,YAAY;IAChB,KAAK,EAAE,MAAM;IACb,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,IAAI,GAAG,GAAG,IAAI,CAAC;QAEf,wEAAwE;QACxE,0EAA0E;QAC1E,wEAAwE;QACxE,2EAA2E;QAC3E,wEAAwE;QACxE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,yBAAyB,EAAE,GAAG,EAAE;YAChD,IAAI,EAAE,CAAC;YACP,OAAO,EAAE,CAAC;QACZ,CAAC,CAAC,CAAC;QACH,qEAAqE;QACrE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,kBAAkB,EAAE,CAAC,EAAE,EAAE,EAAU,EAAE,EAAE;YACvD,IAAI,EAAE,CAAC;YACP,OAAO,GAAG,EAAE,GAAG,CAAC;QAClB,CAAC,CAAC,CAAC;QACH,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,EAAE;YAChC,IAAI,EAAE,CAAC;YACP,OAAO,MAAM,CAAC;QAChB,CAAC,CAAC,CAAC;QACH,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,EAAE;YAC5B,IAAI,EAAE,CAAC;YACP,OAAO,GAAG,CAAC;QACb,CAAC,CAAC,CAAC;QAEH,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;IAC7B,CAAC;CACF,CAAC;AAEF,MAAM,cAAc,GAAS;IAC3B,EAAE,EAAE,YAAY;IAChB,KAAK,EAAE,MAAM;IACb,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,IAAI,GAAG,GAAG,IAAI,CAAC;QAEf,6DAA6D;QAC7D,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,uCAAuC,EAAE,GAAG,EAAE;YAC9D,IAAI,EAAE,CAAC;YACP,OAAO,EAAE,CAAC;QACZ,CAAC,CAAC,CAAC;QACH,iEAAiE;QACjE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,iBAAiB,EAAE,CAAC,EAAE,EAAE,EAAU,EAAE,EAAE;YACtD,IAAI,EAAE,CAAC;YACP,OAAO,EAAE,CAAC;QACZ,CAAC,CAAC,CAAC;QACH,0DAA0D;QAC1D,IAAI,IAAI,GAAG,CAAC;YAAE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;QAEnD,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;IAC7B,CAAC;CACF,CAAC;AAEF,MAAM,kBAAkB,GAAS;IAC/B,EAAE,EAAE,iBAAiB;IACrB,KAAK,EAAE,MAAM;IACb,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,IAAI,GAAG,GAAG,IAAI,CAAC;QACf,wDAAwD;QACxD,MAAM,OAAO,GAAG,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QAC/E,KAAK,MAAM,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,OAAO,EAAE,CAAC;YACjC,MAAM,EAAE,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;YAC7B,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,KAAK,EAAE,EAAE;gBAC9B,IAAI,EAAE,CAAC;gBACP,OAAO,SAAS,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;YAC9B,CAAC,CAAC,CAAC;QACL,CAAC;QACD,OAAO,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;IAChE,CAAC;CACF,CAAC;AAEF,MAAM,cAAc,GAAG,QAAQ,CAAC,YAAY,EAAE,MAAM,EAAE,UAAU,CAAC,CAAC;AAClE,MAAM,UAAU,GAAG,QAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;AACtD,MAAM,gBAAgB,GAAG,QAAQ,CAAC,cAAc,EAAE,YAAY,EAAE,YAAY,CAAC,CAAC;AAC9E,MAAM,UAAU,GAAG,QAAQ,CAAC,QAAQ,EAAE,YAAY,EAAE,MAAM,CAAC,CAAC;AAC5D,MAAM,aAAa,GAAG,QAAQ,CAAC,YAAY,EAAE,YAAY,EAAE,UAAU,CAAC,CAAC;AAEvE,MAAM,YAAY,GAAS;IACzB,EAAE,EAAE,UAAU;IACd,KAAK,EAAE,YAAY;IACnB,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,IAAI,GAAG,GAAG,IAAI,CAAC;QAEf,KAAK,MAAM,MAAM,IAAI,iBAAiB,EAAE,CAAC;YACvC,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,gBAAgB,QAAQ,CAAC,MAAM,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC;YACzE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,MAAc,EAAE,EAAE;gBAC3C,IAAI,EAAE,CAAC;gBACP,OAAO,MAAM,CAAC;YAChB,CAAC,CAAC,CAAC;QACL,CAAC;QAED,KAAK,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;YACjC,MAAM,EAAE,GAAG,WAAW,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,wCAAwC;YAC5E,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,KAAK,EAAE,EAAE;gBAC9B,IAAI,EAAE,CAAC;gBACP,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC;YAC7B,CAAC,CAAC,CAAC;QACL,CAAC;QAED,8EAA8E;QAC9E,OAAO,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;IAChE,CAAC;CACF,CAAC;AAEF,MAAM,kBAAkB,GAAS;IAC/B,EAAE,EAAE,iBAAiB;IACrB,KAAK,EAAE,MAAM;IACb,KAAK,CAAC,IAAI;QACR,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC/B,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,MAAM,IAAI,GAAa,EAAE,CAAC;QAC1B,IAAI,IAAI,GAAG,CAAC,CAAC;QAEb,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,UAAU,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;YAC7C,yEAAyE;YACzE,uEAAuE;YACvE,iDAAiD;YACjD,IAAI,UAAU,CAAC,MAAM,IAAI,EAAE,IAAI,CAAC,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC9D,IAAI,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;oBACzB,IAAI,EAAE,CAAC;oBACP,SAAS;gBACX,CAAC;gBACD,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;YACvB,CAAC;YACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClB,CAAC;QAED,IAAI,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC1B,IAAI,IAAI,GAAG,CAAC;YAAE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;QACnD,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;IAC7B,CAAC;CACF,CAAC;AAEF,MAAM,mBAAmB,GAAS;IAChC,EAAE,EAAE,kBAAkB;IACtB,KAAK,EAAE,MAAM;IACb,KAAK,CAAC,IAAI;QACR,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QACpC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,MAAM,IAAI,GAAa,EAAE,CAAC;QAC1B,IAAI,IAAI,GAAG,CAAC,CAAC;QAEb,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,CAAC,CAAC;YAC9C,IAAI,UAAU,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;gBAC5B,IAAI,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;oBACzB,IAAI,EAAE,CAAC;oBACP,SAAS;gBACX,CAAC;gBACD,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;YACvB,CAAC;YACD,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACnB,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,CAAC;IAC3C,CAAC;CACF,CAAC;AAEF,MAAM,uBAAuB,GAAS;IACpC,EAAE,EAAE,uBAAuB;IAC3B,KAAK,EAAE,YAAY;IACnB,KAAK,CAAC,IAAI;QACR,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QACpC,MAAM,cAAc,GAAa,EAAE,CAAC;QACpC,MAAM,IAAI,GAAa,EAAE,CAAC;QAC1B,IAAI,IAAI,GAAG,CAAC,CAAC;QAEb,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,CAAC,CAAC;YAC9C,IAAI,UAAU,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;gBAC5B,MAAM,WAAW,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,EAAE,UAAU,CAAC,IAAI,IAAI,CAAC,CAAC;gBACrF,IAAI,WAAW,EAAE,CAAC;oBAChB,IAAI,EAAE,CAAC;oBACP,SAAS;gBACX,CAAC;gBACD,cAAc,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YAClC,CAAC;YACD,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACnB,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,CAAC;IAC3C,CAAC;CACF,CAAC;AAEF,8CAA8C;AAC9C,MAAM,CAAC,MAAM,KAAK,GAAoB;IACpC,iEAAiE;IACjE,mBAAmB;IACnB,uBAAuB;IACvB,kBAAkB;IAClB,+BAA+B;IAC/B,kBAAkB;IAClB,cAAc;IACd,UAAU;IACV,UAAU;IACV,gBAAgB;IAChB,aAAa;IACb,YAAY;IACZ,wEAAwE;IACxE,cAAc;IACd,cAAc;CACf,CAAC;AAEF,MAAM,UAAU,OAAO,CAAC,EAAU;IAChC,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC;AACxC,CAAC"}

package/dist/savings.d.ts

@@ -0,0 +1,36 @@
+import type { PricingCatalogue } from './pricing.js';
+import type { CostBreakdown, SavingsReport, UsageProfile } from './types.js';
+/**
+ * Cost of a single call.
+ *
+ * `batchDiscount` defaults to Anthropic's 50% rather than being read from the
+ * model, because this function takes prices rather than a model. Callers that
+ * have the model — which is all of them inside this package — pass its own,
+ * and a provider with no batch API passes 1.
+ */
+export declare function costOfCall(inputTokens: number, outputTokens: number, inputPerMTok: number, outputPerMTok: number, batch: boolean, batchDiscount?: number): CostBreakdown;
+/**
+ * Compares cost before and after optimising.
+ *
+ * Important: output tokens are held constant. A shorter prompt often produces
+ * somewhat shorter answers, but that depends on the task and cannot be
+ * promised, so the saving reported here comes exclusively from input tokens.
+ */
+export declare function computeSavings(tokensBefore: number, tokensAfter: number, usage: UsageProfile, on?: Date, pricing?: PricingCatalogue): SavingsReport;
+/**
+ * Formats a USD amount with precision suited to its magnitude.
+ * Currency formatting stays in `en-US` on purpose: these are US dollar prices
+ * from a US price list, and showing them the same way everywhere avoids
+ * confusion when a report is shared across locales.
+ */
+export declare function formatUsd(value: number): string;
+/**
+ * Formats a USD amount that can legitimately be negative, with an explicit
+ * sign.
+ *
+ * `formatUsd` renders a negative as `$-30.80`, which reads as a typo. In a
+ * comparison every cost line can go either way and the sign carries the whole
+ * meaning, so it goes in front of the currency where a reader expects it.
+ */
+export declare function formatSignedUsd(value: number): string;
+//# sourceMappingURL=savings.d.ts.map

package/dist/savings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"savings.d.ts","sourceRoot":"","sources":["../src/savings.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAE7E;;;;;;;GAOG;AACH,wBAAgB,UAAU,CACxB,WAAW,EAAE,MAAM,EACnB,YAAY,EAAE,MAAM,EACpB,YAAY,EAAE,MAAM,EACpB,aAAa,EAAE,MAAM,EACrB,KAAK,EAAE,OAAO,EACd,aAAa,GAAE,MAA+B,GAC7C,aAAa,CAKf;AAUD;;;;;;GAMG;AACH,wBAAgB,cAAc,CAC5B,YAAY,EAAE,MAAM,EACpB,WAAW,EAAE,MAAM,EACnB,KAAK,EAAE,YAAY,EACnB,EAAE,GAAE,IAAiB,EACrB,OAAO,GAAE,gBAAoC,GAC5C,aAAa,CAsCf;AAED;;;;;GAKG;AACH,wBAAgB,SAAS,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAO/C;AAED;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAIrD"}

package/dist/savings.js

@@ -0,0 +1,83 @@
+import { BUNDLED_CATALOGUE, COST_MULTIPLIERS, effectivePricing, modelFrom, multipliersFor, } from './pricing.js';
+/**
+ * Cost of a single call.
+ *
+ * `batchDiscount` defaults to Anthropic's 50% rather than being read from the
+ * model, because this function takes prices rather than a model. Callers that
+ * have the model — which is all of them inside this package — pass its own,
+ * and a provider with no batch API passes 1.
+ */
+export function costOfCall(inputTokens, outputTokens, inputPerMTok, outputPerMTok, batch, batchDiscount = COST_MULTIPLIERS.batch) {
+    const discount = batch ? batchDiscount : 1;
+    const inputUsd = (inputTokens / 1_000_000) * inputPerMTok * discount;
+    const outputUsd = (outputTokens / 1_000_000) * outputPerMTok * discount;
+    return { inputUsd, outputUsd, totalUsd: inputUsd + outputUsd };
+}
+function scale(cost, factor) {
+    return {
+        inputUsd: cost.inputUsd * factor,
+        outputUsd: cost.outputUsd * factor,
+        totalUsd: cost.totalUsd * factor,
+    };
+}
+/**
+ * Compares cost before and after optimising.
+ *
+ * Important: output tokens are held constant. A shorter prompt often produces
+ * somewhat shorter answers, but that depends on the task and cannot be
+ * promised, so the saving reported here comes exclusively from input tokens.
+ */
+export function computeSavings(tokensBefore, tokensAfter, usage, on = new Date(), pricing = BUNDLED_CATALOGUE) {
+    const model = modelFrom(pricing, usage.model);
+    const { inputPerMTok, outputPerMTok, promoApplied } = effectivePricing(model, on);
+    // A provider with no batch API gets no discount even when the caller ticked
+    // the box: `batchEligible` describes the work, not what the provider sells.
+    const batchDiscount = multipliersFor(model).batch ?? 1;
+    const before = costOfCall(tokensBefore, usage.avgOutputTokens, inputPerMTok, outputPerMTok, usage.batchEligible, batchDiscount);
+    const after = costOfCall(tokensAfter, usage.avgOutputTokens, inputPerMTok, outputPerMTok, usage.batchEligible, batchDiscount);
+    const monthBefore = scale(before, usage.callsPerMonth);
+    const monthAfter = scale(after, usage.callsPerMonth);
+    const monthlySavingsUsd = monthBefore.totalUsd - monthAfter.totalUsd;
+    return {
+        model: model.id,
+        modelDisplayName: model.displayName,
+        promoApplied,
+        perCall: { before, after },
+        perMonth: { before: monthBefore, after: monthAfter },
+        monthlySavingsUsd,
+        monthlySavingsPct: monthBefore.totalUsd > 0 ? (monthlySavingsUsd / monthBefore.totalUsd) * 100 : 0,
+    };
+}
+/**
+ * Formats a USD amount with precision suited to its magnitude.
+ * Currency formatting stays in `en-US` on purpose: these are US dollar prices
+ * from a US price list, and showing them the same way everywhere avoids
+ * confusion when a report is shared across locales.
+ */
+export function formatUsd(value) {
+    if (value === 0)
+        return '$0';
+    const abs = Math.abs(value);
+    if (abs < 0.01)
+        return `$${value.toFixed(5)}`;
+    if (abs < 1)
+        return `$${value.toFixed(4)}`;
+    if (abs < 1000)
+        return `$${value.toFixed(2)}`;
+    return `$${value.toLocaleString('en-US', { maximumFractionDigits: 0 })}`;
+}
+/**
+ * Formats a USD amount that can legitimately be negative, with an explicit
+ * sign.
+ *
+ * `formatUsd` renders a negative as `$-30.80`, which reads as a typo. In a
+ * comparison every cost line can go either way and the sign carries the whole
+ * meaning, so it goes in front of the currency where a reader expects it.
+ */
+export function formatSignedUsd(value) {
+    if (value === 0)
+        return '$0';
+    const rendered = formatUsd(Math.abs(value));
+    return value > 0 ? `+${rendered}` : `-${rendered}`;
+}
+//# sourceMappingURL=savings.js.map

package/dist/savings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"savings.js","sourceRoot":"","sources":["../src/savings.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,iBAAiB,EACjB,gBAAgB,EAChB,gBAAgB,EAChB,SAAS,EACT,cAAc,GACf,MAAM,cAAc,CAAC;AAItB;;;;;;;GAOG;AACH,MAAM,UAAU,UAAU,CACxB,WAAmB,EACnB,YAAoB,EACpB,YAAoB,EACpB,aAAqB,EACrB,KAAc,EACd,aAAa,GAAW,gBAAgB,CAAC,KAAK;IAE9C,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC;IAC3C,MAAM,QAAQ,GAAG,CAAC,WAAW,GAAG,SAAS,CAAC,GAAG,YAAY,GAAG,QAAQ,CAAC;IACrE,MAAM,SAAS,GAAG,CAAC,YAAY,GAAG,SAAS,CAAC,GAAG,aAAa,GAAG,QAAQ,CAAC;IACxE,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,QAAQ,GAAG,SAAS,EAAE,CAAC;AACjE,CAAC;AAED,SAAS,KAAK,CAAC,IAAmB,EAAE,MAAc;IAChD,OAAO;QACL,QAAQ,EAAE,IAAI,CAAC,QAAQ,GAAG,MAAM;QAChC,SAAS,EAAE,IAAI,CAAC,SAAS,GAAG,MAAM;QAClC,QAAQ,EAAE,IAAI,CAAC,QAAQ,GAAG,MAAM;KACjC,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,cAAc,CAC5B,YAAoB,EACpB,WAAmB,EACnB,KAAmB,EACnB,EAAE,GAAS,IAAI,IAAI,EAAE,EACrB,OAAO,GAAqB,iBAAiB;IAE7C,MAAM,KAAK,GAAG,SAAS,CAAC,OAAO,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IAC9C,MAAM,EAAE,YAAY,EAAE,aAAa,EAAE,YAAY,EAAE,GAAG,gBAAgB,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IAClF,4EAA4E;IAC5E,4EAA4E;IAC5E,MAAM,aAAa,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC;IAEvD,MAAM,MAAM,GAAG,UAAU,CACvB,YAAY,EACZ,KAAK,CAAC,eAAe,EACrB,YAAY,EACZ,aAAa,EACb,KAAK,CAAC,aAAa,EACnB,aAAa,CACd,CAAC;IACF,MAAM,KAAK,GAAG,UAAU,CACtB,WAAW,EACX,KAAK,CAAC,eAAe,EACrB,YAAY,EACZ,aAAa,EACb,KAAK,CAAC,aAAa,EACnB,aAAa,CACd,CAAC;IAEF,MAAM,WAAW,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC;IACvD,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC;IACrD,MAAM,iBAAiB,GAAG,WAAW,CAAC,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC;IAErE,OAAO;QACL,KAAK,EAAE,KAAK,CAAC,EAAE;QACf,gBAAgB,EAAE,KAAK,CAAC,WAAW;QACnC,YAAY;QACZ,OAAO,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE;QAC1B,QAAQ,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,UAAU,EAAE;QACpD,iBAAiB;QACjB,iBAAiB,EACf,WAAW,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,iBAAiB,GAAG,WAAW,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;KAClF,CAAC;AACJ,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,SAAS,CAAC,KAAa;IACrC,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAC7B,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAC5B,IAAI,GAAG,GAAG,IAAI;QAAE,OAAO,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;IAC9C,IAAI,GAAG,GAAG,CAAC;QAAE,OAAO,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;IAC3C,IAAI,GAAG,GAAG,IAAI;QAAE,OAAO,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;IAC9C,OAAO,IAAI,KAAK,CAAC,cAAc,CAAC,OAAO,EAAE,EAAE,qBAAqB,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC;AAC3E,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,eAAe,CAAC,KAAa;IAC3C,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAC7B,MAAM,QAAQ,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;IAC5C,OAAO,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,QAAQ,EAAE,CAAC,CAAC,CAAC,IAAI,QAAQ,EAAE,CAAC;AACrD,CAAC"}

package/dist/segment.d.ts

@@ -0,0 +1,8 @@
+import type { Segment } from './types.js';
+/** Splits the text into segments, marking what must not be modified. */
+export declare function segment(text: string): Segment[];
+/** Joins segments back into a single string. */
+export declare function join(segments: Segment[]): string;
+/** Returns only the text of the protected segments, in order. */
+export declare function protectedTexts(segments: Segment[]): string[];
+//# sourceMappingURL=segment.d.ts.map

package/dist/segment.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"segment.d.ts","sourceRoot":"","sources":["../src/segment.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAkB,OAAO,EAAE,MAAM,YAAY,CAAC;AA6C1D,wEAAwE;AACxE,wBAAgB,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,EAAE,CAiD/C;AAED,gDAAgD;AAChD,wBAAgB,IAAI,CAAC,QAAQ,EAAE,OAAO,EAAE,GAAG,MAAM,CAEhD;AAED,iEAAiE;AACjE,wBAAgB,cAAc,CAAC,QAAQ,EAAE,OAAO,EAAE,GAAG,MAAM,EAAE,CAE5D"}

package/dist/segment.js

@@ -0,0 +1,74 @@
+// Order matters: the first pattern to match at a position wins.
+const PATTERNS = [
+    // Fenced code blocks with ``` or ~~~ (unclosed blocks run to the end).
+    { kind: 'fenced-code', regex: /(?:```|~~~)[\s\S]*?(?:```|~~~|$)/g },
+    // Indented blocks of 4+ spaces are left to the rules: they are ambiguous in markdown.
+    { kind: 'url', regex: /\b(?:https?|ftp):\/\/[^\s<>"')\]]+/g, trimTrailing: /[.,;:!?]+$/ },
+    // Template placeholders: {{var}}, {var}, ${var}, {% tag %}, <<VAR>>, %(var)s
+    {
+        kind: 'placeholder',
+        regex: /\{\{[^{}]*\}\}|\{%[\s\S]*?%\}|\$\{[^{}]*\}|<<[A-Z0-9_]+>>|%\([^()]*\)[sdfr]|\{[A-Za-z_][A-Za-z0-9_.]*\}/g,
+    },
+    // Inline code spans.
+    { kind: 'inline-code', regex: /`[^`\n]+`/g },
+    // XML/HTML tags, heavily used to structure prompts.
+    { kind: 'xml-tag', regex: /<\/?[A-Za-z][\w:.-]*(?:\s[^<>]*?)?\/?>/g },
+];
+/** Splits the text into segments, marking what must not be modified. */
+export function segment(text) {
+    const matches = [];
+    for (const { kind, regex, trimTrailing } of PATTERNS) {
+        // Copy the regex so lastIndex is not shared between calls.
+        const re = new RegExp(regex.source, regex.flags);
+        let m;
+        while ((m = re.exec(text)) !== null) {
+            let matched = m[0];
+            if (matched.length === 0) {
+                re.lastIndex++;
+                continue;
+            }
+            if (trimTrailing)
+                matched = matched.replace(trimTrailing, '');
+            if (matched.length === 0)
+                continue;
+            matches.push({ start: m.index, end: m.index + matched.length, kind });
+        }
+    }
+    // Sort by start and, on a tie, longest span first.
+    matches.sort((a, b) => a.start - b.start || b.end - a.end);
+    // Drop overlaps, keeping the first match at each position.
+    const kept = [];
+    let cursor = 0;
+    for (const match of matches) {
+        if (match.start < cursor)
+            continue;
+        kept.push(match);
+        cursor = match.end;
+    }
+    const segments = [];
+    let pos = 0;
+    for (const match of kept) {
+        if (match.start > pos) {
+            segments.push({ kind: 'mutable', text: text.slice(pos, match.start) });
+        }
+        segments.push({
+            kind: 'protected',
+            protection: match.kind,
+            text: text.slice(match.start, match.end),
+        });
+        pos = match.end;
+    }
+    if (pos < text.length) {
+        segments.push({ kind: 'mutable', text: text.slice(pos) });
+    }
+    return segments;
+}
+/** Joins segments back into a single string. */
+export function join(segments) {
+    return segments.map((s) => s.text).join('');
+}
+/** Returns only the text of the protected segments, in order. */
+export function protectedTexts(segments) {
+    return segments.filter((s) => s.kind === 'protected').map((s) => s.text);
+}
+//# sourceMappingURL=segment.js.map

package/dist/segment.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"segment.js","sourceRoot":"","sources":["../src/segment.ts"],"names":[],"mappings":"AAqBA,gEAAgE;AAChE,MAAM,QAAQ,GAAiB;IAC7B,uEAAuE;IACvE,EAAE,IAAI,EAAE,aAAa,EAAE,KAAK,EAAE,mCAAmC,EAAE;IACnE,sFAAsF;IACtF,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,qCAAqC,EAAE,YAAY,EAAE,YAAY,EAAE;IACzF,6EAA6E;IAC7E;QACE,IAAI,EAAE,aAAa;QACnB,KAAK,EACH,0GAA0G;KAC7G;IACD,qBAAqB;IACrB,EAAE,IAAI,EAAE,aAAa,EAAE,KAAK,EAAE,YAAY,EAAE;IAC5C,oDAAoD;IACpD,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,yCAAyC,EAAE;CACtE,CAAC;AAQF,wEAAwE;AACxE,MAAM,UAAU,OAAO,CAAC,IAAY;IAClC,MAAM,OAAO,GAAY,EAAE,CAAC;IAE5B,KAAK,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,YAAY,EAAE,IAAI,QAAQ,EAAE,CAAC;QACrD,2DAA2D;QAC3D,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QACjD,IAAI,CAAyB,CAAC;QAC9B,OAAO,CAAC,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACpC,IAAI,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YACnB,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACzB,EAAE,CAAC,SAAS,EAAE,CAAC;gBACf,SAAS;YACX,CAAC;YACD,IAAI,YAAY;gBAAE,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC;YAC9D,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YACnC,OAAO,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;QACxE,CAAC;IACH,CAAC;IAED,mDAAmD;IACnD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IAE3D,2DAA2D;IAC3D,MAAM,IAAI,GAAY,EAAE,CAAC;IACzB,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,IAAI,KAAK,CAAC,KAAK,GAAG,MAAM;YAAE,SAAS;QACnC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACjB,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC;IACrB,CAAC;IAED,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,KAAK,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;QACzB,IAAI,KAAK,CAAC,KAAK,GAAG,GAAG,EAAE,CAAC;YACtB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QACzE,CAAC;QACD,QAAQ,CAAC,IAAI,CAAC;YACZ,IAAI,EAAE,WAAW;YACjB,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC;SACzC,CAAC,CAAC;QACH,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;IAClB,CAAC;IACD,IAAI,GAAG,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACtB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAC5D,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,gDAAgD;AAChD,MAAM,UAAU,IAAI,CAAC,QAAmB;IACtC,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC9C,CAAC;AAED,iEAAiE;AACjE,MAAM,UAAU,cAAc,CAAC,QAAmB;IAChD,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,WAAW,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;AAC3E,CAAC"}

package/dist/shared-prefix.d.ts

@@ -0,0 +1,63 @@
+import type { TokenCounter } from './types.js';
+/**
+ * The prefix length below which unifying buys nothing, and `Infinity` when there
+ * is no answer.
+ *
+ * Lives here rather than in the CLI because it is the gate on this whole report
+ * and a gate with no test is decoration. It was a private function in the CLI
+ * first; a mutation run deleting the `unknown` branch survived, which is what a
+ * function nothing can reach looks like from the outside.
+ *
+ * `unknown` and a missing minimum both yield `Infinity`, so nothing is reported.
+ * That is the direction to fail: telling somebody to unify a preamble across
+ * twelve files to enable caching their provider may not offer spends their
+ * afternoon, and unlike a wrong number on a report nothing later corrects it.
+ */
+export declare function cacheableMinimum(model: {
+    caching?: string | null;
+    cacheMinTokens?: number | null;
+} | undefined): number;
+export interface PrefixCandidate {
+    /** How the report names it. A path, or a path with a marked prompt's id. */
+    path: string;
+    text: string;
+}
+export interface SharedPrefix {
+    /** Every prompt in the group, in the order they were given. */
+    paths: string[];
+    /** Blocks the group has in common, counted from the start. */
+    blocks: number;
+    /** Estimated tokens in that prefix, as the first prompt writes it. */
+    tokens: number;
+    /**
+     * What stops the prefixes being byte-identical today.
+     *
+     * `whitespace` — collapsing runs of whitespace makes them equal, so a
+     * formatter fixes it and no wording decision is needed.
+     * `wording` — something else differs: capitalisation, punctuation, word order,
+     * or genuinely different words. Someone has to choose.
+     */
+    drift: 'whitespace' | 'wording';
+    /** The shared prefix as the first prompt in the group writes it. */
+    sample: string;
+}
+export interface SharedPrefixOptions {
+    /**
+     * The model's cacheable minimum. A shared prefix below it buys nothing, so it
+     * is not reported — the same refusal `reorder.ts` makes, for the same reason:
+     * a change that recovers nothing is a diff for its own sake.
+     */
+    minTokens?: number;
+    countTokens?: TokenCounter;
+}
+/**
+ * Groups of prompts that could share a cache prefix and do not.
+ *
+ * Grouped by their **first** block, which is not a shortcut. Caching matches from
+ * the beginning of the request, so two prompts whose opening paragraphs differ
+ * share nothing no matter how identical the rest is — a group keyed on anything
+ * later would name prompts that cannot be made to share a prefix without
+ * reordering them, which is a different and far more dangerous change.
+ */
+export declare function sharedPrefixes(prompts: readonly PrefixCandidate[], options?: SharedPrefixOptions): SharedPrefix[];
+//# sourceMappingURL=shared-prefix.d.ts.map

package/dist/shared-prefix.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"shared-prefix.d.ts","sourceRoot":"","sources":["../src/shared-prefix.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAmD/C;;;;;;;;;;;;;GAaG;AACH,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE;IAAE,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAAC,cAAc,CAAC,EAAE,MAAM,GAAG,IAAI,CAAA;CAAE,GAAG,SAAS,GAC7E,MAAM,CAIR;AAED,MAAM,WAAW,eAAe;IAC9B,4EAA4E;IAC5E,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,YAAY;IAC3B,+DAA+D;IAC/D,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,8DAA8D;IAC9D,MAAM,EAAE,MAAM,CAAC;IACf,sEAAsE;IACtE,MAAM,EAAE,MAAM,CAAC;IACf;;;;;;;OAOG;IACH,KAAK,EAAE,YAAY,GAAG,SAAS,CAAC;IAChC,oEAAoE;IACpE,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,mBAAmB;IAClC;;;;OAIG;IACH,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,WAAW,CAAC,EAAE,YAAY,CAAC;CAC5B;AAED;;;;;;;;GAQG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,SAAS,eAAe,EAAE,EACnC,OAAO,GAAE,mBAAwB,GAChC,YAAY,EAAE,CA0EhB"}