@discretion/core 0.1.0

package/dist/src/context/documentProfile.js

@@ -0,0 +1,253 @@
+/**
+ * Stage 3, DOCUMENT TYPE.
+ *
+ * SPEC.md: "DOCUMENT TYPE — detect whether input is prose, source code (and
+ * which language), JSON, YAML, CSV, a log dump, a markdown table, or an email
+ * thread. Each mode shifts weights. Code mode raises secret sensitivity and
+ * lowers person-name sensitivity, since identifiers in code are usually not
+ * people."
+ *
+ * Format detection is evidence-scored rather than first-match: several forms
+ * genuinely overlap (a YAML file is a superset of many key/value shapes, a CV
+ * in prose is full of "Label: value" lines that look like YAML), so each
+ * candidate format accumulates evidence and the strongest wins. The evidence
+ * strings are kept and reported, which is what makes a misclassification
+ * diagnosable instead of mysterious.
+ *
+ * The DOMAIN axis is deliberately separate from format — see the note on
+ * `DocumentDomain`. A medical record can arrive as prose, CSV, or JSON.
+ */
+import { buildStructureIndex } from './structure.js';
+import { foldForMatch } from './triggers.js';
+/** Minimum distinct domain-term hits before a domain is claimed over 'general'. */
+const MIN_DOMAIN_HITS = 2;
+/** Fraction of non-empty lines a form must claim to be considered dominant. */
+const DOMINANT_LINE_RATIO = 0.34;
+// ─────────────────────────────────────────────────────────────────────────────
+// Line-level probes
+// ─────────────────────────────────────────────────────────────────────────────
+const TABLE_SEPARATOR = /^\s*\|?\s*:?-{2,}:?\s*(\|\s*:?-{2,}:?\s*)*\|?\s*$/;
+/** ISO-8601, syslog, and bracketed-timestamp line starts. */
+const LOG_TIMESTAMP = /^\s*[[(]?(\d{4}-\d{2}-\d{2}[T ]\d{2}:\d{2}:\d{2}|[A-Z][a-z]{2} {1,2}\d{1,2} \d{2}:\d{2}:\d{2})/;
+const LOG_LEVEL = /\b(TRACE|DEBUG|INFO|WARN|WARNING|ERROR|FATAL|CRITICAL)\b/;
+/**
+ * Keywords that are strong evidence of source code.
+ *
+ * Deliberately excludes words that are ordinary English: `new`, `for`, `from`,
+ * `if`, `else`, `while`, `return`, `class`, `try`, `throw`, `package` and
+ * `public` all appear constantly in prose. Including them classified "everyone
+ * agreed to the new schedule" as source code, and code mode shifts PERSON
+ * sensitivity down — so the cost of a loose keyword list is missed names.
+ */
+const CODE_KEYWORD = /(^|\s)(const|let|var|function|def|elif|lambda|func|fn|impl|struct|async|await|=>|null|undefined|nullptr|println|printf|console\.log|System\.out)\b/;
+const CODE_PUNCTUATION = /(=>|::|->|\)\s*\{|\{\s*$|;\s*$|\)\s*;)/;
+const COMMENT_LINE = /^\s*(\/\/|#|\/\*|\*|--)/;
+/** RFC-822-style headers that mark an email thread. */
+const EMAIL_HEADER = /^\s*(From|To|Cc|Bcc|Subject|Date|Sent|Reply-To)\s*:\s+\S/i;
+const EMAIL_QUOTE = /^\s*>+\s?\S/;
+function nonEmptyLines(text) {
+    return text.split('\n').filter((l) => l.trim().length > 0);
+}
+function ratio(count, total) {
+    return total === 0 ? 0 : count / total;
+}
+// ─────────────────────────────────────────────────────────────────────────────
+// Format scoring
+// ─────────────────────────────────────────────────────────────────────────────
+function scoreJson(text, lines, scores) {
+    const trimmed = text.trim();
+    const braced = /^[[{]/.test(trimmed) && /[\]}]$/.test(trimmed);
+    const keyLines = lines.filter((l) => /^\s*"[^"]+"\s*:/.test(l)).length;
+    const keyRatio = ratio(keyLines, lines.length);
+    if (!braced && keyRatio < DOMINANT_LINE_RATIO)
+        return;
+    const evidence = [];
+    let score = 0;
+    if (braced) {
+        score += 2;
+        evidence.push('wrapped in braces/brackets');
+    }
+    if (keyRatio >= DOMINANT_LINE_RATIO) {
+        score += 2 + keyRatio;
+        evidence.push(`${Math.round(keyRatio * 100)}% quoted-key lines`);
+    }
+    scores.push({ format: 'json', score, evidence });
+}
+function scoreYaml(lines, index, scores) {
+    const yamlSlots = index.slots.filter((s) => s.kind === 'yaml').length;
+    const yamlRatio = ratio(yamlSlots, lines.length);
+    const documentMarker = lines.some((l) => /^---\s*$/.test(l));
+    const indented = lines.filter((l) => /^\s+\S/.test(l)).length;
+    if (yamlRatio < DOMINANT_LINE_RATIO && !documentMarker)
+        return;
+    const evidence = [];
+    let score = 0;
+    if (yamlRatio >= DOMINANT_LINE_RATIO) {
+        score += 1.5 + yamlRatio;
+        evidence.push(`${Math.round(yamlRatio * 100)}% identifier-key lines`);
+    }
+    if (documentMarker) {
+        score += 1;
+        evidence.push('--- document marker');
+    }
+    if (indented > 0) {
+        score += 0.5;
+        evidence.push('indented block structure');
+    }
+    scores.push({ format: 'yaml', score, evidence });
+}
+function scoreCsv(lines, scores) {
+    if (lines.length < 2)
+        return;
+    const counts = lines.map((l) => (l.match(/,/g) ?? []).length);
+    const first = counts[0];
+    if (first === undefined || first < 1)
+        return;
+    const consistent = counts.filter((c) => c === first).length;
+    const consistentRatio = ratio(consistent, counts.length);
+    if (consistentRatio < 0.8)
+        return;
+    scores.push({
+        format: 'csv',
+        score: 2 + first * 0.1 + consistentRatio,
+        evidence: [`${consistent}/${counts.length} lines with ${first} commas`],
+    });
+}
+function scoreMarkdownTable(lines, scores) {
+    const separators = lines.filter((l) => TABLE_SEPARATOR.test(l) && l.includes('-')).length;
+    if (separators === 0)
+        return;
+    const piped = lines.filter((l) => l.includes('|')).length;
+    scores.push({
+        format: 'markdown-table',
+        score: 2.5 + ratio(piped, lines.length),
+        evidence: [`${separators} table separator row(s)`, `${piped} pipe-delimited lines`],
+    });
+}
+function scoreLog(lines, scores) {
+    const stamped = lines.filter((l) => LOG_TIMESTAMP.test(l)).length;
+    const levelled = lines.filter((l) => LOG_LEVEL.test(l)).length;
+    const stampRatio = ratio(stamped, lines.length);
+    if (stampRatio < DOMINANT_LINE_RATIO && ratio(levelled, lines.length) < DOMINANT_LINE_RATIO)
+        return;
+    const evidence = [];
+    let score = 1;
+    if (stampRatio > 0) {
+        score += 1.5 + stampRatio;
+        evidence.push(`${Math.round(stampRatio * 100)}% timestamped lines`);
+    }
+    if (levelled > 0) {
+        score += ratio(levelled, lines.length);
+        evidence.push(`${levelled} lines with a log level`);
+    }
+    scores.push({ format: 'log', score, evidence });
+}
+function scoreCode(lines, scores) {
+    const keyworded = lines.filter((l) => CODE_KEYWORD.test(l)).length;
+    const punctuated = lines.filter((l) => CODE_PUNCTUATION.test(l)).length;
+    const commented = lines.filter((l) => COMMENT_LINE.test(l)).length;
+    // Punctuation is required, not merely additive. Prose can borrow a keyword
+    // but it does not end lines with `;`, `{` or `);` — that is what separates
+    // a sentence containing "function" from a line of code.
+    if (punctuated === 0)
+        return;
+    const signal = ratio(keyworded + punctuated, lines.length);
+    if (signal < 0.25)
+        return;
+    const evidence = [`${keyworded} keyword lines`, `${punctuated} code-punctuation lines`];
+    if (commented > 0)
+        evidence.push(`${commented} comment lines`);
+    scores.push({ format: 'code', score: 1 + signal * 2, evidence });
+}
+function scoreEmail(lines, scores) {
+    const headers = lines.filter((l) => EMAIL_HEADER.test(l)).length;
+    const quoted = lines.filter((l) => EMAIL_QUOTE.test(l)).length;
+    if (headers < 2 && quoted === 0)
+        return;
+    const evidence = [];
+    let score = 0;
+    if (headers >= 2) {
+        score += 2 + headers * 0.2;
+        evidence.push(`${headers} message headers`);
+    }
+    if (quoted > 0) {
+        score += 0.5;
+        evidence.push(`${quoted} quoted-reply lines`);
+    }
+    scores.push({ format: 'email', score, evidence });
+}
+// ─────────────────────────────────────────────────────────────────────────────
+// Domain scoring
+// ─────────────────────────────────────────────────────────────────────────────
+/**
+ * Does `text` contain `term` as a whole word?
+ *
+ * Plain substring matching found "lan" inside "balance" and counted it as
+ * financial vocabulary. Scripts that do not space their words cannot use a
+ * boundary test, so for those the substring match stands — which is correct,
+ * since that is how those scripts are written.
+ */
+function containsAsWord(text, term) {
+    if (!/^[\p{L}\p{M}\p{N} '-]+$/u.test(term))
+        return text.includes(term);
+    let at = text.indexOf(term);
+    while (at !== -1) {
+        const before = at === 0 ? '' : text[at - 1] ?? '';
+        const after = text[at + term.length] ?? '';
+        if (!/\p{L}/u.test(before) && !/\p{L}/u.test(after))
+            return true;
+        at = text.indexOf(term, at + 1);
+    }
+    return false;
+}
+function detectDomain(folded, lexicon) {
+    let best = { domain: 'general', evidence: [] };
+    let bestCount = 0;
+    for (const [domain, terms] of Object.entries(lexicon)) {
+        const hits = [];
+        for (const term of terms ?? []) {
+            const foldedTerm = foldForMatch(term);
+            if (foldedTerm.length > 0 && containsAsWord(folded, foldedTerm))
+                hits.push(foldedTerm);
+        }
+        if (hits.length > bestCount) {
+            bestCount = hits.length;
+            best = { domain: domain, evidence: hits.slice(0, 8) };
+        }
+    }
+    return bestCount >= MIN_DOMAIN_HITS ? best : { domain: 'general', evidence: [] };
+}
+// ─────────────────────────────────────────────────────────────────────────────
+// Entry point
+// ─────────────────────────────────────────────────────────────────────────────
+/**
+ * Classify a document's format and subject domain.
+ *
+ * `structureIndex` is accepted rather than rebuilt so callers that already
+ * have one (the analyzer does) pay for it once.
+ */
+export function profileDocument(text, options = {}) {
+    const lines = nonEmptyLines(text);
+    const index = options.structureIndex ?? buildStructureIndex(text);
+    const scores = [];
+    scoreMarkdownTable(lines, scores);
+    scoreJson(text, lines, scores);
+    scoreCsv(lines, scores);
+    scoreYaml(lines, index, scores);
+    scoreLog(lines, scores);
+    scoreCode(lines, scores);
+    scoreEmail(lines, scores);
+    scores.sort((a, b) => b.score - a.score);
+    const winner = scores[0];
+    const { domain, evidence: domainEvidence } = detectDomain(foldForMatch(text), options.domainLexicon ?? {});
+    if (winner === undefined) {
+        return {
+            format: lines.length > 0 ? 'prose' : 'unknown',
+            domain,
+            formatEvidence: lines.length > 0 ? ['no structural form matched'] : ['empty document'],
+            domainEvidence,
+        };
+    }
+    return { format: winner.format, domain, formatEvidence: winner.evidence, domainEvidence };
+}
+//# sourceMappingURL=documentProfile.js.map

package/dist/src/context/documentProfile.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"documentProfile.js","sourceRoot":"","sources":["../../../src/context/documentProfile.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,EAAE,mBAAmB,EAAuB,MAAM,gBAAgB,CAAC;AAC1E,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAM7C,mFAAmF;AACnF,MAAM,eAAe,GAAG,CAAC,CAAC;AAE1B,+EAA+E;AAC/E,MAAM,mBAAmB,GAAG,IAAI,CAAC;AAQjC,gFAAgF;AAChF,oBAAoB;AACpB,gFAAgF;AAEhF,MAAM,eAAe,GAAG,mDAAmD,CAAC;AAC5E,6DAA6D;AAC7D,MAAM,aAAa,GACjB,gGAAgG,CAAC;AACnG,MAAM,SAAS,GAAG,0DAA0D,CAAC;AAC7E;;;;;;;;GAQG;AACH,MAAM,YAAY,GAChB,oJAAoJ,CAAC;AACvJ,MAAM,gBAAgB,GAAG,wCAAwC,CAAC;AAClE,MAAM,YAAY,GAAG,yBAAyB,CAAC;AAC/C,uDAAuD;AACvD,MAAM,YAAY,GAAG,2DAA2D,CAAC;AACjF,MAAM,WAAW,GAAG,aAAa,CAAC;AAElC,SAAS,aAAa,CAAC,IAAY;IACjC,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AAC7D,CAAC;AAED,SAAS,KAAK,CAAC,KAAa,EAAE,KAAa;IACzC,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC;AACzC,CAAC;AAED,gFAAgF;AAChF,iBAAiB;AACjB,gFAAgF;AAEhF,SAAS,SAAS,CAAC,IAAY,EAAE,KAAwB,EAAE,MAAqB;IAC9E,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAC/D,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACvE,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAC/C,IAAI,CAAC,MAAM,IAAI,QAAQ,GAAG,mBAAmB;QAAE,OAAO;IAEtD,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,MAAM,EAAE,CAAC;QACX,KAAK,IAAI,CAAC,CAAC;QACX,QAAQ,CAAC,IAAI,CAAC,4BAA4B,CAAC,CAAC;IAC9C,CAAC;IACD,IAAI,QAAQ,IAAI,mBAAmB,EAAE,CAAC;QACpC,KAAK,IAAI,CAAC,GAAG,QAAQ,CAAC;QACtB,QAAQ,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,GAAG,CAAC,oBAAoB,CAAC,CAAC;IACnE,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAC;AACnD,CAAC;AAED,SAAS,SAAS,CAAC,KAAwB,EAAE,KAAqB,EAAE,MAAqB;IACvF,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC,CAAC,MAAM,CAAC;IACtE,MAAM,SAAS,GAAG,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IACjD,MAAM,cAAc,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7D,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC9D,IAAI,SAAS,GAAG,mBAAmB,IAAI,CAAC,cAAc;QAAE,OAAO;IAE/D,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,SAAS,IAAI,mBAAmB,EAAE,CAAC;QACrC,KAAK,IAAI,GAAG,GAAG,SAAS,CAAC;QACzB,QAAQ,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,GAAG,CAAC,wBAAwB,CAAC,CAAC;IACxE,CAAC;IACD,IAAI,cAAc,EAAE,CAAC;QACnB,KAAK,IAAI,CAAC,CAAC;QACX,QAAQ,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC;IACvC,CAAC;IACD,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;QACjB,KAAK,IAAI,GAAG,CAAC;QACb,QAAQ,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC;IAC5C,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAC;AACnD,CAAC;AAED,SAAS,QAAQ,CAAC,KAAwB,EAAE,MAAqB;IAC/D,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO;IAC7B,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC;IAC9D,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IACxB,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,GAAG,CAAC;QAAE,OAAO;IAC7C,MAAM,UAAU,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,MAAM,CAAC;IAC5D,MAAM,eAAe,GAAG,KAAK,CAAC,UAAU,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC;IACzD,IAAI,eAAe,GAAG,GAAG;QAAE,OAAO;IAElC,MAAM,CAAC,IAAI,CAAC;QACV,MAAM,EAAE,KAAK;QACb,KAAK,EAAE,CAAC,GAAG,KAAK,GAAG,GAAG,GAAG,eAAe;QACxC,QAAQ,EAAE,CAAC,GAAG,UAAU,IAAI,MAAM,CAAC,MAAM,eAAe,KAAK,SAAS,CAAC;KACxE,CAAC,CAAC;AACL,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAwB,EAAE,MAAqB;IACzE,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;IAC1F,IAAI,UAAU,KAAK,CAAC;QAAE,OAAO;IAC7B,MAAM,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;IAC1D,MAAM,CAAC,IAAI,CAAC;QACV,MAAM,EAAE,gBAAgB;QACxB,KAAK,EAAE,GAAG,GAAG,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC;QACvC,QAAQ,EAAE,CAAC,GAAG,UAAU,yBAAyB,EAAE,GAAG,KAAK,uBAAuB,CAAC;KACpF,CAAC,CAAC;AACL,CAAC;AAED,SAAS,QAAQ,CAAC,KAAwB,EAAE,MAAqB;IAC/D,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAClE,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,MAAM,UAAU,GAAG,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAChD,IAAI,UAAU,GAAG,mBAAmB,IAAI,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,MAAM,CAAC,GAAG,mBAAmB;QAAE,OAAO;IAEpG,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC;QACnB,KAAK,IAAI,GAAG,GAAG,UAAU,CAAC;QAC1B,QAAQ,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,GAAG,GAAG,CAAC,qBAAqB,CAAC,CAAC;IACtE,CAAC;IACD,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;QACjB,KAAK,IAAI,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;QACvC,QAAQ,CAAC,IAAI,CAAC,GAAG,QAAQ,yBAAyB,CAAC,CAAC;IACtD,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAC;AAClD,CAAC;AAED,SAAS,SAAS,CAAC,KAAwB,EAAE,MAAqB;IAChE,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACnE,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACxE,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACnE,2EAA2E;IAC3E,2EAA2E;IAC3E,wDAAwD;IACxD,IAAI,UAAU,KAAK,CAAC;QAAE,OAAO;IAC7B,MAAM,MAAM,GAAG,KAAK,CAAC,SAAS,GAAG,UAAU,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAC3D,IAAI,MAAM,GAAG,IAAI;QAAE,OAAO;IAE1B,MAAM,QAAQ,GAAa,CAAC,GAAG,SAAS,gBAAgB,EAAE,GAAG,UAAU,yBAAyB,CAAC,CAAC;IAClG,IAAI,SAAS,GAAG,CAAC;QAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,SAAS,gBAAgB,CAAC,CAAC;IAC/D,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,GAAG,MAAM,GAAG,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC;AACnE,CAAC;AAED,SAAS,UAAU,CAAC,KAAwB,EAAE,MAAqB;IACjE,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACjE,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,IAAI,OAAO,GAAG,CAAC,IAAI,MAAM,KAAK,CAAC;QAAE,OAAO;IAExC,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,OAAO,IAAI,CAAC,EAAE,CAAC;QACjB,KAAK,IAAI,CAAC,GAAG,OAAO,GAAG,GAAG,CAAC;QAC3B,QAAQ,CAAC,IAAI,CAAC,GAAG,OAAO,kBAAkB,CAAC,CAAC;IAC9C,CAAC;IACD,IAAI,MAAM,GAAG,CAAC,EAAE,CAAC;QACf,KAAK,IAAI,GAAG,CAAC;QACb,QAAQ,CAAC,IAAI,CAAC,GAAG,MAAM,qBAAqB,CAAC,CAAC;IAChD,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAC;AACpD,CAAC;AAED,gFAAgF;AAChF,iBAAiB;AACjB,gFAAgF;AAEhF;;;;;;;GAOG;AACH,SAAS,cAAc,CAAC,IAAY,EAAE,IAAY;IAChD,IAAI,CAAC,0BAA0B,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAEvE,IAAI,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAC5B,OAAO,EAAE,KAAK,CAAC,CAAC,EAAE,CAAC;QACjB,MAAM,MAAM,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAClD,MAAM,KAAK,GAAG,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QAC3C,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO,IAAI,CAAC;QACjE,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;IAClC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,YAAY,CACnB,MAAc,EACd,OAAsB;IAEtB,IAAI,IAAI,GAAmD,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IAC/F,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACtD,MAAM,IAAI,GAAa,EAAE,CAAC;QAC1B,KAAK,MAAM,IAAI,IAAI,KAAK,IAAI,EAAE,EAAE,CAAC;YAC/B,MAAM,UAAU,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;YACtC,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,IAAI,cAAc,CAAC,MAAM,EAAE,UAAU,CAAC;gBAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QACzF,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;YAC5B,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC;YACxB,IAAI,GAAG,EAAE,MAAM,EAAE,MAAwB,EAAE,QAAQ,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QAC1E,CAAC;IACH,CAAC;IAED,OAAO,SAAS,IAAI,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AACnF,CAAC;AAED,gFAAgF;AAChF,cAAc;AACd,gFAAgF;AAEhF;;;;;GAKG;AACH,MAAM,UAAU,eAAe,CAC7B,IAAY,EACZ,UAAgG,EAAE;IAElG,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,KAAK,GAAG,OAAO,CAAC,cAAc,IAAI,mBAAmB,CAAC,IAAI,CAAC,CAAC;IAElE,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,kBAAkB,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAClC,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;IAC/B,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IACxB,SAAS,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;IAChC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IACxB,SAAS,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IACzB,UAAU,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAE1B,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IACzC,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IAEzB,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,cAAc,EAAE,GAAG,YAAY,CACvD,YAAY,CAAC,IAAI,CAAC,EAClB,OAAO,CAAC,aAAa,IAAI,EAAE,CAC5B,CAAC;IAEF,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QACzB,OAAO;YACL,MAAM,EAAE,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS;YAC9C,MAAM;YACN,cAAc,EAAE,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,4BAA4B,CAAC,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC;YACtF,cAAc;SACf,CAAC;IACJ,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,cAAc,EAAE,MAAM,CAAC,QAAQ,EAAE,cAAc,EAAE,CAAC;AAC5F,CAAC"}

package/dist/src/context/negativeRules.d.ts

@@ -0,0 +1,60 @@
+/**
+ * Stage 3, NEGATIVE CONTEXT.
+ *
+ * SPEC.md: "NEGATIVE CONTEXT — signals that a candidate is NOT sensitive:
+ * inside a code comment describing a format, in a documentation example
+ * block, a known dummy value, lorem ipsum, a test fixture, a UUID in a log
+ * line, a git SHA. These must actively suppress."
+ *
+ * THE SAFETY RULE FOR THIS FILE. Every rule here can cause a real secret to
+ * go unreported, which in this product is the worst possible failure. So each
+ * rule must satisfy three things, and each is enforced by review, not by the
+ * type system alone:
+ *
+ *   1. It states a GENERAL principle about how text is written — never a
+ *      literal drawn from the evaluation corpus. A rule tuned to the corpus
+ *      would make the measured precision gain self-fulfilling and dishonest.
+ *   2. It states the real positive it risks suppressing, in `risk`.
+ *   3. It suppresses only on CONCLUSIVE evidence that the value is a
+ *      different kind of thing. Where the evidence merely weakens the case,
+ *      the rule returns a penalty instead, and the candidate survives at
+ *      lower confidence.
+ *
+ * Note what is deliberately absent: no rule here looks at OTHER candidates'
+ * spans. Deciding that a generic high-entropy match should yield to a
+ * structurally validated JWT covering the same characters is cross-type
+ * overlap resolution, which SPEC.md assigns to Stage 4. Keeping it out is
+ * what stops Stage 3 quietly absorbing the next milestone's work.
+ */
+import type { EntityType } from '../types.js';
+import type { NegativeRule } from './types.js';
+/**
+ * The full rule set, in evaluation order.
+ *
+ * Order is not significant for correctness — every rule is evaluated and the
+ * strongest action wins — but keeping the conclusive structural rules first
+ * makes traces easier to read.
+ */
+export declare const NEGATIVE_RULES: readonly NegativeRule[];
+/**
+ * DELIBERATELY NOT HERE: containment.
+ *
+ * The M7 error taxonomy proposed four further rules that suppress a candidate
+ * because ANOTHER candidate's span covers it — a format-only detector yielding
+ * to a validated one, non-sensitivity propagating across schemes, an NER span
+ * inside a structured Stage 1 span. Together they account for the largest
+ * measured error class in the corpus, and they are all excluded.
+ *
+ * Two reasons, both ratified rather than assumed. First, ARCHITECTURE.md D19:
+ * deciding which of two overlapping type claims wins is Stage 4 resolution,
+ * and Stage 3 pre-empting it was exactly the mistake the GENERIC_SECRET
+ * measurement exposed. Second, the taxonomy's own highest-priority residual is
+ * a SPAN-HYGIENE PREREQUISITE — url-with-credentials and connection-string
+ * spans were measured straddling CSV and line boundaries — and containment
+ * suppression is only ever as safe as the covering span is correct. Shipping
+ * it before that is fixed would let a bad span silently un-redact whatever it
+ * wrongly covers.
+ */
+/** True when the rule may act on this entity type. */
+export declare function ruleApplies(rule: NegativeRule, type: EntityType): boolean;
+//# sourceMappingURL=negativeRules.d.ts.map

package/dist/src/context/negativeRules.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"negativeRules.d.ts","sourceRoot":"","sources":["../../../src/context/negativeRules.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAIH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAC9C,OAAO,KAAK,EAAE,YAAY,EAAe,MAAM,YAAY,CAAC;AA8qB5D;;;;;;GAMG;AACH,eAAO,MAAM,cAAc,EAAE,SAAS,YAAY,EAcjD,CAAC;AAEF;;;;;;;;;;;;;;;;;;GAkBG;AAEH,sDAAsD;AACtD,wBAAgB,WAAW,CAAC,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,UAAU,GAAG,OAAO,CAEzE"}