@profile-as-code/core 0.1.1 → 0.1.2

package/dist/cite.d.ts

@@ -0,0 +1,91 @@
+/**
+ * cite.ts — the deterministic citation matcher (SPEC §7.4, issue #291).
+ *
+ * The engine answers ONE question about an answer-engine record: did this saved
+ * answer cite this entity? It is a pure function over the SSOT and the answer
+ * bytes — no network, no model, no clock (D19) — because the alternative
+ * (substring matching) is why GEO reporting overreports, and the only fix is a
+ * matcher that says exactly which text it matched, why, and when it refuses to
+ * guess.
+ *
+ * The vocabulary is the workspace's DECLARED data, never world knowledge:
+ *   name     identity.name
+ *   alias    identity.orgName (the one declared alternate name the schema has)
+ *   handle   identity.handle (its final token, '@' stripped) + socialHandles
+ *            rows with status: claimed
+ *   hosts    the role: identity URL and the sameAs set (structured-data.ts —
+ *            one derivation, never a second)
+ *   partial  a contiguous multi-token prefix of the name (low confidence)
+ *
+ * Normalization and word-boundary matching live in `normalize.ts`. It is
+ * `cite-check`'s fold, not `geoCheck`'s: `geoCheck` keeps its own case-only
+ * comparison so this matcher cannot move a published GEO score (§8).
+ *
+ * Guards make a text hit NOT a citation (`cited: false`): a shorter term
+ * contained in a longer declared one (`longer-name`), a one-word name with no
+ * declared disambiguator in its sentence (`ambiguous-term` + `no-qualifier`), and
+ * a mention inside a contrast clause (`contrast-context`). `competing:<phrase>`
+ * never decides on its own — it only names the longer reading the text carries
+ * when the hit is already ambiguous, so ordinary prose cannot become a
+ * competitor through an adjacent capital. `confidence: low` is never citable
+ * downstream (declared in §7.4).
+ */
+import type { Profile } from "./types.js";
+/** The rung a span was matched by — one kind per ladder rung (SPEC §7.4). */
+export type CiteSpanKind = "citation-url" | "name" | "alias" | "handle" | "partial";
+/** The best surviving rung. `low` is never a citation in a downstream report. */
+export type CiteConfidence = "high" | "medium" | "low";
+/** One citation row of the runner's input contract. */
+export interface CiteCitation {
+    url: string;
+    title?: string;
+}
+/**
+ * One saved answer (the runner's contract, never produced here). The matcher
+ * reads `answer` + `citations`; the rest is carried through for a report.
+ */
+export interface CiteAnswer {
+    engine?: string | null;
+    askedAt?: string | null;
+    question?: string | null;
+    answer: string;
+    citations: CiteCitation[];
+}
+/** A matched span: a character range in `answer`, or a citation pointer. */
+export interface CiteSpan {
+    /** Original offset (inclusive). 0 for a `citation-url` span (not in the text). */
+    start: number;
+    /** Original offset (exclusive). 0 for a `citation-url` span. */
+    end: number;
+    kind: CiteSpanKind;
+    /** The citation URL — only on `citation-url` spans. */
+    url?: string;
+}
+/** Per-answer result. Never aggregated by the engine (SPEC §7.4). */
+export interface CiteMatch {
+    cited: boolean;
+    confidence: CiteConfidence;
+    spans: CiteSpan[];
+    reasons: string[];
+}
+/**
+ * The vocabulary the matcher resolved from the workspace. Exposed in `--json`
+ * so a report can say WHAT was matched without re-reading the SSOT.
+ */
+export interface CiteVocabulary {
+    name: string;
+    aliases: string[];
+    handles: string[];
+    partials: string[];
+    hosts: string[];
+}
+/** The public vocabulary of a workspace (the `entity` block of `--json`). */
+export declare function citeVocabulary(profile: Profile): CiteVocabulary;
+/**
+ * Match one saved answer against the workspace's declared entity (SPEC §7.4).
+ * Pure: the same profile + answer bytes always produce the same result.
+ */
+export declare function citeMatch(profile: Profile, answer: CiteAnswer): CiteMatch;
+/** Exported for tests: the declared host of a URL under the matcher's rule. */
+export declare function citationHost(url: string): string | undefined;
+//# sourceMappingURL=cite.d.ts.map

package/dist/cite.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cite.d.ts","sourceRoot":"","sources":["../src/cite.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAMH,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AAE1C,6EAA6E;AAC7E,MAAM,MAAM,YAAY,GAAG,cAAc,GAAG,MAAM,GAAG,OAAO,GAAG,QAAQ,GAAG,SAAS,CAAC;AAEpF,iFAAiF;AACjF,MAAM,MAAM,cAAc,GAAG,MAAM,GAAG,QAAQ,GAAG,KAAK,CAAC;AAEvD,uDAAuD;AACvD,MAAM,WAAW,YAAY;IAC3B,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;;GAGG;AACH,MAAM,WAAW,UAAU;IACzB,MAAM,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,QAAQ,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,YAAY,EAAE,CAAC;CAC3B;AAED,4EAA4E;AAC5E,MAAM,WAAW,QAAQ;IACvB,kFAAkF;IAClF,KAAK,EAAE,MAAM,CAAC;IACd,gEAAgE;IAChE,GAAG,EAAE,MAAM,CAAC;IACZ,IAAI,EAAE,YAAY,CAAC;IACnB,uDAAuD;IACvD,GAAG,CAAC,EAAE,MAAM,CAAC;CACd;AAED,qEAAqE;AACrE,MAAM,WAAW,SAAS;IACxB,KAAK,EAAE,OAAO,CAAC;IACf,UAAU,EAAE,cAAc,CAAC;IAC3B,KAAK,EAAE,QAAQ,EAAE,CAAC;IAClB,OAAO,EAAE,MAAM,EAAE,CAAC;CACnB;AAED;;;GAGG;AACH,MAAM,WAAW,cAAc;IAC7B,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,KAAK,EAAE,MAAM,EAAE,CAAC;CACjB;AAsND,6EAA6E;AAC7E,wBAAgB,cAAc,CAAC,OAAO,EAAE,OAAO,GAAG,cAAc,CAE/D;AAgFD;;;GAGG;AACH,wBAAgB,SAAS,CAAC,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,UAAU,GAAG,SAAS,CA0JzE;AAaD,+EAA+E;AAC/E,wBAAgB,YAAY,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAE5D"}

package/dist/cite.js

@@ -0,0 +1,440 @@
+/**
+ * cite.ts — the deterministic citation matcher (SPEC §7.4, issue #291).
+ *
+ * The engine answers ONE question about an answer-engine record: did this saved
+ * answer cite this entity? It is a pure function over the SSOT and the answer
+ * bytes — no network, no model, no clock (D19) — because the alternative
+ * (substring matching) is why GEO reporting overreports, and the only fix is a
+ * matcher that says exactly which text it matched, why, and when it refuses to
+ * guess.
+ *
+ * The vocabulary is the workspace's DECLARED data, never world knowledge:
+ *   name     identity.name
+ *   alias    identity.orgName (the one declared alternate name the schema has)
+ *   handle   identity.handle (its final token, '@' stripped) + socialHandles
+ *            rows with status: claimed
+ *   hosts    the role: identity URL and the sameAs set (structured-data.ts —
+ *            one derivation, never a second)
+ *   partial  a contiguous multi-token prefix of the name (low confidence)
+ *
+ * Normalization and word-boundary matching live in `normalize.ts`. It is
+ * `cite-check`'s fold, not `geoCheck`'s: `geoCheck` keeps its own case-only
+ * comparison so this matcher cannot move a published GEO score (§8).
+ *
+ * Guards make a text hit NOT a citation (`cited: false`): a shorter term
+ * contained in a longer declared one (`longer-name`), a one-word name with no
+ * declared disambiguator in its sentence (`ambiguous-term` + `no-qualifier`), and
+ * a mention inside a contrast clause (`contrast-context`). `competing:<phrase>`
+ * never decides on its own — it only names the longer reading the text carries
+ * when the hit is already ambiguous, so ordinary prose cannot become a
+ * competitor through an adjacent capital. `confidence: low` is never citable
+ * downstream (declared in §7.4).
+ */
+import { byteOrder } from "./byte-order.js";
+import { disambiguationQualifiers } from "./geo.js";
+import { findTerm, foldForMatch, normalizeText, termOccursIn } from "./normalize.js";
+import { entityId, sameAsUrls } from "./structured-data.js";
+const RUNG_ORDER = ["name", "alias", "handle", "partial"];
+const RUNG_RANK = {
+    "citation-url": 0,
+    name: 0,
+    alias: 1,
+    handle: 2,
+    partial: 3,
+};
+const RUNG_REASON = {
+    "citation-url": "citation-url",
+    name: "exact-name",
+    alias: "alias",
+    handle: "handle",
+    partial: "partial",
+};
+/**
+ * The closed contrast lexicon (§7.4, declared limit): a mention inside a
+ * comparison is not an assertion about the entity. EN · ES · PT, matched
+ * case- and accent-insensitively as a phrase, only BEFORE the hit.
+ */
+const CONTRAST_MARKERS = [
+    "unlike",
+    "contrary to",
+    "as opposed to",
+    "rather than",
+    "instead of",
+    "versus",
+    "vs",
+    "compared to",
+    "a diferencia de",
+    "en contraste con",
+    "en lugar de",
+    "em vez de",
+    "ao contrário de",
+    "diferentemente de",
+];
+/** Reason ordering in the output set (SPEC §7.4: a fixed order). */
+const REASON_ORDER = [
+    "citation-url",
+    "identity-domain",
+    "sameAs-domain",
+    "url-invalid",
+    "exact-name",
+    "alias",
+    "handle",
+    "partial",
+    "longer-name",
+    "competing",
+    "ambiguous-term",
+    "no-qualifier",
+    "contrast-context",
+];
+/** Host of a URL, lowercased and with a leading `www.` stripped. */
+function hostOf(url) {
+    try {
+        const host = new URL(url).hostname.toLowerCase();
+        return host.startsWith("www.") ? host.slice(4) : host;
+    }
+    catch {
+        return undefined;
+    }
+}
+/** Citation host belongs to a declared host: equal, or a subdomain of it. */
+function sameSite(citationHost, declaredHost) {
+    return citationHost === declaredHost || citationHost.endsWith(`.${declaredHost}`);
+}
+/** A declared URL's path, without a trailing slash (`""` for the root). */
+function pathOf(url) {
+    try {
+        return new URL(url).pathname.replace(/\/+$/, "");
+    }
+    catch {
+        return undefined;
+    }
+}
+/**
+ * Does the cited URL point at the declared one? A shared host alone is not
+ * enough: `github.com/someone-else` shares `github.com` with a declared
+ * `github.com/elenavasquez`, and host-only matching would overreport exactly
+ * the way substring matching does. A declared URL on the root claims the whole
+ * host; a declared URL with a path claims that path and its descendants.
+ */
+function pathMatches(citationPath, declaredPath) {
+    if (declaredPath === "")
+        return true;
+    return citationPath === declaredPath || citationPath.startsWith(`${declaredPath}/`);
+}
+/**
+ * The declared handle inside a free-text `identity.handle` ("X @cluciani_ve"):
+ * its final whitespace-separated token, leading `@` stripped, trailing
+ * punctuation trimmed. Nothing is parsed beyond that (declared limit).
+ */
+function parseDeclaredHandle(raw) {
+    const tokens = raw.trim().split(/\s+/).filter(Boolean);
+    const last = tokens[tokens.length - 1];
+    if (last === undefined)
+        return undefined;
+    const cleaned = last.replace(/^@+/, "").replace(/[.,;:)\]}]+$/, "");
+    return cleaned.length >= 2 ? cleaned : undefined;
+}
+/** Push `value` unless its normalized form is already present (order preserved). */
+function pushUnique(out, seen, value) {
+    const key = normalizeText(value);
+    if (key === "" || seen.has(key))
+        return;
+    seen.add(key);
+    out.push(value);
+}
+/**
+ * Resolve the workspace's declared citation vocabulary. One source per fact:
+ * `sameAsUrls` and `entityId` are the structured-data set — never re-derived.
+ */
+function buildVocabulary(profile) {
+    const identity = profile.identity;
+    const name = identity.name.trim();
+    const used = new Set([normalizeText(name)]);
+    const aliases = [];
+    const orgName = identity.orgName?.trim();
+    if (orgName !== undefined && orgName !== "")
+        pushUnique(aliases, used, orgName);
+    const handles = [];
+    const handle = identity.handle?.trim();
+    if (handle !== undefined && handle !== "") {
+        const parsed = parseDeclaredHandle(handle);
+        if (parsed !== undefined)
+            pushUnique(handles, used, parsed);
+    }
+    for (const row of profile.socialHandles ?? []) {
+        if (row.status !== "claimed")
+            continue;
+        pushUnique(handles, used, row.handle.trim());
+    }
+    const partials = [];
+    const tokens = name.split(/\s+/).filter(Boolean);
+    // EVERY contiguous multi-token prefix shorter than the name (SPEC §7.4):
+    // "DA SEO Digital Studio" yields "DA SEO" and "DA SEO Digital". A single-token
+    // prefix is deliberately not a rung — it is a common word.
+    for (let length = 2; length < tokens.length; length++) {
+        pushUnique(partials, used, tokens.slice(0, length).join(" "));
+    }
+    const targets = [];
+    const seenTargets = new Set();
+    const addTarget = (url, kind) => {
+        if (url === undefined)
+            return;
+        const host = hostOf(url);
+        const path = pathOf(url);
+        if (host === undefined || path === undefined)
+            return;
+        const key = `${host}${path}`;
+        if (seenTargets.has(key))
+            return;
+        seenTargets.add(key);
+        targets.push({ host, path, kind });
+    };
+    // The entity's own page first: a citation of it is still a citation of the
+    // entity, and the reason says which authority the target came from.
+    addTarget(entityId(identity), "identity");
+    for (const url of sameAsUrls(profile))
+        addTarget(url, "sameAs");
+    const knownTerms = new Set();
+    for (const term of [name, ...aliases, ...handles]) {
+        const key = normalizeText(term);
+        if (key !== "")
+            knownTerms.add(key);
+    }
+    // Qualifiers are declared, never inferred: the location and role titles
+    // geoCheck's name-collision already reads (one home) plus the entity's other
+    // declared terms.
+    const qualifiers = [];
+    const seenQualifiers = new Set();
+    for (const q of disambiguationQualifiers(profile))
+        pushUnique(qualifiers, seenQualifiers, q);
+    for (const term of [name, ...aliases, ...handles])
+        pushUnique(qualifiers, seenQualifiers, term);
+    return {
+        public: {
+            name,
+            aliases,
+            handles,
+            partials,
+            hosts: [...new Set(targets.map((t) => t.host))],
+        },
+        targets,
+        qualifiers,
+        knownTerms,
+    };
+}
+/** The public vocabulary of a workspace (the `entity` block of `--json`). */
+export function citeVocabulary(profile) {
+    return buildVocabulary(profile).public;
+}
+/** A sentence's `[start, end)` bounds around one offset, in the original text. */
+function sentenceBounds(text, index) {
+    const breaks = new Set([".", "!", "?", ";", "\n", "\u2026"]);
+    let start = 0;
+    for (let i = index - 1; i >= 0; i--) {
+        if (breaks.has(text[i] ?? "")) {
+            start = i + 1;
+            break;
+        }
+    }
+    let end = text.length;
+    for (let i = index; i < text.length; i++) {
+        if (breaks.has(text[i] ?? "")) {
+            end = i;
+            break;
+        }
+    }
+    return { start, end };
+}
+/**
+ * The longer capitalised phrase the occurrence sits inside — the text's own
+ * competing reading (`DA SEO` → `DA SEO Digital`). Expansion is by adjacent
+ * tokens that start with an uppercase letter; it stops at punctuation, so
+ * `Pasco, WA` does not expand.
+ */
+function capitalRun(text, start, end) {
+    const TOKEN_CHAR = /[\p{L}\p{N}'\u2019-]/u;
+    const UPPER = /^\p{Lu}/u;
+    let runStart = start;
+    let runEnd = end;
+    for (;;) {
+        if (text[runEnd] !== " ")
+            break;
+        const first = runEnd + 1;
+        if (!UPPER.test(text[first] ?? ""))
+            break;
+        let tokenEnd = first;
+        while (tokenEnd < text.length && TOKEN_CHAR.test(text[tokenEnd] ?? ""))
+            tokenEnd++;
+        runEnd = tokenEnd;
+    }
+    for (;;) {
+        if (runStart === 0 || text[runStart - 1] !== " ")
+            break;
+        const tokenEnd = runStart - 1;
+        let tokenStart = tokenEnd;
+        while (tokenStart > 0 && TOKEN_CHAR.test(text[tokenStart - 1] ?? ""))
+            tokenStart--;
+        // A capital that only opens a sentence ("Follow daseo") is not a longer
+        // proper noun; expanding over it would manufacture a competitor.
+        if (isSentenceInitial(text, tokenStart))
+            break;
+        const token = text.slice(tokenStart, tokenEnd);
+        if (token === "" || !UPPER.test(token))
+            break;
+        runStart = tokenStart;
+    }
+    return runStart === start && runEnd === end ? undefined : { start: runStart, end: runEnd };
+}
+/** Is the token at `index` the first word of its sentence? */
+function isSentenceInitial(text, index) {
+    const breaks = new Set([".", "!", "?", ";", "\n", "\u2026"]);
+    let i = index - 1;
+    while (i >= 0 && (text[i] === " " || text[i] === '""' || text[i] === "'"))
+        i--;
+    return i < 0 || breaks.has(text[i] ?? "");
+}
+/**
+ * Match one saved answer against the workspace's declared entity (SPEC §7.4).
+ * Pure: the same profile + answer bytes always produce the same result.
+ */
+export function citeMatch(profile, answer) {
+    const vocab = buildVocabulary(profile);
+    const text = answer.answer;
+    const folded = foldForMatch(text);
+    const reasons = [];
+    // 1. Citation rung: a saved citation on a declared host needs no text match —
+    //    "cited without naming" is exactly this case.
+    const citationSpans = [];
+    const seenUrls = new Set();
+    for (const citation of answer.citations) {
+        const host = hostOf(citation.url);
+        const path = pathOf(citation.url);
+        if (host === undefined || path === undefined) {
+            // A citation the URL parser rejects (a missing scheme is the common case)
+            // is NOT a silent "not cited": it is named, and the scheme is never
+            // guessed — a dropped citation would change the rate (SPEC §7.4 limit f).
+            reasons.push(`url-invalid:${citation.url}`);
+            continue;
+        }
+        const declared = vocab.targets.find((target) => sameSite(host, target.host) && pathMatches(path, target.path));
+        if (declared === undefined)
+            continue;
+        if (!seenUrls.has(citation.url)) {
+            seenUrls.add(citation.url);
+            citationSpans.push({ start: 0, end: 0, kind: "citation-url", url: citation.url });
+        }
+        reasons.push("citation-url", declared.kind === "identity" ? "identity-domain" : "sameAs-domain");
+    }
+    // 2. Text rungs: every word-bounded occurrence of every declared term.
+    const raw = [];
+    const termLists = {
+        name: [vocab.public.name],
+        alias: vocab.public.aliases,
+        handle: vocab.public.handles,
+        partial: vocab.public.partials,
+    };
+    for (const kind of RUNG_ORDER) {
+        const list = termLists[kind];
+        for (const term of list) {
+            for (const occurrence of findTerm(folded, term)) {
+                raw.push({
+                    span: { start: occurrence.start, end: occurrence.end, kind },
+                    kind: kind,
+                    term,
+                    reasons: [RUNG_REASON[kind]],
+                    cited: true,
+                });
+            }
+        }
+    }
+    // 3. Containment: the longer declared term wins; the shorter keeps its reason
+    //    but is not a second hit ("DA SEO" inside "DA SEO Digital").
+    const ordered = [...raw].sort((a, b) => b.span.end - b.span.start - (a.span.end - a.span.start) ||
+        RUNG_RANK[a.kind] - RUNG_RANK[b.kind] ||
+        a.span.start - b.span.start ||
+        a.span.end - b.span.end);
+    const kept = [];
+    for (const hit of ordered) {
+        const sameSpan = kept.find((k) => k.span.start === hit.span.start && k.span.end === hit.span.end);
+        // Same text matched by two rungs: the higher rung already reports it, so the
+        // duplicate is dropped without a `longer-name` claim (nothing was shorter).
+        if (sameSpan !== undefined)
+            continue;
+        const container = kept.find((k) => k.span.start <= hit.span.start &&
+            k.span.end >= hit.span.end &&
+            k.span.end - k.span.start > hit.span.end - hit.span.start);
+        if (container !== undefined) {
+            reasons.push(`longer-name:${container.term}`);
+            continue;
+        }
+        kept.push(hit);
+    }
+    // 4. Guards — each one says why a hit is NOT a citation.
+    for (const hit of kept) {
+        const sentence = sentenceBounds(text, hit.span.start);
+        const before = foldForMatch(text.slice(sentence.start, hit.span.start));
+        const sentenceText = foldForMatch(text.slice(sentence.start, sentence.end));
+        const contrast = CONTRAST_MARKERS.find((marker) => termOccursIn(before, marker, 0, before.text.length));
+        if (contrast !== undefined) {
+            hit.reasons.push("contrast-context");
+            hit.cited = false;
+        }
+        // The one-word guard is the ambiguity decision. `competing` does NOT decide
+        // anything on its own: it only NAMES the longer reading the text carries
+        // once the hit is already ambiguous, so ordinary prose ("In Ecuador Pasco
+        // is a CLI", "… Studio Founder Ana ships") can never be turned into a
+        // competitor by an adjacent capital (review F1).
+        const singleWord = !/\s/.test(hit.term.trim());
+        if (singleWord && hit.kind !== "handle") {
+            const qualified = vocab.qualifiers.some((qualifier) => normalizeText(qualifier) !== normalizeText(hit.term) &&
+                termOccursIn(sentenceText, qualifier, 0, sentenceText.text.length));
+            if (!qualified) {
+                hit.reasons.push(`ambiguous-term:${normalizeText(hit.term)}`, "no-qualifier");
+                hit.cited = false;
+                const run = capitalRun(text, hit.span.start, hit.span.end);
+                if (run !== undefined) {
+                    const phrase = text.slice(run.start, run.end);
+                    if (!vocab.knownTerms.has(normalizeText(phrase))) {
+                        hit.reasons.push(`competing:${phrase}`);
+                    }
+                }
+            }
+        }
+    }
+    // 5. Project: citations first, then text spans by position; reasons as a set
+    //    in the fixed contract order.
+    const textSpans = [...kept]
+        .sort((a, b) => a.span.start - b.span.start ||
+        a.span.end - b.span.end ||
+        RUNG_RANK[a.kind] - RUNG_RANK[b.kind])
+        .map((hit) => hit.span);
+    for (const hit of kept)
+        reasons.push(...hit.reasons);
+    const citedHits = kept.filter((hit) => hit.cited);
+    const cited = citationSpans.length > 0 || citedHits.length > 0;
+    let confidence = "low";
+    if (citationSpans.length > 0 || citedHits.some((hit) => hit.kind === "name"))
+        confidence = "high";
+    else if (citedHits.some((hit) => hit.kind === "alias" || hit.kind === "handle")) {
+        confidence = "medium";
+    }
+    return {
+        cited,
+        confidence,
+        spans: [...citationSpans, ...textSpans],
+        reasons: orderReasons(reasons),
+    };
+}
+/** De-duplicate and order the reason set as the contract declares. */
+function orderReasons(reasons) {
+    const rank = (reason) => {
+        const index = REASON_ORDER.findIndex((head) => reason === head || reason.startsWith(`${head}:`));
+        return index === -1 ? REASON_ORDER.length : index;
+    };
+    return [...new Set(reasons)].sort((a, b) => rank(a) - rank(b) || byteOrder(a, b));
+}
+/** Exported for tests: the declared host of a URL under the matcher's rule. */
+export function citationHost(url) {
+    return hostOf(url);
+}
+//# sourceMappingURL=cite.js.map

package/dist/cite.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cite.js","sourceRoot":"","sources":["../src/cite.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,OAAO,EAAE,wBAAwB,EAAE,MAAM,UAAU,CAAC;AACpD,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,aAAa,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAC;AACrF,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAC;AA8E5D,MAAM,UAAU,GAA4B,CAAC,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;AAEnF,MAAM,SAAS,GAA2C;IACxD,cAAc,EAAE,CAAC;IACjB,IAAI,EAAE,CAAC;IACP,KAAK,EAAE,CAAC;IACR,MAAM,EAAE,CAAC;IACT,OAAO,EAAE,CAAC;CACX,CAAC;AAEF,MAAM,WAAW,GAA2C;IAC1D,cAAc,EAAE,cAAc;IAC9B,IAAI,EAAE,YAAY;IAClB,KAAK,EAAE,OAAO;IACd,MAAM,EAAE,QAAQ;IAChB,OAAO,EAAE,SAAS;CACnB,CAAC;AAEF;;;;GAIG;AACH,MAAM,gBAAgB,GAAsB;IAC1C,QAAQ;IACR,aAAa;IACb,eAAe;IACf,aAAa;IACb,YAAY;IACZ,QAAQ;IACR,IAAI;IACJ,aAAa;IACb,iBAAiB;IACjB,kBAAkB;IAClB,aAAa;IACb,WAAW;IACX,iBAAiB;IACjB,mBAAmB;CACpB,CAAC;AAEF,oEAAoE;AACpE,MAAM,YAAY,GAAsB;IACtC,cAAc;IACd,iBAAiB;IACjB,eAAe;IACf,aAAa;IACb,YAAY;IACZ,OAAO;IACP,QAAQ;IACR,SAAS;IACT,aAAa;IACb,WAAW;IACX,gBAAgB;IAChB,cAAc;IACd,kBAAkB;CACnB,CAAC;AAEF,oEAAoE;AACpE,SAAS,MAAM,CAAC,GAAW;IACzB,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC;QACjD,OAAO,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IACxD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,SAAS,QAAQ,CAAC,YAAoB,EAAE,YAAoB;IAC1D,OAAO,YAAY,KAAK,YAAY,IAAI,YAAY,CAAC,QAAQ,CAAC,IAAI,YAAY,EAAE,CAAC,CAAC;AACpF,CAAC;AAED,2EAA2E;AAC3E,SAAS,MAAM,CAAC,GAAW;IACzB,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IACnD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED;;;;;;GAMG;AACH,SAAS,WAAW,CAAC,YAAoB,EAAE,YAAoB;IAC7D,IAAI,YAAY,KAAK,EAAE;QAAE,OAAO,IAAI,CAAC;IACrC,OAAO,YAAY,KAAK,YAAY,IAAI,YAAY,CAAC,UAAU,CAAC,GAAG,YAAY,GAAG,CAAC,CAAC;AACtF,CAAC;AAED;;;;GAIG;AACH,SAAS,mBAAmB,CAAC,GAAW;IACtC,MAAM,MAAM,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACvC,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACzC,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC;IACpE,OAAO,OAAO,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;AACnD,CAAC;AAED,oFAAoF;AACpF,SAAS,UAAU,CAAC,GAAa,EAAE,IAAiB,EAAE,KAAa;IACjE,MAAM,GAAG,GAAG,aAAa,CAAC,KAAK,CAAC,CAAC;IACjC,IAAI,GAAG,KAAK,EAAE,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO;IACxC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACd,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAClB,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CAAC,OAAgB;IACvC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC;IAClC,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IAElC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAS,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACpD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC;IACzC,IAAI,OAAO,KAAK,SAAS,IAAI,OAAO,KAAK,EAAE;QAAE,UAAU,CAAC,OAAO,EAAE,IAAI,EAAE,OAAO,CAAC,CAAC;IAEhF,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC;IACvC,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;QAC1C,MAAM,MAAM,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;QAC3C,IAAI,MAAM,KAAK,SAAS;YAAE,UAAU,CAAC,OAAO,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;IAC9D,CAAC;IACD,KAAK,MAAM,GAAG,IAAI,OAAO,CAAC,aAAa,IAAI,EAAE,EAAE,CAAC;QAC9C,IAAI,GAAG,CAAC,MAAM,KAAK,SAAS;YAAE,SAAS;QACvC,UAAU,CAAC,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;IAC/C,CAAC;IAED,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IACjD,yEAAyE;IACzE,+EAA+E;IAC/E,2DAA2D;IAC3D,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,CAAC;QACtD,UAAU,CAAC,QAAQ,EAAE,IAAI,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAChE,CAAC;IAED,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,SAAS,GAAG,CAAC,GAAuB,EAAE,IAA4B,EAAQ,EAAE;QAChF,IAAI,GAAG,KAAK,SAAS;YAAE,OAAO;QAC9B,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QACzB,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QACzB,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO;QACrD,MAAM,GAAG,GAAG,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC;QAC7B,IAAI,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,OAAO;QACjC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACrB,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC;IACrC,CAAC,CAAC;IACF,2EAA2E;IAC3E,oEAAoE;IACpE,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,UAAU,CAAC,CAAC;IAC1C,KAAK,MAAM,GAAG,IAAI,UAAU,CAAC,OAAO,CAAC;QAAE,SAAS,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;IAEhE,MAAM,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;IACrC,KAAK,MAAM,IAAI,IAAI,CAAC,IAAI,EAAE,GAAG,OAAO,EAAE,GAAG,OAAO,CAAC,EAAE,CAAC;QAClD,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAChC,IAAI,GAAG,KAAK,EAAE;YAAE,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACtC,CAAC;IAED,wEAAwE;IACxE,6EAA6E;IAC7E,kBAAkB;IAClB,MAAM,UAAU,GAAa,EAAE,CAAC;IAChC,MAAM,cAAc,GAAG,IAAI,GAAG,EAAU,CAAC;IACzC,KAAK,MAAM,CAAC,IAAI,wBAAwB,CAAC,OAAO,CAAC;QAAE,UAAU,CAAC,UAAU,EAAE,cAAc,EAAE,CAAC,CAAC,CAAC;IAC7F,KAAK,MAAM,IAAI,IAAI,CAAC,IAAI,EAAE,GAAG,OAAO,EAAE,GAAG,OAAO,CAAC;QAAE,UAAU,CAAC,UAAU,EAAE,cAAc,EAAE,IAAI,CAAC,CAAC;IAEhG,OAAO;QACL,MAAM,EAAE;YACN,IAAI;YACJ,OAAO;YACP,OAAO;YACP,QAAQ;YACR,KAAK,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;SAChD;QACD,OAAO;QACP,UAAU;QACV,UAAU;KACX,CAAC;AACJ,CAAC;AAED,6EAA6E;AAC7E,MAAM,UAAU,cAAc,CAAC,OAAgB;IAC7C,OAAO,eAAe,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC;AACzC,CAAC;AAED,kFAAkF;AAClF,SAAS,cAAc,CAAC,IAAY,EAAE,KAAa;IACjD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC;IAC7D,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACpC,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,MAAM;QACR,CAAC;IACH,CAAC;IACD,IAAI,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;IACtB,KAAK,IAAI,CAAC,GAAG,KAAK,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;YAC9B,GAAG,GAAG,CAAC,CAAC;YACR,MAAM;QACR,CAAC;IACH,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC;AACxB,CAAC;AAED;;;;;GAKG;AACH,SAAS,UAAU,CACjB,IAAY,EACZ,KAAa,EACb,GAAW;IAEX,MAAM,UAAU,GAAG,uBAAuB,CAAC;IAC3C,MAAM,KAAK,GAAG,UAAU,CAAC;IACzB,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,IAAI,MAAM,GAAG,GAAG,CAAC;IAEjB,SAAS,CAAC;QACR,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG;YAAE,MAAM;QAChC,MAAM,KAAK,GAAG,MAAM,GAAG,CAAC,CAAC;QACzB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;YAAE,MAAM;QAC1C,IAAI,QAAQ,GAAG,KAAK,CAAC;QACrB,OAAO,QAAQ,GAAG,IAAI,CAAC,MAAM,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;YAAE,QAAQ,EAAE,CAAC;QACnF,MAAM,GAAG,QAAQ,CAAC;IACpB,CAAC;IAED,SAAS,CAAC;QACR,IAAI,QAAQ,KAAK,CAAC,IAAI,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC,KAAK,GAAG;YAAE,MAAM;QACxD,MAAM,QAAQ,GAAG,QAAQ,GAAG,CAAC,CAAC;QAC9B,IAAI,UAAU,GAAG,QAAQ,CAAC;QAC1B,OAAO,UAAU,GAAG,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YAAE,UAAU,EAAE,CAAC;QACnF,wEAAwE;QACxE,iEAAiE;QACjE,IAAI,iBAAiB,CAAC,IAAI,EAAE,UAAU,CAAC;YAAE,MAAM;QAC/C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;QAC/C,IAAI,KAAK,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,MAAM;QAC9C,QAAQ,GAAG,UAAU,CAAC;IACxB,CAAC;IAED,OAAO,QAAQ,KAAK,KAAK,IAAI,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,GAAG,EAAE,MAAM,EAAE,CAAC;AAC7F,CAAC;AAED,8DAA8D;AAC9D,SAAS,iBAAiB,CAAC,IAAY,EAAE,KAAa;IACpD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC;IAC7D,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;IAClB,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC;QAAE,CAAC,EAAE,CAAC;IAC/E,OAAO,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAC5C,CAAC;AAWD;;;GAGG;AACH,MAAM,UAAU,SAAS,CAAC,OAAgB,EAAE,MAAkB;IAC5D,MAAM,KAAK,GAAG,eAAe,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC;IAC3B,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,OAAO,GAAa,EAAE,CAAC;IAE7B,8EAA8E;IAC9E,kDAAkD;IAClD,MAAM,aAAa,GAAe,EAAE,CAAC;IACrC,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IACnC,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;QACxC,MAAM,IAAI,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;QAClC,MAAM,IAAI,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;QAClC,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7C,0EAA0E;YAC1E,oEAAoE;YACpE,0EAA0E;YAC1E,OAAO,CAAC,IAAI,CAAC,eAAe,QAAQ,CAAC,GAAG,EAAE,CAAC,CAAC;YAC5C,SAAS;QACX,CAAC;QACD,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CACjC,CAAC,MAAM,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,WAAW,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAC1E,CAAC;QACF,IAAI,QAAQ,KAAK,SAAS;YAAE,SAAS;QACrC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;YAChC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;YAC3B,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,IAAI,EAAE,cAAc,EAAE,GAAG,EAAE,QAAQ,CAAC,GAAG,EAAE,CAAC,CAAC;QACpF,CAAC;QACD,OAAO,CAAC,IAAI,CACV,cAAc,EACd,QAAQ,CAAC,IAAI,KAAK,UAAU,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,eAAe,CACnE,CAAC;IACJ,CAAC;IAED,uEAAuE;IACvE,MAAM,GAAG,GAAc,EAAE,CAAC;IAC1B,MAAM,SAAS,GAA4D;QACzE,IAAI,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC;QACzB,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,OAAO;QAC3B,MAAM,EAAE,KAAK,CAAC,MAAM,CAAC,OAAO;QAC5B,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,QAAQ;KAC/B,CAAC;IACF,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,MAAM,IAAI,GAAG,SAAS,CAAC,IAA6C,CAAC,CAAC;QACtE,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACxB,KAAK,MAAM,UAAU,IAAI,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,CAAC;gBAChD,GAAG,CAAC,IAAI,CAAC;oBACP,IAAI,EAAE,EAAE,KAAK,EAAE,UAAU,CAAC,KAAK,EAAE,GAAG,EAAE,UAAU,CAAC,GAAG,EAAE,IAAI,EAAE;oBAC5D,IAAI,EAAE,IAA6C;oBACnD,IAAI;oBACJ,OAAO,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;oBAC5B,KAAK,EAAE,IAAI;iBACZ,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,8EAA8E;IAC9E,iEAAiE;IACjE,MAAM,OAAO,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAC3B,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACP,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC;QACvD,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;QACrC,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK;QAC3B,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,CAC1B,CAAC;IACF,MAAM,IAAI,GAAc,EAAE,CAAC;IAC3B,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;QAC1B,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CACxB,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,KAAK,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,IAAI,CAAC,GAAG,CACtE,CAAC;QACF,6EAA6E;QAC7E,4EAA4E;QAC5E,IAAI,QAAQ,KAAK,SAAS;YAAE,SAAS;QACrC,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CACzB,CAAC,CAAC,EAAE,EAAE,CACJ,CAAC,CAAC,IAAI,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK;YAC9B,CAAC,CAAC,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG;YAC1B,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC,IAAI,CAAC,KAAK,CAC5D,CAAC;QACF,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;YAC5B,OAAO,CAAC,IAAI,CAAC,eAAe,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC;YAC9C,SAAS;QACX,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACjB,CAAC;IAED,yDAAyD;IACzD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,QAAQ,GAAG,cAAc,CAAC,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACtD,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QACxE,MAAM,YAAY,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC;QAE5E,MAAM,QAAQ,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAChD,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CACpD,CAAC;QACF,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC;YACrC,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC;QACpB,CAAC;QAED,4EAA4E;QAC5E,yEAAyE;QACzE,0EAA0E;QAC1E,sEAAsE;QACtE,iDAAiD;QACjD,MAAM,UAAU,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;QAC/C,IAAI,UAAU,IAAI,GAAG,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YACxC,MAAM,SAAS,GAAG,KAAK,CAAC,UAAU,CAAC,IAAI,CACrC,CAAC,SAAS,EAAE,EAAE,CACZ,aAAa,CAAC,SAAS,CAAC,KAAK,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC;gBACpD,YAAY,CAAC,YAAY,EAAE,SAAS,EAAE,CAAC,EAAE,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CACrE,CAAC;YACF,IAAI,CAAC,SAAS,EAAE,CAAC;gBACf,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,kBAAkB,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,EAAE,cAAc,CAAC,CAAC;gBAC9E,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC;gBAClB,MAAM,GAAG,GAAG,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC3D,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;oBACtB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC;oBAC9C,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,aAAa,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC;wBACjD,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,aAAa,MAAM,EAAE,CAAC,CAAC;oBAC1C,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,6EAA6E;IAC7E,kCAAkC;IAClC,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,CAAC;SACxB,IAAI,CACH,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACP,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK;QAC3B,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG;QACvB,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CACxC;SACA,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAE1B,KAAK,MAAM,GAAG,IAAI,IAAI;QAAE,OAAO,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,CAAC;IAErD,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAClD,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC;IAC/D,IAAI,UAAU,GAAmB,KAAK,CAAC;IACvC,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,KAAK,MAAM,CAAC;QAAE,UAAU,GAAG,MAAM,CAAC;SAC7F,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,KAAK,OAAO,IAAI,GAAG,CAAC,IAAI,KAAK,QAAQ,CAAC,EAAE,CAAC;QAChF,UAAU,GAAG,QAAQ,CAAC;IACxB,CAAC;IAED,OAAO;QACL,KAAK;QACL,UAAU;QACV,KAAK,EAAE,CAAC,GAAG,aAAa,EAAE,GAAG,SAAS,CAAC;QACvC,OAAO,EAAE,YAAY,CAAC,OAAO,CAAC;KAC/B,CAAC;AACJ,CAAC;AAED,sEAAsE;AACtE,SAAS,YAAY,CAAC,OAAiB;IACrC,MAAM,IAAI,GAAG,CAAC,MAAc,EAAU,EAAE;QACtC,MAAM,KAAK,GAAG,YAAY,CAAC,SAAS,CAClC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,KAAK,IAAI,IAAI,MAAM,CAAC,UAAU,CAAC,GAAG,IAAI,GAAG,CAAC,CAC3D,CAAC;QACF,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC;IACpD,CAAC,CAAC;IACF,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;AACpF,CAAC;AAED,+EAA+E;AAC/E,MAAM,UAAU,YAAY,CAAC,GAAW;IACtC,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC;AACrB,CAAC"}

package/dist/completeness.d.ts

@@ -14,7 +14,7 @@
  *   evidence     claims that carry a source + registry entries (`evidenceItem`)
  *   deployment   share of trackable units with a ledger record (`ledger`)
  *
- * Policy is parameterized (dev-methodology §3): a user says "my target is the
+ * Policy is parameterized: a user says "my target is the
  * GitHub milestone set, not everything" by dropping an item (weight 0) — a
  * parameter, never a code path. No weights are hardcoded outside the catalog.
  */

package/dist/completeness.js

@@ -14,7 +14,7 @@
  *   evidence     claims that carry a source + registry entries (`evidenceItem`)
  *   deployment   share of trackable units with a ledger record (`ledger`)
  *
- * Policy is parameterized (dev-methodology §3): a user says "my target is the
+ * Policy is parameterized: a user says "my target is the
  * GitHub milestone set, not everything" by dropping an item (weight 0) — a
  * parameter, never a code path. No weights are hardcoded outside the catalog.
  */

package/dist/generated/ai-policy.gen.d.ts

@@ -0,0 +1,36 @@
+/**
+ * GENERATED by scripts/gen-platforms/gen.mjs from platforms/registry.yaml — DO NOT EDIT BY HAND (#290).
+ *
+ * The AI crawler policy has exactly one data home: the `ai:` block of
+ * `platforms/registry.yaml`, where every entry names the operator's own page
+ * and the date it was read. `pasco export ai-policy` renders these rows into
+ * robots.txt / llms.txt / AI-POLICY.md; a UA list kept in the CLI would rot
+ * without a failing command. Run `pnpm gen:platforms` after editing the
+ * registry; CI asserts this file is fresh.
+ */
+/** The two robots.txt groups the policy is expressed as — never one wildcard. */
+export type AiCrawlerFamily = "retrieval" | "training";
+/** One robots.txt token and the evidence behind its classification. */
+export interface AiCrawler {
+    /** Registry slug — the entry a reader is pointed at. */
+    readonly slug: string;
+    /** The robots.txt `User-agent` token, verbatim. */
+    readonly ua: string;
+    /** retrieval = the output is a citation · training = foundation-model corpus. */
+    readonly family: AiCrawlerFamily;
+    /** The company/foundation that runs it. */
+    readonly operator: string;
+    /** That operator's own page documenting the token. */
+    readonly source: string;
+    /** ISO date `source` was read (the `--check` staleness anchor). */
+    readonly as_of: string;
+}
+/** The policy contract version (a consumer refuses a version it does not know). */
+export declare const AI_POLICY_VERSION = 1;
+/** The date the policy was last reviewed (the generated files' dated stamp). */
+export declare const AI_POLICY_AS_OF = "2026-10-27";
+/** The stated position, verbatim from the SPEC §12 M9 floor (gate-checked). */
+export declare const AI_POLICY_STANCE = "**private-first default**: `robots.txt` on published surfaces **blocks** AI-training crawlers (CCBot / vendor AI crawlers) and user data ships **without** an open license; entering corpora or open-licensing data requires the profile owner's explicit opt-in (consent field \u2014 schema work deferred with Q10).";
+/** The crawler rows, in registry order (retrieval first, then training). */
+export declare const AI_CRAWLERS: readonly AiCrawler[];
+//# sourceMappingURL=ai-policy.gen.d.ts.map

package/dist/generated/ai-policy.gen.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ai-policy.gen.d.ts","sourceRoot":"","sources":["../../src/generated/ai-policy.gen.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,iFAAiF;AACjF,MAAM,MAAM,eAAe,GAAG,WAAW,GAAG,UAAU,CAAC;AAEvD,uEAAuE;AACvE,MAAM,WAAW,SAAS;IACxB,wDAAwD;IACxD,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,mDAAmD;IACnD,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAC;IACpB,iFAAiF;IACjF,QAAQ,CAAC,MAAM,EAAE,eAAe,CAAC;IACjC,2CAA2C;IAC3C,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,sDAAsD;IACtD,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,mEAAmE;IACnE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;CACxB;AAED,mFAAmF;AACnF,eAAO,MAAM,iBAAiB,IAAI,CAAC;AAEnC,gFAAgF;AAChF,eAAO,MAAM,eAAe,eAAe,CAAC;AAE5C,+EAA+E;AAC/E,eAAO,MAAM,gBAAgB,2TAAsT,CAAC;AAEpV,4EAA4E;AAC5E,eAAO,MAAM,WAAW,EAAE,SAAS,SAAS,EAY3C,CAAC"}

package/dist/generated/ai-policy.gen.js

@@ -0,0 +1,31 @@
+/**
+ * GENERATED by scripts/gen-platforms/gen.mjs from platforms/registry.yaml — DO NOT EDIT BY HAND (#290).
+ *
+ * The AI crawler policy has exactly one data home: the `ai:` block of
+ * `platforms/registry.yaml`, where every entry names the operator's own page
+ * and the date it was read. `pasco export ai-policy` renders these rows into
+ * robots.txt / llms.txt / AI-POLICY.md; a UA list kept in the CLI would rot
+ * without a failing command. Run `pnpm gen:platforms` after editing the
+ * registry; CI asserts this file is fresh.
+ */
+/** The policy contract version (a consumer refuses a version it does not know). */
+export const AI_POLICY_VERSION = 1;
+/** The date the policy was last reviewed (the generated files' dated stamp). */
+export const AI_POLICY_AS_OF = "2026-10-27";
+/** The stated position, verbatim from the SPEC §12 M9 floor (gate-checked). */
+export const AI_POLICY_STANCE = "**private-first default**: `robots.txt` on published surfaces **blocks** AI-training crawlers (CCBot / vendor AI crawlers) and user data ships **without** an open license; entering corpora or open-licensing data requires the profile owner's explicit opt-in (consent field — schema work deferred with Q10).";
+/** The crawler rows, in registry order (retrieval first, then training). */
+export const AI_CRAWLERS = [
+    { slug: "googlebot", ua: "Googlebot", family: "retrieval", operator: "Google", source: "https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers", as_of: "2026-10-27" },
+    { slug: "oai-searchbot", ua: "OAI-SearchBot", family: "retrieval", operator: "OpenAI", source: "https://developers.openai.com/api/docs/bots", as_of: "2026-10-27" },
+    { slug: "chatgpt-user", ua: "ChatGPT-User", family: "retrieval", operator: "OpenAI", source: "https://developers.openai.com/api/docs/bots", as_of: "2026-10-27" },
+    { slug: "perplexitybot", ua: "PerplexityBot", family: "retrieval", operator: "Perplexity", source: "https://docs.perplexity.ai/docs/resources/perplexity-crawlers", as_of: "2026-10-27" },
+    { slug: "perplexity-user", ua: "Perplexity-User", family: "retrieval", operator: "Perplexity", source: "https://docs.perplexity.ai/docs/resources/perplexity-crawlers", as_of: "2026-10-27" },
+    { slug: "claude-searchbot", ua: "Claude-SearchBot", family: "retrieval", operator: "Anthropic", source: "https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler", as_of: "2026-10-27" },
+    { slug: "claude-user", ua: "Claude-User", family: "retrieval", operator: "Anthropic", source: "https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler", as_of: "2026-10-27" },
+    { slug: "gptbot", ua: "GPTBot", family: "training", operator: "OpenAI", source: "https://developers.openai.com/api/docs/bots", as_of: "2026-10-27" },
+    { slug: "claudebot", ua: "ClaudeBot", family: "training", operator: "Anthropic", source: "https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler", as_of: "2026-10-27" },
+    { slug: "google-extended", ua: "Google-Extended", family: "training", operator: "Google", source: "https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers", as_of: "2026-10-27" },
+    { slug: "ccbot", ua: "CCBot", family: "training", operator: "Common Crawl", source: "https://commoncrawl.org/ccbot", as_of: "2026-10-27" },
+];
+//# sourceMappingURL=ai-policy.gen.js.map