@mailwoman/resolver-wof-sqlite 9.1.0 → 9.3.0

package/out/candidate/ancestors-sidecar.js

@@ -0,0 +1,140 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 3b of the candidate build — the containment sidecar (closure rows + interval labels).
+ */
+import { placetypeDepth } from "#ancestry/index";
+import { CANDIDATE_ANCESTOR_COLUMNS, CANDIDATE_ANCESTOR_TABLE, CANDIDATE_INTERVAL_TABLE, createCandidateAncestorTable, createCandidateIntervalTable, MAX_ANCESTOR_DEPTH, } from "#candidate-ancestors-schema";
+/**
+ * Pass 3b — the ancestors sidecar: closure rows + interval labels (candidate-ancestors-schema.ts owns the encoding
+ * decision and the DAG/absence semantics). Reads the same source `ancestors` table the region stamp reads,
+ * denormalizing each edge with the parent's name/key from `attrs`, streamed `ORDER BY id` so the clustered `(spr_id,
+ * depth)` insert is sorted — the contiguous-leaves discipline of the candidate table itself.
+ *
+ * Excluded by policy: self rows, and placetypes outside the containment ladder (continent, empire, …: `placetypeDepth` 0)
+ * — they discriminate nothing a consumer of this sidecar checks. An edge to a parent with no current `spr` row has no
+ * name to denormalize; it is dropped and counted rather than stored blind.
+ */
+export async function buildAncestorsSidecar(ctx) {
+    const { src, out, attrs, ptID, progress } = ctx;
+    progress("ancestors", "building containment sidecar (closure rows + interval labels)");
+    await createCandidateAncestorTable(ctx.out);
+    await createCandidateIntervalTable(ctx.out);
+    const insAncestor = out.prepare(`INSERT INTO ${CANDIDATE_ANCESTOR_TABLE} VALUES (${CANDIDATE_ANCESTOR_COLUMNS.map(() => "?").join(", ")})`);
+    // The canonical-parent forest the interval labels are computed over. One parent per place — the
+    // depth-1 edge (finest containment tier, lowest ancestor id; the `regionOf` MIN-stability
+    // convention). ALL parents stay in the closure rows; only the interval tree canonicalizes.
+    const canonicalParentOf = new Map();
+    const childrenOf = new Map();
+    const forest = new Set();
+    let ancestorRows = 0;
+    let ancestorPlaces = 0;
+    let droppedParents = 0;
+    // Per-child edge buffer; the stream below is grouped by child id, so each flush owns one place.
+    let childID = -1;
+    let edges = [];
+    const flush = () => {
+        if (childID < 0 || !edges.length)
+            return;
+        // Deterministic nearest-first: containment depth descending, then ancestor id ascending —
+        // the FTS backend's `ancestorLineage` ordering, made stable across rebuilds.
+        edges.sort((a, b) => placetypeDepth(b.apt) - placetypeDepth(a.apt) || a.aid - b.aid);
+        if (edges.length > MAX_ANCESTOR_DEPTH) {
+            edges = edges.slice(0, MAX_ANCESTOR_DEPTH);
+        }
+        ancestorPlaces++;
+        for (const [i, edge] of edges.entries()) {
+            const parent = attrs.get(edge.aid);
+            insAncestor.run(childID, i + 1, edge.aid, ptID(edge.apt), parent.name, parent.pkey);
+            ancestorRows++;
+        }
+        const canonical = edges[0].aid;
+        canonicalParentOf.set(childID, canonical);
+        const siblings = childrenOf.get(canonical);
+        if (siblings) {
+            siblings.push(childID);
+        }
+        else {
+            childrenOf.set(canonical, [childID]);
+        }
+        forest.add(childID);
+        forest.add(canonical);
+    };
+    out.exec("BEGIN");
+    for (const r of src
+        .prepare("SELECT id, ancestor_id, ancestor_placetype FROM ancestors WHERE ancestor_id != id ORDER BY id")
+        .iterate()) {
+        const id = Number(r.id);
+        if (id !== childID) {
+            flush();
+            childID = id;
+            edges = [];
+        }
+        if (!attrs.has(id))
+            continue;
+        const apt = String(r.ancestor_placetype ?? "");
+        if (placetypeDepth(apt) === 0)
+            continue;
+        const aid = Number(r.ancestor_id);
+        if (!attrs.has(aid)) {
+            droppedParents++;
+            continue;
+        }
+        edges.push({ aid, apt });
+    }
+    flush();
+    out.exec("COMMIT");
+    // Interval labels: pre/post-order DFS over the canonical-parent forest. Root order and child
+    // order are id-ascending so the labels are stable across rebuilds of the same source.
+    const preOf = new Map();
+    const postOf = new Map();
+    for (const kids of childrenOf.values()) {
+        // oxlint-disable-next-line unicorn/no-array-sort -- sorts an array this pass just built
+        kids.sort((a, b) => a - b);
+    }
+    const roots = [...forest].filter((id) => !canonicalParentOf.has(id));
+    // oxlint-disable-next-line unicorn/no-array-sort -- sorts an array this pass just built
+    roots.sort((a, b) => a - b);
+    let counter = 0;
+    for (const root of roots) {
+        preOf.set(root, counter++);
+        const stack = [{ id: root, next: 0 }];
+        while (stack.length) {
+            const top = stack.at(-1);
+            const kids = childrenOf.get(top.id);
+            if (kids && top.next < kids.length) {
+                const kid = kids[top.next++];
+                // Each child holds exactly one canonical parent, so a labeled node here means the
+                // grouping upstream broke — skip rather than corrupt the numbering.
+                if (preOf.has(kid))
+                    continue;
+                preOf.set(kid, counter++);
+                stack.push({ id: kid, next: 0 });
+            }
+            else {
+                postOf.set(top.id, counter++);
+                stack.pop();
+            }
+        }
+    }
+    // A canonical-parent CYCLE (corrupt source ancestry) leaves its members unreachable from any
+    // root: they simply receive no label, and containment against them reads unverifiable — the
+    // absence semantics the schema module states. Counted so a jump is visible across rebuilds.
+    const cycleSkipped = forest.size - preOf.size;
+    const insInterval = out.prepare(`INSERT INTO ${CANDIDATE_INTERVAL_TABLE} VALUES (?, ?, ?)`);
+    const labeled = [...preOf.keys()];
+    // oxlint-disable-next-line unicorn/no-array-sort -- sorts an array this pass just built
+    labeled.sort((a, b) => a - b);
+    out.exec("BEGIN");
+    for (const id of labeled) {
+        insInterval.run(id, preOf.get(id), postOf.get(id));
+    }
+    out.exec("COMMIT");
+    progress("ancestors", `${ancestorRows.toLocaleString()} closure rows across ${ancestorPlaces.toLocaleString()} places; ` +
+        `${preOf.size.toLocaleString()} interval labels` +
+        (droppedParents ? `; ${droppedParents.toLocaleString()} edges dropped (parent has no current spr row)` : "") +
+        (cycleSkipped ? `; ${cycleSkipped.toLocaleString()} places skipped (canonical-parent cycle)` : ""));
+    return { ancestorRows, ancestorPlaces, intervalPlaces: preOf.size };
+}
+//# sourceMappingURL=ancestors-sidecar.js.map

package/out/candidate/ancestors-sidecar.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ancestors-sidecar.js","sourceRoot":"","sources":["../../lib/candidate/ancestors-sidecar.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAIH,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAA;AAChD,OAAO,EACN,0BAA0B,EAC1B,wBAAwB,EACxB,wBAAwB,EACxB,4BAA4B,EAC5B,4BAA4B,EAC5B,kBAAkB,GAClB,MAAM,6BAA6B,CAAA;AAKpC;;;;;;;;;GASG;AACH,MAAM,CAAC,KAAK,UAAU,qBAAqB,CAAC,GAM3C;IACA,MAAM,EAAE,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,GAAG,CAAA;IAE/C,QAAQ,CAAC,WAAW,EAAE,+DAA+D,CAAC,CAAA;IACtF,MAAM,4BAA4B,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;IAC3C,MAAM,4BAA4B,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;IAE3C,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAC9B,eAAe,wBAAwB,YAAY,0BAA0B,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAC1G,CAAA;IAED,gGAAgG;IAChG,0FAA0F;IAC1F,2FAA2F;IAC3F,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAkB,CAAA;IACnD,MAAM,UAAU,GAAG,IAAI,GAAG,EAAoB,CAAA;IAC9C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAA;IAEhC,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,IAAI,cAAc,GAAG,CAAC,CAAA;IACtB,IAAI,cAAc,GAAG,CAAC,CAAA;IAEtB,gGAAgG;IAChG,IAAI,OAAO,GAAG,CAAC,CAAC,CAAA;IAChB,IAAI,KAAK,GAAwC,EAAE,CAAA;IAEnD,MAAM,KAAK,GAAG,GAAS,EAAE;QACxB,IAAI,OAAO,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM;YAAE,OAAM;QAExC,0FAA0F;QAC1F,6EAA6E;QAC7E,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,cAAc,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QAEpF,IAAI,KAAK,CAAC,MAAM,GAAG,kBAAkB,EAAE,CAAC;YACvC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAA;QAC3C,CAAC;QAED,cAAc,EAAE,CAAA;QAEhB,KAAK,MAAM,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;YACzC,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAE,CAAA;YAEnC,WAAW,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAA;YAEnF,YAAY,EAAE,CAAA;QACf,CAAC;QAED,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,GAAG,CAAA;QAE/B,iBAAiB,CAAC,GAAG,CAAC,OAAO,EAAE,SAAS,CAAC,CAAA;QAEzC,MAAM,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;QAE1C,IAAI,QAAQ,EAAE,CAAC;YACd,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QACvB,CAAC;aAAM,CAAC;YACP,UAAU,CAAC,GAAG,CAAC,SAAS,EAAE,CAAC,OAAO,CAAC,CAAC,CAAA;QACrC,CAAC;QAED,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;QACnB,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;IACtB,CAAC,CAAA;IAED,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,IAAI,GAAG;SACjB,OAAO,CAAC,+FAA+F,CAAC;SACxG,OAAO,EAAE,EAAE,CAAC;QACb,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;QAEvB,IAAI,EAAE,KAAK,OAAO,EAAE,CAAC;YACpB,KAAK,EAAE,CAAA;YACP,OAAO,GAAG,EAAE,CAAA;YACZ,KAAK,GAAG,EAAE,CAAA;QACX,CAAC;QAED,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YAAE,SAAQ;QAE5B,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,kBAAkB,IAAI,EAAE,CAAC,CAAA;QAE9C,IAAI,cAAc,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,SAAQ;QAEvC,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,WAAW,CAAC,CAAA;QAEjC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACrB,cAAc,EAAE,CAAA;YAEhB,SAAQ;QACT,CAAC;QAED,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAA;IACzB,CAAC;IAED,KAAK,EAAE,CAAA;IACP,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAElB,6FAA6F;IAC7F,sFAAsF;IACtF,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAA;IACvC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAA;IAExC,KAAK,MAAM,IAAI,IAAI,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC;QACxC,wFAAwF;QACxF,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAC3B,CAAC;IAED,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,iBAAiB,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAA;IAEpE,wFAAwF;IACxF,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAE3B,IAAI,OAAO,GAAG,CAAC,CAAA;IAEf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QAC1B,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC,CAAA;QAC1B,MAAM,KAAK,GAAwC,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAA;QAE1E,OAAO,KAAK,CAAC,MAAM,EAAE,CAAC;YACrB,MAAM,GAAG,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,CAAA;YACzB,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;YAEnC,IAAI,IAAI,IAAI,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;gBACpC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAE,CAAA;gBAE7B,kFAAkF;gBAClF,oEAAoE;gBACpE,IAAI,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC;oBAAE,SAAQ;gBAE5B,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,EAAE,CAAC,CAAA;gBACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAA;YACjC,CAAC;iBAAM,CAAC;gBACP,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,OAAO,EAAE,CAAC,CAAA;gBAC7B,KAAK,CAAC,GAAG,EAAE,CAAA;YACZ,CAAC;QACF,CAAC;IACF,CAAC;IAED,6FAA6F;IAC7F,4FAA4F;IAC5F,4FAA4F;IAC5F,MAAM,YAAY,GAAG,MAAM,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;IAE7C,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAAC,eAAe,wBAAwB,mBAAmB,CAAC,CAAA;IAC3F,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,CAAA;IAEjC,wFAAwF;IACxF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAE7B,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEjB,KAAK,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;QAC1B,WAAW,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,GAAG,CAAC,EAAE,CAAE,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAE,CAAC,CAAA;IACrD,CAAC;IAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAElB,QAAQ,CACP,WAAW,EACX,GAAG,YAAY,CAAC,cAAc,EAAE,wBAAwB,cAAc,CAAC,cAAc,EAAE,WAAW;QACjG,GAAG,KAAK,CAAC,IAAI,CAAC,cAAc,EAAE,kBAAkB;QAChD,CAAC,cAAc,CAAC,CAAC,CAAC,KAAK,cAAc,CAAC,cAAc,EAAE,gDAAgD,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5G,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,YAAY,CAAC,cAAc,EAAE,0CAA0C,CAAC,CAAC,CAAC,EAAE,CAAC,CACnG,CAAA;IAED,OAAO,EAAE,YAAY,EAAE,cAAc,EAAE,cAAc,EAAE,KAAK,CAAC,IAAI,EAAE,CAAA;AACpE,CAAC"}

package/out/candidate/country-display-names.d.ts

@@ -0,0 +1,35 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 1b of the candidate build — fold ICU's country display names onto the country rows.
+ */
+import type { PlaceAttrs, StageRow } from "#candidate/place-attrs";
+/**
+ * Fold every country surface ICU knows onto that country's candidate row (#1678 thread 1).
+ *
+ * A bare `格鲁吉亚` (Georgia the country) resolved to NOTHING while `佐治亚州` (Georgia the US state) resolved correctly, and
+ * the model gave both the same wrong `locality` tag — so the tag was never the variable. Measured 2026-08-15: 140 of
+ * 237 country rows are synthetic and carry a canonical English name and nothing else; WOF holds no Chinese country
+ * names at all; and the GeoNames alias fold filters through a Latin-script regex, so neither existing source could ever
+ * supply them.
+ *
+ * `Intl.DisplayNames` already knows every one — ~280 regions, ~5,244 surfaces, from the same ICU the runtime uses for
+ * every other locale-sensitive operation. No download, no vendored corpus, no snapshot to drift.
+ *
+ * `is_primary = 0`: these are NAMES THE WORLD USES, not the country's canonical name. The display `name` stays whatever
+ * the gazetteer already had, so resolving `格鲁吉亚` answers with the Georgia country row rather than renaming it.
+ *
+ * Returns the row count so the caller can report it — a zero means ICU supplied nothing, which is a different fact from
+ * the pass not having run.
+ */
+export declare function stageCountryDisplayNames(ctx: {
+    attrs: Map<number, PlaceAttrs>;
+    iso2ByID: Map<number, string>;
+    countryPtID: number;
+    stageRow: StageRow;
+    tx: {
+        exec(sql: string): void;
+    };
+}): number;
+//# sourceMappingURL=country-display-names.d.ts.map

package/out/candidate/country-display-names.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"country-display-names.d.ts","sourceRoot":"","sources":["../../lib/candidate/country-display-names.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAIH,OAAO,KAAK,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,wBAAwB,CAAA;AAGlE;;;;;;;;;;;;;;;;;GAiBG;AACH,wBAAgB,wBAAwB,CAAC,GAAG,EAAE;IAC7C,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,UAAU,CAAC,CAAA;IAC9B,QAAQ,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC7B,WAAW,EAAE,MAAM,CAAA;IACnB,QAAQ,EAAE,QAAQ,CAAA;IAClB,EAAE,EAAE;QAAE,IAAI,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAAA;KAAE,CAAA;CAC/B,GAAG,MAAM,CA0CT"}

package/out/candidate/country-display-names.js

@@ -0,0 +1,59 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 1b of the candidate build — fold ICU's country display names onto the country rows.
+ */
+import { enumerateCountryDisplayNames } from "@mailwoman/codex/country";
+import { normalizeLocalityForKey } from "#street/normalize";
+/**
+ * Fold every country surface ICU knows onto that country's candidate row (#1678 thread 1).
+ *
+ * A bare `格鲁吉亚` (Georgia the country) resolved to NOTHING while `佐治亚州` (Georgia the US state) resolved correctly, and
+ * the model gave both the same wrong `locality` tag — so the tag was never the variable. Measured 2026-08-15: 140 of
+ * 237 country rows are synthetic and carry a canonical English name and nothing else; WOF holds no Chinese country
+ * names at all; and the GeoNames alias fold filters through a Latin-script regex, so neither existing source could ever
+ * supply them.
+ *
+ * `Intl.DisplayNames` already knows every one — ~280 regions, ~5,244 surfaces, from the same ICU the runtime uses for
+ * every other locale-sensitive operation. No download, no vendored corpus, no snapshot to drift.
+ *
+ * `is_primary = 0`: these are NAMES THE WORLD USES, not the country's canonical name. The display `name` stays whatever
+ * the gazetteer already had, so resolving `格鲁吉亚` answers with the Georgia country row rather than renaming it.
+ *
+ * Returns the row count so the caller can report it — a zero means ICU supplied nothing, which is a different fact from
+ * the pass not having run.
+ */
+export function stageCountryDisplayNames(ctx) {
+    // One country row per ISO2. Where a code has several (historic rows surviving the is_current filter), the most
+    // populous wins — the same tiebreak the ranking uses everywhere else.
+    const countryByISO2 = new Map();
+    for (const [sid, a] of ctx.attrs) {
+        if (a.ptid !== ctx.countryPtID)
+            continue;
+        const iso2 = ctx.iso2ByID.get(a.cid);
+        if (!iso2 || iso2 === "??")
+            continue;
+        const held = countryByISO2.get(iso2);
+        if (!held || a.pop > held.a.pop) {
+            countryByISO2.set(iso2, { sid, a });
+        }
+    }
+    let staged = 0;
+    ctx.tx.exec("BEGIN");
+    for (const { iso2, name } of enumerateCountryDisplayNames()) {
+        const target = countryByISO2.get(iso2);
+        if (!target)
+            continue;
+        const k = normalizeLocalityForKey(name);
+        // The country's own key is already staged as its primary; INSERT OR IGNORE at materialization dedupes the
+        // rest, so this only skips the obvious self-alias.
+        if (!k || k === target.a.pkey)
+            continue;
+        ctx.stageRow(k, target.a, target.sid, 0);
+        staged++;
+    }
+    ctx.tx.exec("COMMIT");
+    return staged;
+}
+//# sourceMappingURL=country-display-names.js.map

package/out/candidate/country-display-names.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"country-display-names.js","sourceRoot":"","sources":["../../lib/candidate/country-display-names.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,4BAA4B,EAAE,MAAM,0BAA0B,CAAA;AAGvE,OAAO,EAAE,uBAAuB,EAAE,MAAM,mBAAmB,CAAA;AAE3D;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,UAAU,wBAAwB,CAAC,GAMxC;IACA,+GAA+G;IAC/G,sEAAsE;IACtE,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0C,CAAA;IAEvE,KAAK,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QAClC,IAAI,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC,WAAW;YAAE,SAAQ;QAExC,MAAM,IAAI,GAAG,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;QAEpC,IAAI,CAAC,IAAI,IAAI,IAAI,KAAK,IAAI;YAAE,SAAQ;QAEpC,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAEpC,IAAI,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC;YACjC,aAAa,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAA;QACpC,CAAC;IACF,CAAC;IAED,IAAI,MAAM,GAAG,CAAC,CAAA;IAEd,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEpB,KAAK,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,4BAA4B,EAAE,EAAE,CAAC;QAC7D,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAEtC,IAAI,CAAC,MAAM;YAAE,SAAQ;QAErB,MAAM,CAAC,GAAG,uBAAuB,CAAC,IAAI,CAAC,CAAA;QAEvC,0GAA0G;QAC1G,mDAAmD;QACnD,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,CAAC,CAAC,IAAI;YAAE,SAAQ;QAEvC,GAAG,CAAC,QAAQ,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,EAAE,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;QAExC,MAAM,EAAE,CAAA;IACT,CAAC;IAED,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAErB,OAAO,MAAM,CAAA;AACd,CAAC"}

package/out/candidate/extract-fold.d.ts

@@ -0,0 +1,32 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 4 of the candidate build — fold a postcode or locality extract into the staging table.
+ */
+import { DatabaseClient } from "@mailwoman/sqlite/client";
+import type { CandidateDatabase } from "#candidate-schema";
+import type { StageRow } from "#candidate/place-attrs";
+/**
+ * Fold ONE extract (`spr` rows at `extractPlacetype` carrying real coordinates) in, then pass 4b: the alias names
+ * hanging off that same extract's `names` table.
+ *
+ * Self-contained by construction — it shares only the staging statement and the code dictionaries with the admin passes
+ * above it, and nothing downstream reads anything it produces except the two counters it returns.
+ */
+export declare function foldExtract(ctx: {
+    /**
+     * The staging connection. The extract itself is opened read-only here and closed before returning.
+     */
+    out: DatabaseClient<CandidateDatabase>;
+    extractPath: string;
+    extractPlacetype: "postalcode" | "locality";
+    ccID: (code: string | null) => number;
+    ptID: (pt: string | null) => number;
+    stageRow: StageRow;
+    progress: (phase: string, message: string) => void;
+}): {
+    primaries: number;
+    aliases: number;
+};
+//# sourceMappingURL=extract-fold.d.ts.map

package/out/candidate/extract-fold.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-fold.d.ts","sourceRoot":"","sources":["../../lib/candidate/extract-fold.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAA;AAGzD,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AAC1D,OAAO,KAAK,EAAc,QAAQ,EAAE,MAAM,wBAAwB,CAAA;AAIlE;;;;;;GAMG;AACH,wBAAgB,WAAW,CAAC,GAAG,EAAE;IAChC;;OAEG;IACH,GAAG,EAAE,cAAc,CAAC,iBAAiB,CAAC,CAAA;IACtC,WAAW,EAAE,MAAM,CAAA;IACnB,gBAAgB,EAAE,YAAY,GAAG,UAAU,CAAA;IAC3C,IAAI,EAAE,CAAC,IAAI,EAAE,MAAM,GAAG,IAAI,KAAK,MAAM,CAAA;IACrC,IAAI,EAAE,CAAC,EAAE,EAAE,MAAM,GAAG,IAAI,KAAK,MAAM,CAAA;IACnC,QAAQ,EAAE,QAAQ,CAAA;IAClB,QAAQ,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,KAAK,IAAI,CAAA;CAClD,GAAG;IAAE,SAAS,EAAE,MAAM,CAAC;IAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAwGzC"}

package/out/candidate/extract-fold.js

@@ -0,0 +1,104 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 4 of the candidate build — fold a postcode or locality extract into the staging table.
+ */
+import { DatabaseClient } from "@mailwoman/sqlite/client";
+import { tableExists } from "@mailwoman/sqlite/introspection";
+import { normalizeLocalityForKey } from "#street/normalize";
+/**
+ * Fold ONE extract (`spr` rows at `extractPlacetype` carrying real coordinates) in, then pass 4b: the alias names
+ * hanging off that same extract's `names` table.
+ *
+ * Self-contained by construction — it shares only the staging statement and the code dictionaries with the admin passes
+ * above it, and nothing downstream reads anything it produces except the two counters it returns.
+ */
+export function foldExtract(ctx) {
+    const { out, extractPath, extractPlacetype, ccID, ptID, stageRow, progress } = ctx;
+    progress(extractPlacetype === "postalcode" ? "postcodes" : "localities", `reading ${extractPath}`);
+    using pc = new DatabaseClient(extractPath, { readOnly: true });
+    const pcPtid = ptID(extractPlacetype);
+    // Per-extract, not the admin `attrs` map: pass 1 only ever sees the admin DB, so the alias pass
+    // below has nothing to join against unless this primary loop records what it staged.
+    const pcAttrs = new Map();
+    let primaries = 0;
+    let aliases = 0;
+    out.exec("BEGIN");
+    for (const r of pc
+        .prepare(`SELECT id, name, country, latitude, longitude,
+				min_latitude AS mnlat, min_longitude AS mnlon, max_latitude AS mxlat, max_longitude AS mxlon
+			 FROM spr WHERE placetype = ? AND latitude != 0 AND longitude != 0`)
+        .iterate(extractPlacetype)) {
+        const name = String(r.name ?? "");
+        const key = normalizeLocalityForKey(name);
+        if (!key)
+            continue;
+        const lat = r.latitude;
+        const lon = r.longitude;
+        // region_id 0 (a postcode is unique by name+country — no same-name disambiguation); neg_rank 0
+        // (no population). bbox = the postcode's own min/max (falls back to the centroid point).
+        const a = {
+            cid: ccID(r.country),
+            rid: 0,
+            ptid: pcPtid,
+            name,
+            lat,
+            lon,
+            mnLat: r.mnlat || lat,
+            mnLon: r.mnlon || lon,
+            mxLat: r.mxlat || lat,
+            mxLon: r.mxlon || lon,
+            pop: 0,
+            neg: 0,
+            pkey: key,
+            // A postcode has no toponym fame — nobody writes an encyclopedia article about SW1A 2AA — and
+            // the score source carries no `postalcode` rows to join against anyway. NULL is the truthful
+            // value: unmeasured, so the ranking key leaves postcode rows exactly where they were.
+            imp: null,
+        };
+        pcAttrs.set(Number(r.id), a);
+        stageRow(key, a, Number(r.id), 1);
+        primaries++;
+    }
+    out.exec("COMMIT");
+    // --- pass 4b: postcode ALIAS names (#1495) ---
+    //
+    // The delivery-city names GeoNames supplies for a ZIP ("Brooklyn" for 11201) are written into
+    // the extract's `names` table by `postcode/centroid-fills.ts`'s `geonamesNameFill`. Everything
+    // downstream of `names` picked them up EXCEPT this build: `fts.ts` unions `spr.name` with every
+    // `names` row into `place_search.alt_names`, so the FTS backend resolved "Brooklyn" → 11201
+    // while the candidate backend — whose every row IS an exact-tier row — had no key for it at
+    // all. Pass 2 does the equivalent fold for admin places, but reads the ADMIN `place_search`,
+    // and `attrs` holds admin ids only, so a postcode extract could never reach it.
+    //
+    // Same discipline as pass 2: `is_primary = 0` (so `rankByPrimaryPreference` treats it as an
+    // alias, not a canonical postcode name), the row stays denormalized onto the POSTCODE's own
+    // spr_id/coords/bbox, and the display `name` stays the postcode — resolving "brooklyn" answers
+    // with place 11201, it does not rename the place to its delivery city.
+    const hasNames = tableExists(pc, "names");
+    if (hasNames) {
+        out.exec("BEGIN");
+        for (const r of pc.prepare("SELECT id, name FROM names").iterate()) {
+            const a = pcAttrs.get(Number(r.id));
+            if (!a)
+                continue;
+            const k = normalizeLocalityForKey(String(r.name ?? ""));
+            // The postcode's own key is already staged as the primary; `INSERT OR IGNORE` at
+            // materialization dedupes repeats, so this only skips the obvious self-alias.
+            if (!k || k === a.pkey)
+                continue;
+            stageRow(k, a, Number(r.id), 0);
+            aliases++;
+        }
+        out.exec("COMMIT");
+    }
+    else {
+        // Never a silent zero: real extracts come from `createUnifiedSchema`, which always creates
+        // `names`. A extract without it has no alias surface to lose, but say so rather than reporting
+        // "0 aliases" from a table that was never read.
+        progress("postcode-aliases", `${extractPath} has no \`names\` table — no delivery-city aliases to fold`);
+    }
+    return { primaries, aliases };
+}
+//# sourceMappingURL=extract-fold.js.map

package/out/candidate/extract-fold.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-fold.js","sourceRoot":"","sources":["../../lib/candidate/extract-fold.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAA;AACzD,OAAO,EAAE,WAAW,EAAE,MAAM,iCAAiC,CAAA;AAK7D,OAAO,EAAE,uBAAuB,EAAE,MAAM,mBAAmB,CAAA;AAE3D;;;;;;GAMG;AACH,MAAM,UAAU,WAAW,CAAC,GAW3B;IACA,MAAM,EAAE,GAAG,EAAE,WAAW,EAAE,gBAAgB,EAAE,IAAI,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,GAAG,GAAG,CAAA;IAElF,QAAQ,CAAC,gBAAgB,KAAK,YAAY,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,YAAY,EAAE,WAAW,WAAW,EAAE,CAAC,CAAA;IAElG,MAAM,EAAE,GAAG,IAAI,cAAc,CAAc,WAAW,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAA;IAC3E,MAAM,MAAM,GAAG,IAAI,CAAC,gBAAgB,CAAC,CAAA;IACrC,gGAAgG;IAChG,qFAAqF;IACrF,MAAM,OAAO,GAAG,IAAI,GAAG,EAAsB,CAAA;IAC7C,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,OAAO,GAAG,CAAC,CAAA;IAEf,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,IAAI,EAAE;SAChB,OAAO,CACP;;sEAEmE,CACnE;SACA,OAAO,CAAC,gBAAgB,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,CAAA;QACjC,MAAM,GAAG,GAAG,uBAAuB,CAAC,IAAI,CAAC,CAAA;QAEzC,IAAI,CAAC,GAAG;YAAE,SAAQ;QAElB,MAAM,GAAG,GAAG,CAAC,CAAC,QAAkB,CAAA;QAChC,MAAM,GAAG,GAAG,CAAC,CAAC,SAAmB,CAAA;QAEjC,+FAA+F;QAC/F,yFAAyF;QACzF,MAAM,CAAC,GAAe;YACrB,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC,OAAwB,CAAC;YACrC,GAAG,EAAE,CAAC;YACN,IAAI,EAAE,MAAM;YACZ,IAAI;YACJ,GAAG;YACH,GAAG;YACH,KAAK,EAAG,CAAC,CAAC,KAAgB,IAAI,GAAG;YACjC,KAAK,EAAG,CAAC,CAAC,KAAgB,IAAI,GAAG;YACjC,KAAK,EAAG,CAAC,CAAC,KAAgB,IAAI,GAAG;YACjC,KAAK,EAAG,CAAC,CAAC,KAAgB,IAAI,GAAG;YACjC,GAAG,EAAE,CAAC;YACN,GAAG,EAAE,CAAC;YACN,IAAI,EAAE,GAAG;YACT,8FAA8F;YAC9F,6FAA6F;YAC7F,sFAAsF;YACtF,GAAG,EAAE,IAAI;SACT,CAAA;QAED,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;QAC5B,QAAQ,CAAC,GAAG,EAAE,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;QAEjC,SAAS,EAAE,CAAA;IACZ,CAAC;IAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAElB,gDAAgD;IAChD,EAAE;IACF,8FAA8F;IAC9F,+FAA+F;IAC/F,gGAAgG;IAChG,4FAA4F;IAC5F,4FAA4F;IAC5F,6FAA6F;IAC7F,gFAAgF;IAChF,EAAE;IACF,4FAA4F;IAC5F,4FAA4F;IAC5F,+FAA+F;IAC/F,uEAAuE;IACvE,MAAM,QAAQ,GAAG,WAAW,CAAC,EAAE,EAAE,OAAO,CAAC,CAAA;IAEzC,IAAI,QAAQ,EAAE,CAAC;QACd,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAEjB,KAAK,MAAM,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,4BAA4B,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC;YACpE,MAAM,CAAC,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;YAEnC,IAAI,CAAC,CAAC;gBAAE,SAAQ;YAEhB,MAAM,CAAC,GAAG,uBAAuB,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,CAAA;YAEvD,iFAAiF;YACjF,8EAA8E;YAC9E,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI;gBAAE,SAAQ;YAEhC,QAAQ,CAAC,CAAC,EAAE,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;YAE/B,OAAO,EAAE,CAAA;QACV,CAAC;QAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IACnB,CAAC;SAAM,CAAC;QACP,2FAA2F;QAC3F,+FAA+F;QAC/F,gDAAgD;QAChD,QAAQ,CAAC,kBAAkB,EAAE,GAAG,WAAW,4DAA4D,CAAC,CAAA;IACzG,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAA;AAC9B,CAAC"}

package/out/candidate/name-roles.d.ts

@@ -0,0 +1,57 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 3c of the candidate build — the `name_role` detectors and the thresholds they are judged by.
+ */
+import type { DatabaseClient } from "@mailwoman/sqlite/client";
+import type { CandidateDatabase } from "#candidate-schema";
+import type { PlaceAttrs } from "#candidate/place-attrs";
+import type { WOFDatabase } from "#schema";
+/**
+ * Key-count threshold for the gloss anomaly detector (#1730) — the sweep's own boundary: 4,000 places carried >= 50
+ * keys, and a legitimate famous place at that count (New York, 176 keys) is separated by the PROMINENCE check, never by
+ * this number alone.
+ */
+export declare const GLOSS_KEY_THRESHOLD = 50;
+/**
+ * Placetypes the gloss detector never flags. A country or region legitimately carries a name in every language — that
+ * is what an exonym set IS — so key volume discriminates nothing there. The detector's population is the non-admin
+ * tail, where a place named by a common noun ("Poisson", "Sunday") accumulating 200+ translations is a
+ * machine-translated gloss set, not fame.
+ */
+export declare const GLOSS_EXCLUDED_PLACETYPES: ReadonlySet<string>;
+/**
+ * Pass 3c — the #1730 name-role prototype: two independent detectors over the staged rows, WRITE-ONLY in this
+ * generation (no ranking consumer; the rank penalty is its own D-rule-conditional step with the `gloss_key` board as
+ * regression check). Both stamp `is_primary = 0` rows only — a place's canonical name and the `place_abbr` region
+ * abbreviations are never a gloss or a variant.
+ *
+ * - `gloss` is ANOMALY-based, and stamps only the certain core: key volume at/over the threshold + a non-admin placetype
+ *
+ *   - NO measured prominence (population absent AND importance unmeasured). Provenance cannot separate a gloss from an
+ *     exonym — WOF imported both as `x_preferred` — and prominence is what rescues New York/Paris.
+ * - `abbr` is PROVENANCE-based — the #936 signal: a WOF `variant` name in one of the country's official languages (or
+ *   English), measured there at a 13× key-collision rate. A source without a `names` table (fixture-scale admin DBs)
+ *   skips this detector loudly.
+ *
+ * Returns the stamp counts plus the census the prototype exists to report: how much of the ≥-threshold key tail carries
+ * any role.
+ */
+export declare function stampNameRoles(ctx: {
+    src: DatabaseClient<WOFDatabase>;
+    out: DatabaseClient<CandidateDatabase>;
+    attrs: Map<number, PlaceAttrs>;
+    keyCounts: Map<number, number>;
+    glossThreshold: number;
+    ptcodes: Map<string, number>;
+    ccodes: Map<string, number>;
+    progress: (phase: string, message: string) => void;
+}): {
+    roleGloss: number;
+    roleAbbr: number;
+    roleVariant: number;
+    keyTailPlaces: number;
+    keyTailWithRole: number;
+};
+//# sourceMappingURL=name-roles.d.ts.map

package/out/candidate/name-roles.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"name-roles.d.ts","sourceRoot":"","sources":["../../lib/candidate/name-roles.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAGH,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAA;AAG9D,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AAE1D,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,wBAAwB,CAAA;AACxD,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,SAAS,CAAA;AAG1C;;;;GAIG;AACH,eAAO,MAAM,mBAAmB,KAAK,CAAA;AAErC;;;;;GAKG;AACH,eAAO,MAAM,yBAAyB,EAAE,WAAW,CAAC,MAAM,CASxD,CAAA;AAEF;;;;;;;;;;;;;;;;GAgBG;AACH,wBAAgB,cAAc,CAAC,GAAG,EAAE;IACnC,GAAG,EAAE,cAAc,CAAC,WAAW,CAAC,CAAA;IAChC,GAAG,EAAE,cAAc,CAAC,iBAAiB,CAAC,CAAA;IACtC,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,UAAU,CAAC,CAAA;IAC9B,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC9B,cAAc,EAAE,MAAM,CAAA;IACtB,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC5B,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC3B,QAAQ,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,KAAK,IAAI,CAAA;CAClD,GAAG;IAAE,SAAS,EAAE,MAAM,CAAC;IAAC,QAAQ,EAAE,MAAM,CAAC;IAAC,WAAW,EAAE,MAAM,CAAC;IAAC,aAAa,EAAE,MAAM,CAAC;IAAC,eAAe,EAAE,MAAM,CAAA;CAAE,CA0K/G"}

package/out/candidate/name-roles.js

@@ -0,0 +1,166 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Pass 3c of the candidate build — the `name_role` detectors and the thresholds they are judged by.
+ */
+import { isOfficialLanguage } from "@mailwoman/codex/country";
+import { tableExists } from "@mailwoman/sqlite/introspection";
+import { isOwnNameVariant } from "#candidate/own-name";
+import { normalizeLocalityForKey } from "#street/normalize";
+/**
+ * Key-count threshold for the gloss anomaly detector (#1730) — the sweep's own boundary: 4,000 places carried >= 50
+ * keys, and a legitimate famous place at that count (New York, 176 keys) is separated by the PROMINENCE check, never by
+ * this number alone.
+ */
+export const GLOSS_KEY_THRESHOLD = 50;
+/**
+ * Placetypes the gloss detector never flags. A country or region legitimately carries a name in every language — that
+ * is what an exonym set IS — so key volume discriminates nothing there. The detector's population is the non-admin
+ * tail, where a place named by a common noun ("Poisson", "Sunday") accumulating 200+ translations is a
+ * machine-translated gloss set, not fame.
+ */
+export const GLOSS_EXCLUDED_PLACETYPES = new Set([
+    "country",
+    "dependency",
+    "disputed",
+    "empire",
+    "macroregion",
+    "region",
+    "macrocounty",
+    "county",
+]);
+/**
+ * Pass 3c — the #1730 name-role prototype: two independent detectors over the staged rows, WRITE-ONLY in this
+ * generation (no ranking consumer; the rank penalty is its own D-rule-conditional step with the `gloss_key` board as
+ * regression check). Both stamp `is_primary = 0` rows only — a place's canonical name and the `place_abbr` region
+ * abbreviations are never a gloss or a variant.
+ *
+ * - `gloss` is ANOMALY-based, and stamps only the certain core: key volume at/over the threshold + a non-admin placetype
+ *
+ *   - NO measured prominence (population absent AND importance unmeasured). Provenance cannot separate a gloss from an
+ *     exonym — WOF imported both as `x_preferred` — and prominence is what rescues New York/Paris.
+ * - `abbr` is PROVENANCE-based — the #936 signal: a WOF `variant` name in one of the country's official languages (or
+ *   English), measured there at a 13× key-collision rate. A source without a `names` table (fixture-scale admin DBs)
+ *   skips this detector loudly.
+ *
+ * Returns the stamp counts plus the census the prototype exists to report: how much of the ≥-threshold key tail carries
+ * any role.
+ */
+export function stampNameRoles(ctx) {
+    const { src, out, attrs, keyCounts, glossThreshold, progress } = ctx;
+    progress("roles", "stamping name roles (gloss anomaly + abbr provenance)");
+    const ptNameByID = new Map([...ctx.ptcodes].map(([placetype, id]) => [id, placetype]));
+    const iso2ByCID = new Map([...ctx.ccodes].map(([code, id]) => [id, code]));
+    let keyTailPlaces = 0;
+    const glossSids = [];
+    for (const [sid, count] of keyCounts) {
+        if (count < glossThreshold)
+            continue;
+        keyTailPlaces++;
+        const a = attrs.get(sid);
+        if (!a)
+            continue;
+        if (GLOSS_EXCLUDED_PLACETYPES.has(ptNameByID.get(a.ptid) ?? ""))
+            continue;
+        if (a.pop > 0 || a.imp != null)
+            continue;
+        glossSids.push(sid);
+    }
+    out.exec("CREATE TEMP TABLE role_key (spr_id INTEGER NOT NULL, name_key TEXT NOT NULL, PRIMARY KEY (spr_id, name_key)) WITHOUT ROWID");
+    const insRoleKey = out.prepare("INSERT OR IGNORE INTO role_key VALUES (?, ?)");
+    // Zero abbr stamps from a skipped detector is a different fact from zero variants found.
+    const hasSourceNames = tableExists(src, "names");
+    if (hasSourceNames) {
+        out.exec("BEGIN");
+        // Two provenance routes into the same stamp: WOF's abbreviation/short name KINDS arrive in the
+        // LANGUAGE column ('abbr'/'short' — 280 rows, measured 2026-08-18, Toledo's 'TO' among them) and
+        // qualify by kind alone; everything else qualifies as a variant in an official language.
+        for (const r of src
+            .prepare("SELECT id, name, language FROM names WHERE privateuse = 'variant' OR language IN ('abbr', 'short')")
+            .iterate()) {
+            const a = attrs.get(Number(r.id));
+            if (!a)
+                continue;
+            const language = String(r.language ?? "");
+            if (language !== "abbr" && language !== "short") {
+                const iso2 = iso2ByCID.get(a.cid) ?? "??";
+                if (language !== "eng" && !isOfficialLanguage(iso2, language))
+                    continue;
+            }
+            const k = normalizeLocalityForKey(String(r.name ?? ""));
+            if (!k || k === a.pkey)
+                continue;
+            insRoleKey.run(Number(r.id), k);
+        }
+        out.exec("COMMIT");
+    }
+    else {
+        progress("roles", "source carries no `names` table — abbr detector skipped (gloss still runs)");
+    }
+    // Stamp order is precedence: the provenance-based abbr first, then the own-name variant verdict,
+    // then gloss fills what neither claimed.
+    const roleAbbr = Number(out
+        .prepare(`UPDATE cand_stage SET name_role = 'abbr'
+				 WHERE is_primary = 0 AND EXISTS (
+					SELECT 1 FROM role_key rk WHERE rk.spr_id = cand_stage.spr_id AND rk.name_key = cand_stage.name_key)`)
+        .run().changes);
+    // --- variant detector (#1882): the alias surface is the holder's OWN primary name in another
+    // orthography — romanization, spacing/diacritic variant, or abbreviation expansion. The verdict
+    // is per (alias key, primary key) pair, so it runs in JS over the still-unstamped alias rows;
+    // an uncovered script answers no-verdict and stamps nothing (own-name.ts owns the predicate and
+    // its measured threshold). Runs BEFORE gloss on purpose: a surface that IS the place's own name
+    // is not a translation, whatever the key volume says.
+    out.exec("CREATE TEMP TABLE variant_key (spr_id INTEGER NOT NULL, name_key TEXT NOT NULL, PRIMARY KEY (spr_id, name_key)) WITHOUT ROWID");
+    const insVariantKey = out.prepare("INSERT OR IGNORE INTO variant_key VALUES (?, ?)");
+    let variantScanned = 0;
+    out.exec("BEGIN");
+    for (const r of out
+        .prepare("SELECT DISTINCT spr_id, name_key FROM cand_stage WHERE is_primary = 0 AND name_role IS NULL")
+        .iterate()) {
+        variantScanned++;
+        const a = attrs.get(Number(r.spr_id));
+        if (!a?.pkey)
+            continue;
+        if (isOwnNameVariant(String(a.pkey), String(r.name_key))) {
+            insVariantKey.run(Number(r.spr_id), String(r.name_key));
+        }
+    }
+    out.exec("COMMIT");
+    const roleVariant = Number(out
+        .prepare(`UPDATE cand_stage SET name_role = 'variant'
+				 WHERE is_primary = 0 AND name_role IS NULL AND EXISTS (
+					SELECT 1 FROM variant_key vk WHERE vk.spr_id = cand_stage.spr_id AND vk.name_key = cand_stage.name_key)`)
+        .run().changes);
+    out.exec("DROP TABLE variant_key");
+    progress("roles", `variant: ${roleVariant.toLocaleString()} of ${variantScanned.toLocaleString()} unstamped alias keys are own-name variants`);
+    out.exec("CREATE TEMP TABLE role_place (spr_id INTEGER PRIMARY KEY) WITHOUT ROWID");
+    out.exec("CREATE TEMP TABLE key_tail (spr_id INTEGER PRIMARY KEY) WITHOUT ROWID");
+    const insRolePlace = out.prepare("INSERT OR IGNORE INTO role_place VALUES (?)");
+    const insKeyTail = out.prepare("INSERT OR IGNORE INTO key_tail VALUES (?)");
+    out.exec("BEGIN");
+    for (const sid of glossSids) {
+        insRolePlace.run(sid);
+    }
+    for (const [sid, count] of keyCounts) {
+        if (count >= glossThreshold) {
+            insKeyTail.run(sid);
+        }
+    }
+    out.exec("COMMIT");
+    const roleGloss = Number(out
+        .prepare(`UPDATE cand_stage SET name_role = 'gloss'
+				 WHERE is_primary = 0 AND name_role IS NULL AND spr_id IN (SELECT spr_id FROM role_place)`)
+        .run().changes);
+    const keyTailWithRole = Number(out
+        .prepare(`SELECT count(DISTINCT spr_id) AS n FROM cand_stage
+				 WHERE name_role IS NOT NULL AND spr_id IN (SELECT spr_id FROM key_tail)`)
+        .get()["n"]);
+    out.exec("DROP TABLE role_key");
+    out.exec("DROP TABLE role_place");
+    out.exec("DROP TABLE key_tail");
+    progress("roles", `${roleAbbr.toLocaleString()} abbr + ${roleGloss.toLocaleString()} gloss rows stamped; ` +
+        `key tail (>= ${glossThreshold} keys): ${keyTailWithRole.toLocaleString()} of ${keyTailPlaces.toLocaleString()} places carry a role`);
+    return { roleGloss, roleAbbr, roleVariant, keyTailPlaces, keyTailWithRole };
+}
+//# sourceMappingURL=name-roles.js.map

package/out/candidate/name-roles.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"name-roles.js","sourceRoot":"","sources":["../../lib/candidate/name-roles.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAA;AAE7D,OAAO,EAAE,WAAW,EAAE,MAAM,iCAAiC,CAAA;AAG7D,OAAO,EAAE,gBAAgB,EAAE,MAAM,qBAAqB,CAAA;AAGtD,OAAO,EAAE,uBAAuB,EAAE,MAAM,mBAAmB,CAAA;AAE3D;;;;GAIG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAG,EAAE,CAAA;AAErC;;;;;GAKG;AACH,MAAM,CAAC,MAAM,yBAAyB,GAAwB,IAAI,GAAG,CAAC;IACrE,SAAS;IACT,YAAY;IACZ,UAAU;IACV,QAAQ;IACR,aAAa;IACb,QAAQ;IACR,aAAa;IACb,QAAQ;CACR,CAAC,CAAA;AAEF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,UAAU,cAAc,CAAC,GAS9B;IACA,MAAM,EAAE,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,cAAc,EAAE,QAAQ,EAAE,GAAG,GAAG,CAAA;IACpE,QAAQ,CAAC,OAAO,EAAE,uDAAuD,CAAC,CAAA;IAE1E,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC,CAAC,CAAA;IACtF,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,CAAC,CAAA;IAE1E,IAAI,aAAa,GAAG,CAAC,CAAA;IACrB,MAAM,SAAS,GAAa,EAAE,CAAA;IAE9B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,SAAS,EAAE,CAAC;QACtC,IAAI,KAAK,GAAG,cAAc;YAAE,SAAQ;QAEpC,aAAa,EAAE,CAAA;QACf,MAAM,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QAExB,IAAI,CAAC,CAAC;YAAE,SAAQ;QAEhB,IAAI,yBAAyB,CAAC,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YAAE,SAAQ;QAEzE,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,IAAI;YAAE,SAAQ;QACxC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACpB,CAAC;IAED,GAAG,CAAC,IAAI,CACP,4HAA4H,CAC5H,CAAA;IAED,MAAM,UAAU,GAAG,GAAG,CAAC,OAAO,CAAC,8CAA8C,CAAC,CAAA;IAE9E,yFAAyF;IACzF,MAAM,cAAc,GAAG,WAAW,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;IAEhD,IAAI,cAAc,EAAE,CAAC;QACpB,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAEjB,+FAA+F;QAC/F,iGAAiG;QACjG,yFAAyF;QACzF,KAAK,MAAM,CAAC,IAAI,GAAG;aACjB,OAAO,CAAC,oGAAoG,CAAC;aAC7G,OAAO,EAAE,EAAE,CAAC;YACb,MAAM,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;YAEjC,IAAI,CAAC,CAAC;gBAAE,SAAQ;YAChB,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAA;YAEzC,IAAI,QAAQ,KAAK,MAAM,IAAI,QAAQ,KAAK,OAAO,EAAE,CAAC;gBACjD,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,IAAI,CAAA;gBAEzC,IAAI,QAAQ,KAAK,KAAK,IAAI,CAAC,kBAAkB,CAAC,IAAI,EAAE,QAAQ,CAAC;oBAAE,SAAQ;YACxE,CAAC;YAED,MAAM,CAAC,GAAG,uBAAuB,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,CAAA;YAEvD,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI;gBAAE,SAAQ;YAChC,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;QAChC,CAAC;QAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IACnB,CAAC;SAAM,CAAC;QACP,QAAQ,CAAC,OAAO,EAAE,4EAA4E,CAAC,CAAA;IAChG,CAAC;IAED,iGAAiG;IACjG,yCAAyC;IACzC,MAAM,QAAQ,GAAG,MAAM,CACtB,GAAG;SACD,OAAO,CACP;;0GAEsG,CACtG;SACA,GAAG,EAAE,CAAC,OAAO,CACf,CAAA;IAED,8FAA8F;IAC9F,gGAAgG;IAChG,8FAA8F;IAC9F,gGAAgG;IAChG,gGAAgG;IAChG,sDAAsD;IACtD,GAAG,CAAC,IAAI,CACP,+HAA+H,CAC/H,CAAA;IAED,MAAM,aAAa,GAAG,GAAG,CAAC,OAAO,CAAC,iDAAiD,CAAC,CAAA;IACpF,IAAI,cAAc,GAAG,CAAC,CAAA;IAEtB,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,IAAI,GAAG;SACjB,OAAO,CAAC,6FAA6F,CAAC;SACtG,OAAO,EAAE,EAAE,CAAC;QACb,cAAc,EAAE,CAAA;QAChB,MAAM,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAA;QAErC,IAAI,CAAC,CAAC,EAAE,IAAI;YAAE,SAAQ;QAEtB,IAAI,gBAAgB,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC;YAC1D,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAA;QACxD,CAAC;IACF,CAAC;IAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAElB,MAAM,WAAW,GAAG,MAAM,CACzB,GAAG;SACD,OAAO,CACP;;6GAEyG,CACzG;SACA,GAAG,EAAE,CAAC,OAAO,CACf,CAAA;IAED,GAAG,CAAC,IAAI,CAAC,wBAAwB,CAAC,CAAA;IAElC,QAAQ,CACP,OAAO,EACP,YAAY,WAAW,CAAC,cAAc,EAAE,OAAO,cAAc,CAAC,cAAc,EAAE,6CAA6C,CAC3H,CAAA;IAED,GAAG,CAAC,IAAI,CAAC,yEAAyE,CAAC,CAAA;IACnF,GAAG,CAAC,IAAI,CAAC,uEAAuE,CAAC,CAAA;IACjF,MAAM,YAAY,GAAG,GAAG,CAAC,OAAO,CAAC,6CAA6C,CAAC,CAAA;IAC/E,MAAM,UAAU,GAAG,GAAG,CAAC,OAAO,CAAC,2CAA2C,CAAC,CAAA;IAC3E,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAEjB,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;QAC7B,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;IACtB,CAAC;IAED,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,SAAS,EAAE,CAAC;QACtC,IAAI,KAAK,IAAI,cAAc,EAAE,CAAC;YAC7B,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QACpB,CAAC;IACF,CAAC;IAED,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAElB,MAAM,SAAS,GAAG,MAAM,CACvB,GAAG;SACD,OAAO,CACP;8FAC0F,CAC1F;SACA,GAAG,EAAE,CAAC,OAAO,CACf,CAAA;IAED,MAAM,eAAe,GAAG,MAAM,CAC7B,GAAG;SACD,OAAO,CACP;6EACyE,CACzE;SACA,GAAG,EAAG,CAAC,GAAG,CAAC,CACb,CAAA;IAED,GAAG,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAA;IAC/B,GAAG,CAAC,IAAI,CAAC,uBAAuB,CAAC,CAAA;IACjC,GAAG,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAA;IAE/B,QAAQ,CACP,OAAO,EACP,GAAG,QAAQ,CAAC,cAAc,EAAE,WAAW,SAAS,CAAC,cAAc,EAAE,uBAAuB;QACvF,gBAAgB,cAAc,WAAW,eAAe,CAAC,cAAc,EAAE,OAAO,aAAa,CAAC,cAAc,EAAE,sBAAsB,CACrI,CAAA;IAED,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,WAAW,EAAE,aAAa,EAAE,eAAe,EAAE,CAAA;AAC5E,CAAC"}