@mailwoman/corpus 9.4.0 → 10.0.0

package/out/tools/golden/relabel-street/row.d.ts

@@ -0,0 +1,121 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Golden-set street-suffix relabel — v0.1.2 → v0.1.3.
+ *
+ *   ## Why this exists
+ *
+ *   The golden answer key and the training corpus disagreed about ONE thing, and the v9.0.0
+ *   promotion eval read the disagreement as a model regression (`us.street` 87.4 vs a floor of
+ *   87.8). The corpus SPLITS a US street into `street` + `street_suffix` — TIGER's adapter
+ *   decomposes at `corpus/src/adapters/tiger/street-decompose.ts`, the `street-affix` slice recipe
+ *   teaches it from USPS Pub-28, and `ComponentTag` carries `street_suffix` as a first-class tag.
+ *   The golden set FOLDED it: 2,216 US rows carry a `street`, and exactly 2 of them label a
+ *   `street_suffix`. Operator ruling, 2026-08-06: **the split is canonical**; the golden is the
+ *   stale side. This tool moves the answer key onto the corpus convention.
+ *
+ *   ## The instrument
+ *
+ *   `matchTrailingSuffix` from `@mailwoman/codex/us` — the USPS Pub-28 Appendix C table, which is
+ *   also what the corpus slice recipe splits on. The table is NOT re-implemented here, and the
+ *   libpostal dictionary TIGER reads is deliberately not used: measured on this golden set the two
+ *   disagree on 51 US rows, and the disagreements run in the codex table's favour (libpostal's
+ *   `directionals.txt` lists `center|c`, so TIGER reads the `C` of "C STREET" as a directional
+ *   prefix and then emits no suffix at all).
+ *
+ *   ## What it changes, and what it refuses to
+ *
+ *   Applied only to rows whose `country` is `US`. Three branches, mirroring the SHAPE of TIGER's
+ *   `decomposeStreet` on codex tables:
+ *
+ *   - **street type** — the last whitespace-separated word is a Pub-28 suffix, and something is left
+ *       over: "Main St" → `street: "Main"`, `street_suffix: "St"` (1,559 rows).
+ *   - **street type + post-directional** — the last word is a directional AND the one before it is a
+ *       Pub-28 suffix: "Pennsylvania Avenue NW" → `street: "Pennsylvania"`,
+ *       `street_suffix: "Avenue NW"` (347 rows). The post-directional joins the suffix rather than
+ *       becoming a tag of its own, because that is what the corpus adapter emits; there is no
+ *       `street_postfix` tag to move it to.
+ *   - **everything else is left folded** and reported. In particular a BARE post-directional tail
+ *       ("Seymour East", "BROADWAY N" — 16 rows) is NOT split: a directional is not a Pub-28 suffix,
+ *       and the observed rows in that class are unit-contaminated ("1ST AVE SW BOX E", where the
+ *       trailing "E" is a box letter).
+ *
+ *   FR rows are untouched, deliberately and permanently as far as this tool is concerned. French
+ *   street typology puts the type FIRST ("Rue de la Paix") and the golden labels only 7 of 665 FR
+ *   street rows with a `street_prefix`; whether FR should split at all is a different question with
+ *   a different table behind it, and nothing here should be read as having answered it.
+ *
+ *   ## Surface bytes
+ *
+ *   The split is a break at a whitespace run in the ORIGINAL string — no trimming, no case
+ *   normalization, no re-joining of tokens. `street + gap + street_suffix` reconstructs the input
+ *   byte-for-byte, so the whitespace between them belongs to neither span (the same shape the corpus
+ *   adapter's spans have). The tool asserts this per row and refuses to write a file if it ever
+ *   fails.
+ */
+/**
+ * A golden-set row, as stored one-per-line in `us.jsonl` / `fr.jsonl` / `adversarial.jsonl`. Only the fields this tool
+ * reads are modeled; every other key rides through untouched.
+ */
+export interface GoldenStreetRow {
+    raw: string;
+    components: Record<string, string>;
+    country?: string;
+    source?: string;
+    notes?: string;
+    [key: string]: unknown;
+}
+/**
+ * What the tool decided about one row. Every value except the two `split-*` classes means "left folded".
+ */
+export type GoldenRelabelClass = "split-suffix" | "split-suffix-postdirectional" | "split-prefix-only" | "already-split" | "single-token" | "suffix-only-street" | "postdirectional-tail-only" | "no-suffix-match" | "no-street" | "not-us" | "untrimmed-street";
+/**
+ * A review trigger on a row the tool DID change. A flag is never an adjudication — it marks the row for the operator's
+ * deck, and the split is applied either way.
+ */
+export interface GoldenRelabelFlag {
+    kind: "name-prone-suffix" | "venue-context" | "remainder-is-affix";
+    detail: string;
+}
+/**
+ * Row-level relabel outcome.
+ */
+export interface GoldenRelabelResult {
+    /**
+     * The row to write. Identical object reference when nothing changed.
+     */
+    row: GoldenStreetRow;
+    changed: boolean;
+    rowClass: GoldenRelabelClass;
+    flags: GoldenRelabelFlag[];
+    /**
+     * A leading directional was lifted into `street_prefix` on this row.
+     */
+    prefixSplit: boolean;
+    /**
+     * The street span as it stood in the parent version — recorded for the review deck.
+     */
+    beforeStreet?: string;
+}
+/**
+ * Options for {@linkcode relabelGoldenStreetRow}.
+ */
+export interface RelabelStreetRowOptions {
+    /**
+     * Also lift a folded LEADING directional out into `street_prefix`. Default true.
+     *
+     * On by default because the fold applies both ways and the answer key has to be corrected on both, or the correction
+     * is not a correction: 207 of the 1,682 split dev rows (12.3%) still opened with a directional after the suffix move
+     * — "N Desmet Avenue" would have graded `street: "N Desmet"` against a model that says `street_prefix: "N", street:
+     * "Desmet"`. Turn it OFF only to measure what the prefix fold alone costs.
+     */
+    splitPrefix?: boolean;
+}
+/**
+ * Decide, and apply, the US street-span split for ONE golden row. Pure: never mutates its argument, and returns the
+ * same object reference when the row is left alone.
+ */
+export declare function relabelGoldenStreetRow(row: GoldenStreetRow, options?: RelabelStreetRowOptions): GoldenRelabelResult;
+//# sourceMappingURL=row.d.ts.map

package/out/tools/golden/relabel-street/row.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"row.d.ts","sourceRoot":"","sources":["../../../../lib/tools/golden/relabel-street/row.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAwDG;AAYH;;;GAGG;AACH,MAAM,WAAW,eAAe;IAC/B,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAA;CACtB;AAED;;GAEG;AACH,MAAM,MAAM,kBAAkB,GAC3B,cAAc,GACd,8BAA8B,GAC9B,mBAAmB,GACnB,eAAe,GACf,cAAc,GACd,oBAAoB,GACpB,2BAA2B,GAC3B,iBAAiB,GACjB,WAAW,GACX,QAAQ,GACR,kBAAkB,CAAA;AAErB;;;GAGG;AACH,MAAM,WAAW,iBAAiB;IACjC,IAAI,EAAE,mBAAmB,GAAG,eAAe,GAAG,oBAAoB,CAAA;IAClE,MAAM,EAAE,MAAM,CAAA;CACd;AAED;;GAEG;AACH,MAAM,WAAW,mBAAmB;IACnC;;OAEG;IACH,GAAG,EAAE,eAAe,CAAA;IACpB,OAAO,EAAE,OAAO,CAAA;IAChB,QAAQ,EAAE,kBAAkB,CAAA;IAC5B,KAAK,EAAE,iBAAiB,EAAE,CAAA;IAC1B;;OAEG;IACH,WAAW,EAAE,OAAO,CAAA;IACpB;;OAEG;IACH,YAAY,CAAC,EAAE,MAAM,CAAA;CACrB;AAkFD;;GAEG;AACH,MAAM,WAAW,uBAAuB;IACvC;;;;;;;OAOG;IACH,WAAW,CAAC,EAAE,OAAO,CAAA;CACrB;AAED;;;GAGG;AACH,wBAAgB,sBAAsB,CACrC,GAAG,EAAE,eAAe,EACpB,OAAO,GAAE,uBAA4B,GACnC,mBAAmB,CAmIrB"}

package/out/tools/golden/relabel-street/row.js

@@ -0,0 +1,232 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Golden-set street-suffix relabel — v0.1.2 → v0.1.3.
+ *
+ *   ## Why this exists
+ *
+ *   The golden answer key and the training corpus disagreed about ONE thing, and the v9.0.0
+ *   promotion eval read the disagreement as a model regression (`us.street` 87.4 vs a floor of
+ *   87.8). The corpus SPLITS a US street into `street` + `street_suffix` — TIGER's adapter
+ *   decomposes at `corpus/src/adapters/tiger/street-decompose.ts`, the `street-affix` slice recipe
+ *   teaches it from USPS Pub-28, and `ComponentTag` carries `street_suffix` as a first-class tag.
+ *   The golden set FOLDED it: 2,216 US rows carry a `street`, and exactly 2 of them label a
+ *   `street_suffix`. Operator ruling, 2026-08-06: **the split is canonical**; the golden is the
+ *   stale side. This tool moves the answer key onto the corpus convention.
+ *
+ *   ## The instrument
+ *
+ *   `matchTrailingSuffix` from `@mailwoman/codex/us` — the USPS Pub-28 Appendix C table, which is
+ *   also what the corpus slice recipe splits on. The table is NOT re-implemented here, and the
+ *   libpostal dictionary TIGER reads is deliberately not used: measured on this golden set the two
+ *   disagree on 51 US rows, and the disagreements run in the codex table's favour (libpostal's
+ *   `directionals.txt` lists `center|c`, so TIGER reads the `C` of "C STREET" as a directional
+ *   prefix and then emits no suffix at all).
+ *
+ *   ## What it changes, and what it refuses to
+ *
+ *   Applied only to rows whose `country` is `US`. Three branches, mirroring the SHAPE of TIGER's
+ *   `decomposeStreet` on codex tables:
+ *
+ *   - **street type** — the last whitespace-separated word is a Pub-28 suffix, and something is left
+ *       over: "Main St" → `street: "Main"`, `street_suffix: "St"` (1,559 rows).
+ *   - **street type + post-directional** — the last word is a directional AND the one before it is a
+ *       Pub-28 suffix: "Pennsylvania Avenue NW" → `street: "Pennsylvania"`,
+ *       `street_suffix: "Avenue NW"` (347 rows). The post-directional joins the suffix rather than
+ *       becoming a tag of its own, because that is what the corpus adapter emits; there is no
+ *       `street_postfix` tag to move it to.
+ *   - **everything else is left folded** and reported. In particular a BARE post-directional tail
+ *       ("Seymour East", "BROADWAY N" — 16 rows) is NOT split: a directional is not a Pub-28 suffix,
+ *       and the observed rows in that class are unit-contaminated ("1ST AVE SW BOX E", where the
+ *       trailing "E" is a box letter).
+ *
+ *   FR rows are untouched, deliberately and permanently as far as this tool is concerned. French
+ *   street typology puts the type FIRST ("Rue de la Paix") and the golden labels only 7 of 665 FR
+ *   street rows with a `street_prefix`; whether FR should split at all is a different question with
+ *   a different table behind it, and nothing here should be read as having answered it.
+ *
+ *   ## Surface bytes
+ *
+ *   The split is a break at a whitespace run in the ORIGINAL string — no trimming, no case
+ *   normalization, no re-joining of tokens. `street + gap + street_suffix` reconstructs the input
+ *   byte-for-byte, so the whitespace between them belongs to neither span (the same shape the corpus
+ *   adapter's spans have). The tool asserts this per row and refuses to write a file if it ever
+ *   fails.
+ */
+import { isStreetDirectionalToken, matchTrailingSuffix, NAME_PRONE_US_SUFFIXES, } from "@mailwoman/codex/us";
+import { escapeRegExp } from "@mailwoman/core/strings/regexp";
+/**
+ * Split `s` at its LAST whitespace run, returning the three pieces verbatim. Null when there is no interior whitespace,
+ * when the head would be empty, or when `s` carries leading/trailing whitespace (a golden row is stored trimmed; an
+ * untrimmed one is reported rather than silently normalized).
+ */
+function splitLastWord(s) {
+    if (s !== s.trim() || !s)
+        return null;
+    const match = /^(.*\S)(\s+)(\S+)$/.exec(s);
+    if (!match)
+        return null;
+    return { head: match[1], gap: match[2], tail: match[3] };
+}
+/**
+ * Split `s` at its FIRST whitespace run — the leading-directional counterpart of {@link splitLastWord}. `head` is the
+ * first word, `tail` the rest, both verbatim.
+ */
+function splitFirstWord(s) {
+    if (s !== s.trim() || !s)
+        return null;
+    const match = /^(\S+)(\s+)(\S.*)$/.exec(s);
+    if (!match)
+        return null;
+    return { head: match[1], gap: match[2], tail: match[3] };
+}
+// ── Row-level relabel ──────────────────────────────────────────────────────
+/**
+ * Rebuild `components` with `street_suffix` inserted immediately after `street`, so the written row reads in address
+ * order rather than with the new tag appended at the end.
+ */
+function withStreetSpans(components, spans) {
+    const out = {};
+    for (const [key, value] of Object.entries(components)) {
+        if (key === "street") {
+            if (spans.prefix) {
+                out.street_prefix = spans.prefix;
+            }
+            out.street = spans.street;
+            if (spans.suffix) {
+                out.street_suffix = spans.suffix;
+            }
+            continue;
+        }
+        if (key === "street_prefix" || key === "street_suffix")
+            continue;
+        out[key] = value;
+    }
+    return out;
+}
+/**
+ * Decide, and apply, the US street-span split for ONE golden row. Pure: never mutates its argument, and returns the
+ * same object reference when the row is left alone.
+ */
+export function relabelGoldenStreetRow(row, options = {}) {
+    const splitPrefix = options.splitPrefix ?? true;
+    const keep = (rowClass) => ({
+        row,
+        changed: false,
+        rowClass,
+        flags: [],
+        prefixSplit: false,
+    });
+    if ((row.country ?? "").toUpperCase() !== "US")
+        return keep("not-us");
+    const street = row.components.street;
+    if (!street)
+        return keep("no-street");
+    if (row.components.street_suffix)
+        return keep("already-split");
+    if (street !== street.trim())
+        return keep("untrimmed-street");
+    const flags = [];
+    let rowClass;
+    let name = street;
+    let suffix;
+    let suffixGap = "";
+    let canonical;
+    const split = splitLastWord(street);
+    if (!split) {
+        rowClass = matchTrailingSuffix(street) ? "suffix-only-street" : "single-token";
+    }
+    else if (isStreetDirectionalToken(split.tail)) {
+        // Street type + post-directional ("Pennsylvania Avenue NW"). The corpus adapter emits the pair as
+        // ONE suffix span, and there is no post-directional tag to move it to.
+        const inner = splitLastWord(split.head);
+        const typeMatch = inner ? matchTrailingSuffix(inner.tail) : null;
+        if (!inner || !typeMatch) {
+            rowClass = "postdirectional-tail-only";
+        }
+        else {
+            rowClass = "split-suffix-postdirectional";
+            name = inner.head;
+            suffix = `${inner.tail}${split.gap}${split.tail}`;
+            suffixGap = inner.gap;
+            canonical = typeMatch.canonical;
+        }
+    }
+    else {
+        const typeMatch = matchTrailingSuffix(street);
+        if (!typeMatch) {
+            rowClass = "no-suffix-match";
+        }
+        else {
+            rowClass = "split-suffix";
+            name = split.head;
+            suffix = split.tail;
+            suffixGap = split.gap;
+            canonical = typeMatch.canonical;
+        }
+    }
+    // Leading directional → street_prefix, on whatever name survived the suffix move. Independent of the
+    // suffix branch, because "N Main" is as folded as "N Main St" is.
+    let prefix;
+    let prefixGap = "";
+    if (splitPrefix && !row.components.street_prefix) {
+        const lead = splitFirstWord(name);
+        if (lead && isStreetDirectionalToken(lead.head)) {
+            prefix = lead.head;
+            prefixGap = lead.gap;
+            name = lead.tail;
+        }
+    }
+    if (!suffix && !prefix)
+        return keep(rowClass);
+    if (!suffix) {
+        rowClass = "split-prefix-only";
+    }
+    // The invariant this tool exists to keep: the spans plus the whitespace between them ARE the original
+    // span, byte for byte. Anything else means a token was rewritten.
+    const rebuilt = `${prefix ? prefix + prefixGap : ""}${name}${suffix ? suffixGap + suffix : ""}`;
+    if (rebuilt !== street) {
+        throw new Error(`golden-relabel: span reconstruction failed for ${JSON.stringify(street)}`);
+    }
+    if (canonical && NAME_PRONE_US_SUFFIXES.has(canonical)) {
+        flags.push({
+            kind: "name-prone-suffix",
+            detail: `${canonical} heads proper names as often as street types`,
+        });
+    }
+    const venue = row.components.venue;
+    if (suffix && venue && new RegExp(`(^|\\W)${escapeRegExp(suffix)}(\\W|$)`, "i").test(venue)) {
+        flags.push({
+            kind: "venue-context",
+            detail: `venue ${JSON.stringify(venue)} also carries ${JSON.stringify(suffix)}`,
+        });
+    }
+    // Narrow on purpose: a name that happens to be a Pub-28 canonical is NOT interesting ("Mountain Rd",
+    // "Valley Dr", "Mills Ln" are ordinary streets, and flagging them buried the deck — 108 rows of noise
+    // on the first run). A name that is a bare DIRECTIONAL is: "East Rd" leaves `street: "East"`, which is
+    // a direction, not a name.
+    if (isStreetDirectionalToken(name)) {
+        flags.push({
+            kind: "remainder-is-affix",
+            detail: `street would become the bare directional ${JSON.stringify(name)}`,
+        });
+    }
+    return {
+        row: {
+            ...row,
+            components: withStreetSpans(row.components, {
+                ...(prefix ? { prefix } : row.components.street_prefix ? { prefix: row.components.street_prefix } : {}),
+                street: name,
+                ...(suffix ? { suffix } : {}),
+            }),
+        },
+        changed: true,
+        rowClass,
+        flags,
+        prefixSplit: Boolean(prefix),
+        beforeStreet: street,
+    };
+}
+//# sourceMappingURL=row.js.map

package/out/tools/golden/relabel-street/row.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"row.js","sourceRoot":"","sources":["../../../../lib/tools/golden/relabel-street/row.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAwDG;AAEH,OAAO,EACN,wBAAwB,EACxB,mBAAmB,EAEnB,sBAAsB,GACtB,MAAM,qBAAqB,CAAA;AAC5B,OAAO,EAAE,YAAY,EAAE,MAAM,gCAAgC,CAAA;AAkF7D;;;;GAIG;AACH,SAAS,aAAa,CAAC,CAAS;IAC/B,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAAE,OAAO,IAAI,CAAA;IACrC,MAAM,KAAK,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IAE1C,IAAI,CAAC,KAAK;QAAE,OAAO,IAAI,CAAA;IAEvB,OAAO,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,CAAA;AAC5D,CAAC;AAED;;;GAGG;AACH,SAAS,cAAc,CAAC,CAAS;IAChC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAAE,OAAO,IAAI,CAAA;IACrC,MAAM,KAAK,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IAE1C,IAAI,CAAC,KAAK;QAAE,OAAO,IAAI,CAAA;IAEvB,OAAO,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,CAAA;AAC5D,CAAC;AAED,8EAA8E;AAE9E;;;GAGG;AACH,SAAS,eAAe,CACvB,UAAkC,EAClC,KAA2D;IAE3D,MAAM,GAAG,GAA2B,EAAE,CAAA;IAEtC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC;QACvD,IAAI,GAAG,KAAK,QAAQ,EAAE,CAAC;YACtB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;gBAClB,GAAG,CAAC,aAAa,GAAG,KAAK,CAAC,MAAM,CAAA;YACjC,CAAC;YAED,GAAG,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAA;YAEzB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;gBAClB,GAAG,CAAC,aAAa,GAAG,KAAK,CAAC,MAAM,CAAA;YACjC,CAAC;YAED,SAAQ;QACT,CAAC;QAED,IAAI,GAAG,KAAK,eAAe,IAAI,GAAG,KAAK,eAAe;YAAE,SAAQ;QAChE,GAAG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAA;IACjB,CAAC;IAED,OAAO,GAAG,CAAA;AACX,CAAC;AAiBD;;;GAGG;AACH,MAAM,UAAU,sBAAsB,CACrC,GAAoB,EACpB,OAAO,GAA4B,EAAE;IAErC,MAAM,WAAW,GAAG,OAAO,CAAC,WAAW,IAAI,IAAI,CAAA;IAE/C,MAAM,IAAI,GAAG,CAAC,QAA4B,EAAuB,EAAE,CAAC,CAAC;QACpE,GAAG;QACH,OAAO,EAAE,KAAK;QACd,QAAQ;QACR,KAAK,EAAE,EAAE;QACT,WAAW,EAAE,KAAK;KAClB,CAAC,CAAA;IAEF,IAAI,CAAC,GAAG,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC,QAAQ,CAAC,CAAA;IACrE,MAAM,MAAM,GAAG,GAAG,CAAC,UAAU,CAAC,MAAM,CAAA;IAEpC,IAAI,CAAC,MAAM;QAAE,OAAO,IAAI,CAAC,WAAW,CAAC,CAAA;IAErC,IAAI,GAAG,CAAC,UAAU,CAAC,aAAa;QAAE,OAAO,IAAI,CAAC,eAAe,CAAC,CAAA;IAE9D,IAAI,MAAM,KAAK,MAAM,CAAC,IAAI,EAAE;QAAE,OAAO,IAAI,CAAC,kBAAkB,CAAC,CAAA;IAE7D,MAAM,KAAK,GAAwB,EAAE,CAAA;IACrC,IAAI,QAA4B,CAAA;IAChC,IAAI,IAAI,GAAG,MAAM,CAAA;IACjB,IAAI,MAA0B,CAAA;IAC9B,IAAI,SAAS,GAAG,EAAE,CAAA;IAClB,IAAI,SAAqC,CAAA;IAEzC,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,CAAC,CAAA;IAEnC,IAAI,CAAC,KAAK,EAAE,CAAC;QACZ,QAAQ,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAC,cAAc,CAAA;IAC/E,CAAC;SAAM,IAAI,wBAAwB,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QACjD,kGAAkG;QAClG,uEAAuE;QACvE,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QACvC,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;QAEhE,IAAI,CAAC,KAAK,IAAI,CAAC,SAAS,EAAE,CAAC;YAC1B,QAAQ,GAAG,2BAA2B,CAAA;QACvC,CAAC;aAAM,CAAC;YACP,QAAQ,GAAG,8BAA8B,CAAA;YACzC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;YACjB,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,GAAG,KAAK,CAAC,GAAG,GAAG,KAAK,CAAC,IAAI,EAAE,CAAA;YACjD,SAAS,GAAG,KAAK,CAAC,GAAG,CAAA;YACrB,SAAS,GAAG,SAAS,CAAC,SAAS,CAAA;QAChC,CAAC;IACF,CAAC;SAAM,CAAC;QACP,MAAM,SAAS,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAA;QAE7C,IAAI,CAAC,SAAS,EAAE,CAAC;YAChB,QAAQ,GAAG,iBAAiB,CAAA;QAC7B,CAAC;aAAM,CAAC;YACP,QAAQ,GAAG,cAAc,CAAA;YACzB,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;YACjB,MAAM,GAAG,KAAK,CAAC,IAAI,CAAA;YACnB,SAAS,GAAG,KAAK,CAAC,GAAG,CAAA;YACrB,SAAS,GAAG,SAAS,CAAC,SAAS,CAAA;QAChC,CAAC;IACF,CAAC;IAED,qGAAqG;IACrG,kEAAkE;IAClE,IAAI,MAA0B,CAAA;IAC9B,IAAI,SAAS,GAAG,EAAE,CAAA;IAElB,IAAI,WAAW,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,aAAa,EAAE,CAAC;QAClD,MAAM,IAAI,GAAG,cAAc,CAAC,IAAI,CAAC,CAAA;QAEjC,IAAI,IAAI,IAAI,wBAAwB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACjD,MAAM,GAAG,IAAI,CAAC,IAAI,CAAA;YAClB,SAAS,GAAG,IAAI,CAAC,GAAG,CAAA;YACpB,IAAI,GAAG,IAAI,CAAC,IAAI,CAAA;QACjB,CAAC;IACF,CAAC;IAED,IAAI,CAAC,MAAM,IAAI,CAAC,MAAM;QAAE,OAAO,IAAI,CAAC,QAAQ,CAAC,CAAA;IAE7C,IAAI,CAAC,MAAM,EAAE,CAAC;QACb,QAAQ,GAAG,mBAAmB,CAAA;IAC/B,CAAC;IAED,sGAAsG;IACtG,kEAAkE;IAClE,MAAM,OAAO,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,SAAS,GAAG,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAA;IAE/F,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;QACxB,MAAM,IAAI,KAAK,CAAC,kDAAkD,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,EAAE,CAAC,CAAA;IAC5F,CAAC;IAED,IAAI,SAAS,IAAI,sBAAsB,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;QACxD,KAAK,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,mBAAmB;YACzB,MAAM,EAAE,GAAG,SAAS,8CAA8C;SAClE,CAAC,CAAA;IACH,CAAC;IAED,MAAM,KAAK,GAAG,GAAG,CAAC,UAAU,CAAC,KAAK,CAAA;IAElC,IAAI,MAAM,IAAI,KAAK,IAAI,IAAI,MAAM,CAAC,UAAU,YAAY,CAAC,MAAM,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QAC7F,KAAK,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,eAAe;YACrB,MAAM,EAAE,SAAS,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,iBAAiB,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,EAAE;SAC/E,CAAC,CAAA;IACH,CAAC;IAED,qGAAqG;IACrG,sGAAsG;IACtG,uGAAuG;IACvG,2BAA2B;IAC3B,IAAI,wBAAwB,CAAC,IAAI,CAAC,EAAE,CAAC;QACpC,KAAK,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,oBAAoB;YAC1B,MAAM,EAAE,4CAA4C,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE;SAC1E,CAAC,CAAA;IACH,CAAC;IAED,OAAO;QACN,GAAG,EAAE;YACJ,GAAG,GAAG;YACN,UAAU,EAAE,eAAe,CAAC,GAAG,CAAC,UAAU,EAAE;gBAC3C,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,GAAG,CAAC,UAAU,CAAC,aAAa,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACvG,MAAM,EAAE,IAAI;gBACZ,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aAC7B,CAAC;SACF;QACD,OAAO,EAAE,IAAI;QACb,QAAQ;QACR,KAAK;QACL,WAAW,EAAE,OAAO,CAAC,MAAM,CAAC;QAC5B,YAAY,EAAE,MAAM;KACpB,CAAA;AACF,CAAC"}

package/out/tools/index.d.ts

@@ -11,18 +11,18 @@ export * from "#tools/audit";
 export * from "#tools/corpus-stats";
 export * from "#tools/ingest-csv";
 export * from "#tools/overlay-manifest";
-export * from "#tools/fetch/download";
+export * from "#tools/fetch/download/index";
 export * from "#tools/fetch/index";
-export * from "#tools/golden-expand";
-export * from "#tools/golden-promote";
-export * from "#tools/golden-relabel-street";
+export * from "#tools/golden/expand";
+export * from "#tools/golden/promote";
+export * from "#tools/golden/relabel-street/index";
 export * from "#tools/jsonl-to-parquet";
-export * from "#tools/lint-slice";
-export * from "#tools/lint-slice-vocab";
-export * from "#tools/slice-kryptonite";
+export * from "#tools/lint/slice/index";
+export * from "#tools/lint/slice/vocab";
+export * from "#tools/slice/kryptonite";
 export * from "#tools/overture-subvenue";
 export * from "#tools/postcode-triples";
-export * from "#tools/sub-venue-lexicon";
-export * from "#tools/sub-venue-promotions";
-export * from "#tools/slice-translit";
+export * from "#tools/sub/venue/lexicon";
+export * from "#tools/sub/venue/promotions";
+export * from "#tools/slice/translit";
 //# sourceMappingURL=index.d.ts.map

package/out/tools/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../lib/tools/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,oBAAoB,CAAA;AAClC,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,uBAAuB,CAAA;AACrC,cAAc,oBAAoB,CAAA;AAClC,cAAc,sBAAsB,CAAA;AACpC,cAAc,uBAAuB,CAAA;AACrC,cAAc,8BAA8B,CAAA;AAC5C,cAAc,yBAAyB,CAAA;AACvC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,uBAAuB,CAAA"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../lib/tools/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,oBAAoB,CAAA;AAClC,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,oBAAoB,CAAA;AAClC,cAAc,sBAAsB,CAAA;AACpC,cAAc,uBAAuB,CAAA;AACrC,cAAc,oCAAoC,CAAA;AAClD,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,uBAAuB,CAAA"}

package/out/tools/index.js

@@ -11,18 +11,18 @@ export * from "#tools/audit";
 export * from "#tools/corpus-stats";
 export * from "#tools/ingest-csv";
 export * from "#tools/overlay-manifest";
-export * from "#tools/fetch/download";
+export * from "#tools/fetch/download/index";
 export * from "#tools/fetch/index";
-export * from "#tools/golden-expand";
-export * from "#tools/golden-promote";
-export * from "#tools/golden-relabel-street";
+export * from "#tools/golden/expand";
+export * from "#tools/golden/promote";
+export * from "#tools/golden/relabel-street/index";
 export * from "#tools/jsonl-to-parquet";
-export * from "#tools/lint-slice";
-export * from "#tools/lint-slice-vocab";
-export * from "#tools/slice-kryptonite";
+export * from "#tools/lint/slice/index";
+export * from "#tools/lint/slice/vocab";
+export * from "#tools/slice/kryptonite";
 export * from "#tools/overture-subvenue";
 export * from "#tools/postcode-triples";
-export * from "#tools/sub-venue-lexicon";
-export * from "#tools/sub-venue-promotions";
-export * from "#tools/slice-translit";
+export * from "#tools/sub/venue/lexicon";
+export * from "#tools/sub/venue/promotions";
+export * from "#tools/slice/translit";
 //# sourceMappingURL=index.js.map

package/out/tools/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../../lib/tools/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,oBAAoB,CAAA;AAClC,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,uBAAuB,CAAA;AACrC,cAAc,oBAAoB,CAAA;AAClC,cAAc,sBAAsB,CAAA;AACpC,cAAc,uBAAuB,CAAA;AACrC,cAAc,8BAA8B,CAAA;AAC5C,cAAc,yBAAyB,CAAA;AACvC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,uBAAuB,CAAA"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../lib/tools/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,oBAAoB,CAAA;AAClC,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,mBAAmB,CAAA;AACjC,cAAc,yBAAyB,CAAA;AACvC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,oBAAoB,CAAA;AAClC,cAAc,sBAAsB,CAAA;AACpC,cAAc,uBAAuB,CAAA;AACrC,cAAc,oCAAoC,CAAA;AAClD,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,yBAAyB,CAAA;AACvC,cAAc,0BAA0B,CAAA;AACxC,cAAc,6BAA6B,CAAA;AAC3C,cAAc,uBAAuB,CAAA"}

package/out/tools/ingest-csv.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"ingest-csv.d.ts","sourceRoot":"","sources":["../../lib/tools/ingest-csv.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAyTH;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC,KAAK,EAAE,MAAM,CAAA;IACb,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,QAAQ,CAAC,EAAE,OAAO,CAAA;IAClB,MAAM,CAAC,EAAE,OAAO,CAAA;CAChB;AAED;;;;;GAKG;AACH,wBAAsB,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,IAAI,CAAC,CAiBxE"}
+{"version":3,"file":"ingest-csv.d.ts","sourceRoot":"","sources":["../../lib/tools/ingest-csv.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAuTH;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC,KAAK,EAAE,MAAM,CAAA;IACb,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,QAAQ,CAAC,EAAE,OAAO,CAAA;IAClB,MAAM,CAAC,EAAE,OAAO,CAAA;CAChB;AAED;;;;;GAKG;AACH,wBAAsB,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,IAAI,CAAC,CAiBxE"}

package/out/tools/ingest-csv.js

@@ -91,13 +91,11 @@ async function runIngest(opts) {
     // `skipEmpty` drops blank lines that readline would have turned into all-null rows. The early `break` closes the
     // file descriptor.
     const rows = () => CSVSpliterator.fromAsync(opts.inputPath, {
-        mode: "array",
         header: false,
         columnDelimiter: opts.separator,
         drop: opts.skipLines,
-        enableQuoteHandling: true,
     });
-    // --- Pass 1: read header + sample rows for type inference ---
+    // Read the header and sample rows before inferring the SQLite schema.
     process.stderr.write(`Reading ${opts.inputPath} for schema inference...\n`);
     let headerRow = null;
     const sampleRows = [];
@@ -116,7 +114,7 @@ async function runIngest(opts) {
     if (!headerRow && opts.hasHeader) {
         throw new Error("No header line found in CSV");
     }
-    // --- Determine column names ---
+    // Normalize and deduplicate the input column names.
     let rawHeaders;
     if (opts.hasHeader && headerRow) {
         rawHeaders = headerRow;
@@ -127,7 +125,7 @@ async function runIngest(opts) {
         rawHeaders = Array.from({ length: numCols }, (_, i) => `col_${i}`);
     }
     const colNames = dedupColumns(rawHeaders.map(normalizeColumnName));
-    // --- Infer column types ---
+    // Infer each column's SQLite type from its sample values.
     const columns = colNames.map((name, i) => {
         const samples = sampleRows.map((row) => {
             const raw = row[i] ?? "";
@@ -137,7 +135,7 @@ async function runIngest(opts) {
         info.name = name;
         return info;
     });
-    // --- Generate SQL ---
+    // Generate the table definition and insert statement from the inferred schema.
     const colDefs = columns.map((c) => `"${c.name}" ${c.type}`).join(",\n  ");
     // Raw DDL by design: the column set + types are INFERRED from the CSV at runtime (colDefs above),
     // so a Kysely builder loop would just wrap the same dynamic strings with ceremony and no type safety.
@@ -149,7 +147,7 @@ async function runIngest(opts) {
         process.stderr.write("--dry-run: stopping before import\n");
         return;
     }
-    // --- Create database + import ---
+    // Create the database, then import every input row.
     const { DatabaseClient } = await import("@mailwoman/sqlite/client");
     await makeDirectories(dirname(opts.outputPath));
     // `Database` — the EMPTY schema — deliberately, not by default: `createTableSQL` is built from the columns and

package/out/tools/ingest-csv.js.map

@@ -1 +1 @@
-{"version":3,"file":"ingest-csv.js","sourceRoot":"","sources":["../../lib/tools/ingest-csv.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAEH,OAAO,EAAE,aAAa,EAAE,MAAM,+BAA+B,CAAA;AAC7D,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM,4BAA4B,CAAA;AAChE,OAAO,EAAE,eAAe,EAAE,kBAAkB,EAAE,MAAM,4BAA4B,CAAA;AAGhF,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,SAAS,CAAA;AAC1D,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAE5C,mCAAmC;AAEnC;;GAEG;AACH,SAAS,mBAAmB,CAAC,GAAW;IACvC,OAAO,CACN,GAAG;SACD,IAAI,EAAE;SACN,UAAU,CAAC,UAAU,EAAE,EAAE,CAAC;SAC1B,WAAW,EAAE;SACb,UAAU,CAAC,aAAa,EAAE,GAAG,CAAC;SAC9B,UAAU,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,SAAS,CACvC,CAAA;AACF,CAAC;AAED,SAAS,YAAY,CAAC,KAAe;IACpC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAA;IAEtC,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,GAAG,CAAC,CAAC,CAAA;QAEzB,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,EAAE,CAAA;IACnD,CAAC,CAAC,CAAA;AACH,CAAC;AAcD,SAAS,cAAc,CAAC,GAAW;IAClC,IAAI,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IAElB,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC3D,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;IACzC,CAAC;IAED,oCAAoC;IACpC,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,WAAW,EAAE,CAAC;QACxG,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,OAAO,CAAC,CAAA;AACT,CAAC;AAED,SAAS,eAAe,CAAC,OAA0B;IAClD,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,SAAS,GAAG,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;QACzB,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;YAChB,SAAS,EAAE,CAAA;YAEX,SAAQ;QACT,CAAC;QAED,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACvB,QAAQ,EAAE,CAAA;QACX,CAAC;aAAM,IAAI,eAAe,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACpC,SAAS,EAAE,CAAA;QACZ,CAAC;aAAM,CAAC;YACP,SAAS,EAAE,CAAA;QACZ,CAAC;IACF,CAAC;IAED,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;IAC5B,MAAM,IAAI,GAAkB,SAAS,GAAG,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,GAAG,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAA;IAE5G,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,SAAS,GAAG,KAAK,IAAI,GAAG,EAAE,CAAA;AAC9D,CAAC;AAiBD,KAAK,UAAU,SAAS,CAAC,IAAmB;IAC3C,+GAA+G;IAC/G,8GAA8G;IAC9G,mHAAmH;IACnH,iHAAiH;IACjH,mBAAmB;IACnB,MAAM,IAAI,GAAG,GAA4B,EAAE,CAC1C,cAAc,CAAC,SAAS,CAAW,IAAI,CAAC,SAAS,EAAE;QAClD,IAAI,EAAE,OAAO;QACb,MAAM,EAAE,KAAK;QACb,eAAe,EAAE,IAAI,CAAC,SAAS;QAC/B,IAAI,EAAE,IAAI,CAAC,SAAS;QACpB,mBAAmB,EAAE,IAAI;KACzB,CAAC,CAAA;IAEH,+DAA+D;IAC/D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,WAAW,IAAI,CAAC,SAAS,4BAA4B,CAAC,CAAA;IAE3E,IAAI,SAAS,GAAoB,IAAI,CAAA;IACrC,MAAM,UAAU,GAAe,EAAE,CAAA;IAEjC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,IAAI,EAAE,EAAE,CAAC;QAChC,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;YAClC,SAAS,GAAG,GAAG,CAAA;YAEf,SAAQ;QACT,CAAC;QAED,IAAI,UAAU,CAAC,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;YACzC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACrB,CAAC;aAAM,CAAC;YACP,MAAK;QACN,CAAC;IACF,CAAC;IAED,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;QAClC,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAA;IAC/C,CAAC;IAED,iCAAiC;IACjC,IAAI,UAAoB,CAAA;IAExB,IAAI,IAAI,CAAC,SAAS,IAAI,SAAS,EAAE,CAAC;QACjC,UAAU,GAAG,SAAS,CAAA;IACvB,CAAC;SAAM,CAAC;QACP,gDAAgD;QAChD,MAAM,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;QAC1C,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,CAAC,CAAA;IACnE,CAAC;IAED,MAAM,QAAQ,GAAG,YAAY,CAAC,UAAU,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,CAAA;IAElE,6BAA6B;IAC7B,MAAM,OAAO,GAAiB,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;QACtD,MAAM,OAAO,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE;YACtC,MAAM,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;YAExB,OAAO,cAAc,CAAC,GAAG,CAAC,CAAA;QAC3B,CAAC,CAAC,CAAA;QAEF,MAAM,IAAI,GAAG,eAAe,CAAC,OAAO,CAAC,CAAA;QACrC,IAAI,CAAC,IAAI,GAAG,IAAI,CAAA;QAEhB,OAAO,IAAI,CAAA;IACZ,CAAC,CAAC,CAAA;IAEF,uBAAuB;IACvB,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IACzE,kGAAkG;IAClG,sGAAsG;IACtG,MAAM,cAAc,GAAG,+BAA+B,IAAI,CAAC,SAAS,UAAU,OAAO,MAAM,CAAA;IAE3F,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAE7D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,0BAA0B,UAAU,CAAC,MAAM,iBAAiB,CAAC,CAAA;IAClF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,cAAc,MAAM,CAAC,CAAA;IAE7C,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;QACjB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAA;QAE3D,OAAM;IACP,CAAC;IAED,mCAAmC;IACnC,MAAM,EAAE,cAAc,EAAE,GAAG,MAAM,MAAM,CAAC,0BAA0B,CAAC,CAAA;IACnE,MAAM,eAAe,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAA;IAE/C,+GAA+G;IAC/G,iHAAiH;IACjH,2DAA2D;IAC3D,MAAM,EAAE,GAAG,IAAI,cAAc,CAAW,IAAI,CAAC,UAAU,CAAC,CAAA;IACxD,EAAE,CAAC,IAAI,CAAC,2BAA2B,CAAC,CAAA,CAAC,yBAAyB;IAC9D,EAAE,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAA;IAEnC,EAAE,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;IAEvB,mFAAmF;IACnF,wCAAwC;IACxC,mFAAmF;IACnF,yDAAyD;IACzD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAA;IAE3C,MAAM,UAAU,GAAG,EAAE,CAAC,OAAO,CAC5B,gBAAgB,IAAI,CAAC,SAAS,MAAM,QAAQ,aAAa,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAC7F,CAAA;IAED,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,aAAa,GAAG,KAAK,CAAA;IAEzB,0FAA0F;IAC1F,MAAM,QAAQ,GAAG,GAAG,EAAE;QACrB,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAEhB,IAAI,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;gBACzB,UAAU,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,CAAA;YACvB,CAAC;YAED,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAClB,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;YACnB,MAAM,KAAK,CAAA;QACZ,CAAC;IACF,CAAC,CAAA;IAED,MAAM,KAAK,GAAsB,EAAE,CAAA;IACnC,MAAM,UAAU,GAAG,MAAM,CAAA;IAEzB,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,EAAE,EAAE,CAAC;QACnC,IAAI,IAAI,CAAC,SAAS,IAAI,CAAC,aAAa,EAAE,CAAC;YACtC,aAAa,GAAG,IAAI,CAAA;YAEpB,SAAQ;QACT,CAAC;QAED,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAClC,MAAM,CAAC,GAAG,cAAc,CAAC,CAAC,CAAC,CAAA;YAE3B,IAAI,CAAC,KAAK,IAAI;gBAAE,OAAO,IAAI,CAAA;YAC3B,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,CAAA;YAEtB,IAAI,GAAG,EAAE,IAAI,KAAK,SAAS,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;gBAAE,OAAO,MAAM,CAAC,QAAQ,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;YAE/E,IAAI,GAAG,EAAE,IAAI,KAAK,MAAM,IAAI,eAAe,CAAC,IAAI,CAAC,CAAC,CAAC;gBAAE,OAAO,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;YAEhF,OAAO,CAAC,CAAA;QACT,CAAC,CAAC,CAAA;QAEF,kCAAkC;QAClC,OAAO,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;YACvC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QAED,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAA;QAE9B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QAElB,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU,EAAE,CAAC;YAChC,QAAQ,EAAE,CAAA;YACV,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAA;YACxB,KAAK,CAAC,MAAM,GAAG,CAAC,CAAA;YAEhB,IAAI,QAAQ,GAAG,OAAO,KAAK,CAAC,EAAE,CAAC;gBAC9B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,QAAQ,GAAG,SAAS,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,aAAa,CAAC,CAAA;YAC1E,CAAC;QACF,CAAC;IACF,CAAC;IAED,kBAAkB;IAClB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QAClB,QAAQ,EAAE,CAAA;QACV,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAA;IACzB,CAAC;IAED,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,cAAc,QAAQ,CAAC,cAAc,EAAE,eAAe,IAAI,CAAC,SAAS,KAAK,CAAC,CAAA;IAE/F,wEAAwE;IACxE,MAAM,YAAY,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC,CAAA;IAE3D,IAAI,YAAY,EAAE,CAAC;QAClB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,sBAAsB,YAAY,CAAC,IAAI,QAAQ,CAAC,CAAA;QAErE,EAAE,CAAC,IAAI,CACN,kCAAkC,IAAI,CAAC,SAAS,IAAI,YAAY,CAAC,IAAI,QAAQ,IAAI,CAAC,SAAS,MAAM,YAAY,CAAC,IAAI,KAAK,CACvH,CAAA;IACF,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;IAE3C,MAAM,QAAQ,GAAG;QAChB,WAAW,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACrC,UAAU,EAAE,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC;QACpC,UAAU,EAAE,IAAI,CAAC,SAAS;QAC1B,OAAO,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC;QACnF,SAAS,EAAE,QAAQ;QACnB,QAAQ,EAAE,IAAK,CAAC,IAAI;KACpB,CAAA;IAED,MAAM,YAAY,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,OAAO,EAAE,gBAAgB,CAAC,CAAA;IACvE,MAAM,kBAAkB,CAAC,QAAQ,EAAE,YAAY,CAAC,CAAA;IAEhD,OAAO,CAAC,MAAM,CAAC,KAAK,CACnB,SAAS,QAAQ,CAAC,cAAc,EAAE,WAAW,IAAI,CAAC,UAAU,KAAK,aAAa,CAAC,SAAS,CAAC,IAAK,CAAC,IAAI,CAAC,KAAK,CACzG,CAAA;AACF,CAAC;AAgBD;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAyB;IACxD,IAAI,CAAC,CAAC,MAAM,UAAU,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QACxC,MAAM,IAAI,KAAK,CAAC,mBAAmB,OAAO,CAAC,KAAK,EAAE,CAAC,CAAA;IACpD,CAAC;IAED,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAA;IAE/D,MAAM,SAAS,CAAC;QACf,SAAS,EAAE,OAAO,CAAC,KAAK;QACxB,SAAS,EAAE,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,UAAU,CAAC,gBAAgB,EAAE,GAAG,CAAC;QACrE,UAAU,EAAE,OAAO,CAAC,MAAM,IAAI,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,OAAO,GAAG,KAAK,CAAC;QAC3E,UAAU,EAAE,OAAO,CAAC,MAAM,IAAI,GAAG;QACjC,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,GAAG;QACnC,SAAS,EAAE,OAAO,CAAC,IAAI,IAAI,CAAC;QAC5B,SAAS,EAAE,CAAC,OAAO,CAAC,QAAQ;QAC5B,MAAM,EAAE,OAAO,CAAC,MAAM,IAAI,KAAK;KAC/B,CAAC,CAAA;AACH,CAAC;AAED,YAAY"}
+{"version":3,"file":"ingest-csv.js","sourceRoot":"","sources":["../../lib/tools/ingest-csv.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAEH,OAAO,EAAE,aAAa,EAAE,MAAM,+BAA+B,CAAA;AAC7D,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM,4BAA4B,CAAA;AAChE,OAAO,EAAE,eAAe,EAAE,kBAAkB,EAAE,MAAM,4BAA4B,CAAA;AAGhF,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,SAAS,CAAA;AAC1D,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAE5C,mCAAmC;AAEnC;;GAEG;AACH,SAAS,mBAAmB,CAAC,GAAW;IACvC,OAAO,CACN,GAAG;SACD,IAAI,EAAE;SACN,UAAU,CAAC,UAAU,EAAE,EAAE,CAAC;SAC1B,WAAW,EAAE;SACb,UAAU,CAAC,aAAa,EAAE,GAAG,CAAC;SAC9B,UAAU,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,SAAS,CACvC,CAAA;AACF,CAAC;AAED,SAAS,YAAY,CAAC,KAAe;IACpC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAA;IAEtC,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,GAAG,CAAC,CAAC,CAAA;QAEzB,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,EAAE,CAAA;IACnD,CAAC,CAAC,CAAA;AACH,CAAC;AAcD,SAAS,cAAc,CAAC,GAAW;IAClC,IAAI,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IAElB,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC3D,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;IACzC,CAAC;IAED,oCAAoC;IACpC,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,WAAW,EAAE,CAAC;QACxG,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,OAAO,CAAC,CAAA;AACT,CAAC;AAED,SAAS,eAAe,CAAC,OAA0B;IAClD,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,SAAS,GAAG,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;QACzB,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;YAChB,SAAS,EAAE,CAAA;YAEX,SAAQ;QACT,CAAC;QAED,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACvB,QAAQ,EAAE,CAAA;QACX,CAAC;aAAM,IAAI,eAAe,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACpC,SAAS,EAAE,CAAA;QACZ,CAAC;aAAM,CAAC;YACP,SAAS,EAAE,CAAA;QACZ,CAAC;IACF,CAAC;IAED,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;IAC5B,MAAM,IAAI,GAAkB,SAAS,GAAG,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,GAAG,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAA;IAE5G,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,SAAS,GAAG,KAAK,IAAI,GAAG,EAAE,CAAA;AAC9D,CAAC;AAiBD,KAAK,UAAU,SAAS,CAAC,IAAmB;IAC3C,+GAA+G;IAC/G,8GAA8G;IAC9G,mHAAmH;IACnH,iHAAiH;IACjH,mBAAmB;IACnB,MAAM,IAAI,GAAG,GAA4B,EAAE,CAC1C,cAAc,CAAC,SAAS,CAAW,IAAI,CAAC,SAAS,EAAE;QAClD,MAAM,EAAE,KAAK;QACb,eAAe,EAAE,IAAI,CAAC,SAAS;QAC/B,IAAI,EAAE,IAAI,CAAC,SAAS;KACpB,CAAC,CAAA;IAEH,sEAAsE;IACtE,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,WAAW,IAAI,CAAC,SAAS,4BAA4B,CAAC,CAAA;IAE3E,IAAI,SAAS,GAAoB,IAAI,CAAA;IACrC,MAAM,UAAU,GAAe,EAAE,CAAA;IAEjC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,IAAI,EAAE,EAAE,CAAC;QAChC,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;YAClC,SAAS,GAAG,GAAG,CAAA;YAEf,SAAQ;QACT,CAAC;QAED,IAAI,UAAU,CAAC,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;YACzC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACrB,CAAC;aAAM,CAAC;YACP,MAAK;QACN,CAAC;IACF,CAAC;IAED,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;QAClC,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAA;IAC/C,CAAC;IAED,oDAAoD;IACpD,IAAI,UAAoB,CAAA;IAExB,IAAI,IAAI,CAAC,SAAS,IAAI,SAAS,EAAE,CAAC;QACjC,UAAU,GAAG,SAAS,CAAA;IACvB,CAAC;SAAM,CAAC;QACP,gDAAgD;QAChD,MAAM,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;QAC1C,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,CAAC,CAAA;IACnE,CAAC;IAED,MAAM,QAAQ,GAAG,YAAY,CAAC,UAAU,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,CAAA;IAElE,0DAA0D;IAC1D,MAAM,OAAO,GAAiB,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;QACtD,MAAM,OAAO,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE;YACtC,MAAM,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;YAExB,OAAO,cAAc,CAAC,GAAG,CAAC,CAAA;QAC3B,CAAC,CAAC,CAAA;QAEF,MAAM,IAAI,GAAG,eAAe,CAAC,OAAO,CAAC,CAAA;QACrC,IAAI,CAAC,IAAI,GAAG,IAAI,CAAA;QAEhB,OAAO,IAAI,CAAA;IACZ,CAAC,CAAC,CAAA;IAEF,+EAA+E;IAC/E,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IACzE,kGAAkG;IAClG,sGAAsG;IACtG,MAAM,cAAc,GAAG,+BAA+B,IAAI,CAAC,SAAS,UAAU,OAAO,MAAM,CAAA;IAE3F,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAE7D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,0BAA0B,UAAU,CAAC,MAAM,iBAAiB,CAAC,CAAA;IAClF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,cAAc,MAAM,CAAC,CAAA;IAE7C,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;QACjB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAA;QAE3D,OAAM;IACP,CAAC;IAED,oDAAoD;IACpD,MAAM,EAAE,cAAc,EAAE,GAAG,MAAM,MAAM,CAAC,0BAA0B,CAAC,CAAA;IACnE,MAAM,eAAe,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAA;IAE/C,+GAA+G;IAC/G,iHAAiH;IACjH,2DAA2D;IAC3D,MAAM,EAAE,GAAG,IAAI,cAAc,CAAW,IAAI,CAAC,UAAU,CAAC,CAAA;IACxD,EAAE,CAAC,IAAI,CAAC,2BAA2B,CAAC,CAAA,CAAC,yBAAyB;IAC9D,EAAE,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAA;IAEnC,EAAE,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;IAEvB,mFAAmF;IACnF,wCAAwC;IACxC,mFAAmF;IACnF,yDAAyD;IACzD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAA;IAE3C,MAAM,UAAU,GAAG,EAAE,CAAC,OAAO,CAC5B,gBAAgB,IAAI,CAAC,SAAS,MAAM,QAAQ,aAAa,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAC7F,CAAA;IAED,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,aAAa,GAAG,KAAK,CAAA;IAEzB,0FAA0F;IAC1F,MAAM,QAAQ,GAAG,GAAG,EAAE;QACrB,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAEhB,IAAI,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;gBACzB,UAAU,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,CAAA;YACvB,CAAC;YAED,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAClB,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;YACnB,MAAM,KAAK,CAAA;QACZ,CAAC;IACF,CAAC,CAAA;IAED,MAAM,KAAK,GAAsB,EAAE,CAAA;IACnC,MAAM,UAAU,GAAG,MAAM,CAAA;IAEzB,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,EAAE,EAAE,CAAC;QACnC,IAAI,IAAI,CAAC,SAAS,IAAI,CAAC,aAAa,EAAE,CAAC;YACtC,aAAa,GAAG,IAAI,CAAA;YAEpB,SAAQ;QACT,CAAC;QAED,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAClC,MAAM,CAAC,GAAG,cAAc,CAAC,CAAC,CAAC,CAAA;YAE3B,IAAI,CAAC,KAAK,IAAI;gBAAE,OAAO,IAAI,CAAA;YAC3B,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,CAAA;YAEtB,IAAI,GAAG,EAAE,IAAI,KAAK,SAAS,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;gBAAE,OAAO,MAAM,CAAC,QAAQ,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;YAE/E,IAAI,GAAG,EAAE,IAAI,KAAK,MAAM,IAAI,eAAe,CAAC,IAAI,CAAC,CAAC,CAAC;gBAAE,OAAO,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;YAEhF,OAAO,CAAC,CAAA;QACT,CAAC,CAAC,CAAA;QAEF,kCAAkC;QAClC,OAAO,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;YACvC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QAED,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAA;QAE9B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QAElB,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU,EAAE,CAAC;YAChC,QAAQ,EAAE,CAAA;YACV,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAA;YACxB,KAAK,CAAC,MAAM,GAAG,CAAC,CAAA;YAEhB,IAAI,QAAQ,GAAG,OAAO,KAAK,CAAC,EAAE,CAAC;gBAC9B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,QAAQ,GAAG,SAAS,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,aAAa,CAAC,CAAA;YAC1E,CAAC;QACF,CAAC;IACF,CAAC;IAED,kBAAkB;IAClB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QAClB,QAAQ,EAAE,CAAA;QACV,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAA;IACzB,CAAC;IAED,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,cAAc,QAAQ,CAAC,cAAc,EAAE,eAAe,IAAI,CAAC,SAAS,KAAK,CAAC,CAAA;IAE/F,wEAAwE;IACxE,MAAM,YAAY,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC,CAAA;IAE3D,IAAI,YAAY,EAAE,CAAC;QAClB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,sBAAsB,YAAY,CAAC,IAAI,QAAQ,CAAC,CAAA;QAErE,EAAE,CAAC,IAAI,CACN,kCAAkC,IAAI,CAAC,SAAS,IAAI,YAAY,CAAC,IAAI,QAAQ,IAAI,CAAC,SAAS,MAAM,YAAY,CAAC,IAAI,KAAK,CACvH,CAAA;IACF,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;IAE3C,MAAM,QAAQ,GAAG;QAChB,WAAW,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACrC,UAAU,EAAE,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC;QACpC,UAAU,EAAE,IAAI,CAAC,SAAS;QAC1B,OAAO,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC;QACnF,SAAS,EAAE,QAAQ;QACnB,QAAQ,EAAE,IAAK,CAAC,IAAI;KACpB,CAAA;IAED,MAAM,YAAY,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,OAAO,EAAE,gBAAgB,CAAC,CAAA;IACvE,MAAM,kBAAkB,CAAC,QAAQ,EAAE,YAAY,CAAC,CAAA;IAEhD,OAAO,CAAC,MAAM,CAAC,KAAK,CACnB,SAAS,QAAQ,CAAC,cAAc,EAAE,WAAW,IAAI,CAAC,UAAU,KAAK,aAAa,CAAC,SAAS,CAAC,IAAK,CAAC,IAAI,CAAC,KAAK,CACzG,CAAA;AACF,CAAC;AAgBD;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAyB;IACxD,IAAI,CAAC,CAAC,MAAM,UAAU,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QACxC,MAAM,IAAI,KAAK,CAAC,mBAAmB,OAAO,CAAC,KAAK,EAAE,CAAC,CAAA;IACpD,CAAC;IAED,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAA;IAE/D,MAAM,SAAS,CAAC;QACf,SAAS,EAAE,OAAO,CAAC,KAAK;QACxB,SAAS,EAAE,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,UAAU,CAAC,gBAAgB,EAAE,GAAG,CAAC;QACrE,UAAU,EAAE,OAAO,CAAC,MAAM,IAAI,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,OAAO,GAAG,KAAK,CAAC;QAC3E,UAAU,EAAE,OAAO,CAAC,MAAM,IAAI,GAAG;QACjC,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,GAAG;QACnC,SAAS,EAAE,OAAO,CAAC,IAAI,IAAI,CAAC;QAC5B,SAAS,EAAE,CAAC,OAAO,CAAC,QAAQ;QAC5B,MAAM,EAAE,OAAO,CAAC,MAAM,IAAI,KAAK;KAC/B,CAAC,CAAA;AACH,CAAC;AAED,YAAY"}

package/out/tools/lint/slice/index.d.ts

@@ -0,0 +1,93 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Corpus linter. Compares a new slice against pre-computed corpus statistics and flags patterns
+ *   that would cause the class of failure we hit with v0.6.2's "5th Avenue Theatre" adversarial
+ *   venue templates.
+ *
+ *   Per DeepSeek turn 9 design (2026-05-29). v1 checks:
+ *
+ *   1. **Token-label distribution outliers.** For each token in the new slice, compare the slice's
+ *        majority label to the corpus's majority label. Flag when the corpus has a
+ *        confidently-established majority (>66%) AND the slice's majority differs AND both have
+ *        non-trivial counts (slice ≥ 50, corpus ≥ 200).
+ *   2. **Label-vacuum tokens.** Token labeled with a tag that has ZERO instances in the corpus for that
+ *        token, despite the token being well-represented in the corpus. Stronger signal than #1 —
+ *        we're introducing a novel association, not shifting a distribution.
+ *   3. **Bigram-label collisions.** Identical (token_bigram, label_bigram) appears in slice while the
+ *        same token_bigram has a DIFFERENT majority label_bigram in the corpus. The "5th Avenue"
+ *        with [B-venue, I-venue] vs corpus's [B-house_number, I-street] case.
+ *   4. **Common-form anti-pattern rules.** Applies `lint-rules.json` — token-regex → forbidden-labels
+ *        mappings — flagging matches.
+ *   5. **Basic sanity.** Truncated rows (tokens.length !== labels.length), all-O rows >90% of slice.
+ *
+ *   Output: markdown report on stdout, optional JSON sidecar via `outJSON`. The command exits 0 if
+ *   no errors, 1 if any errors (warnings don't check). Per the design, the MANIFEST entry for a
+ *   flagged slice should require `lint_acknowledged: true` before training consumes it.
+ *
+ *   Usage: mailwoman dev lint corpus-slice\
+ *   --slice <new-slice.parquet>\
+ *   --stats <corpus-stats.json>\
+ *   [--rules <rules.json>]\
+ *   [--out-md /tmp/lint-report.md]\
+ *   [--out-json /tmp/lint-report.json]
+ */
+/**
+ * Options for {@linkcode lintCorpusSlice}.
+ */
+export interface LintCorpusSliceOptions {
+    /**
+     * The new slice parquet to lint.
+     */
+    slicePath: string;
+    /**
+     * Pre-computed corpus stats JSON (see `corpus-stats.ts`).
+     */
+    statsPath: string;
+    /**
+     * Anti-pattern rules JSON. Default: the `lint-rules.json` beside this module.
+     */
+    rulesPath?: string;
+    /**
+     * Write the markdown report here as well as stdout.
+     */
+    outMd?: string;
+    /**
+     * Write a JSON sidecar of the flags + summary here.
+     */
+    outJSON?: string;
+}
+/**
+ * One lint flag emitted by a check.
+ */
+export interface LintSliceFlag {
+    check: string;
+    severity: "error" | "warn";
+    token?: string;
+    bigram?: string;
+    sliceLabel?: string;
+    corpusLabel?: string;
+    sliceCount?: number;
+    corpusCount?: number;
+    detail: string;
+    ruleID?: string;
+}
+/**
+ * Findings summary returned by {@linkcode lintCorpusSlice}.
+ */
+export interface LintCorpusSliceSummary {
+    errors: number;
+    warnings: number;
+    findings: LintSliceFlag[];
+    /**
+     * The rendered markdown report (also printed to stdout).
+     */
+    report: string;
+}
+/**
+ * Lint a slice against corpus stats + the anti-pattern rules; print the markdown report to stdout.
+ */
+export declare function lintCorpusSlice(options: LintCorpusSliceOptions, report?: (line: string) => void): Promise<LintCorpusSliceSummary>;
+//# sourceMappingURL=index.d.ts.map

package/out/tools/lint/slice/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../../lib/tools/lint/slice/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAmCG;AA8CH;;GAEG;AACH,MAAM,WAAW,sBAAsB;IACtC;;OAEG;IACH,SAAS,EAAE,MAAM,CAAA;IACjB;;OAEG;IACH,SAAS,EAAE,MAAM,CAAA;IACjB;;OAEG;IACH,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAA;IACd;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,CAAA;CAChB;AAmFD;;GAEG;AACH,MAAM,WAAW,aAAa;IAC7B,KAAK,EAAE,MAAM,CAAA;IACb,QAAQ,EAAE,OAAO,GAAG,MAAM,CAAA;IAC1B,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,MAAM,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;CACf;AAED;;GAEG;AACH,MAAM,WAAW,sBAAsB;IACtC,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,MAAM,CAAA;IAChB,QAAQ,EAAE,aAAa,EAAE,CAAA;IACzB;;OAEG;IACH,MAAM,EAAE,MAAM,CAAA;CACd;AAqND;;GAEG;AACH,wBAAsB,eAAe,CACpC,OAAO,EAAE,sBAAsB,EAC/B,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAC7B,OAAO,CAAC,sBAAsB,CAAC,CA6DjC"}