@mailwoman/corpus 2.0.0

package/out/src/format.js

@@ -0,0 +1,151 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Render a `ComponentTag`-keyed dict into a country-localized string.
+ *
+ *   Phase 1's adapters carry ground-truth components but need a plausible `raw` string for the model
+ *   to learn from. This module bridges Mailwoman's `ComponentTag` schema to OpenCage's
+ *   `address-formatting` templates (vendored via `@fragaria/address-formatter`, MIT) so adapters
+ *   can synthesize idiomatic per-country output without each one reinventing the template logic.
+ *
+ *   `@fragaria/address-formatter` returns multi-line strings (newline-separated). This module honors
+ *   that by default and exposes a `separator` option to coerce to single-line when an adapter wants
+ *   `", "` joined output for the corpus.
+ *
+ *   Limitations (documented, not blockers for Phase 1):
+ *
+ *   - `unit`: OpenCage's vocabulary doesn't have a slot, so units are appended to the road line
+ *       (`"Pennsylvania Ave NW Apt 4B"`). Phase 2 or beyond can change this if needed.
+ *   - `intersection_a` / `intersection_b`: joined as `"<a> & <b>"` into the road field.
+ *   - `cedex` (FR): folded into `postcode` (e.g. `"75008 CEDEX 08"`) so the FR template renders it in
+ *       the right slot.
+ *   - JP-specific tags (`prefecture`, `municipality`, ...): no mapping yet — left for Phase 6 when JP
+ *       becomes a live locale.
+ */
+import addressFormatter from "@fragaria/address-formatter";
+/**
+ * After rendering, drop any component whose value isn't actually present in the formatted `raw`.
+ * OpenCage's per-country templates legitimately omit some inputs:
+ *
+ * - **FR**: regions are absorbed by the postcode and rarely render verbatim (`"75008 Paris, France"`
+ *   — `"Île-de-France"` is dropped).
+ * - **US**: state names are abbreviated (`"Oregon"` → `"OR"`); the verbatim form may not survive.
+ *
+ * Alignment downstream requires `components[tag]` to occur in `raw` (within fuzzy-match tolerance).
+ * If a component was dropped or transformed beyond Levenshtein reach, keeping it in the dict
+ * guarantees a quarantine reject. Reconciliation prunes the dropouts up front so the row arrives at
+ * alignment with a self-consistent (`raw`, `components`) pair.
+ *
+ * Comparison is case- and whitespace-insensitive. The retained value in `components` is the
+ * **original** input — alignment will fuzzy-match it against the raw character span.
+ */
+export function reconcileComponents(components, raw) {
+    const haystack = raw.toLowerCase().replace(/\s+/g, " ");
+    const out = {};
+    for (const [k, v] of Object.entries(components)) {
+        if (!v)
+            continue;
+        const needle = v.toLowerCase().replace(/\s+/g, " ");
+        if (haystack.includes(needle)) {
+            out[k] = v;
+        }
+    }
+    return out;
+}
+/**
+ * Render a component dict into an idiomatic per-country address string.
+ *
+ * Returns an empty string if `components` is empty after translation. Throws nothing — bad inputs
+ * (empty dict, unsupported tag) silently degrade to the longest meaningful prefix.
+ */
+export function formatAddress(components, country, opts = {}) {
+    const ocComponents = toOpenCageComponents(components, country);
+    if (Object.keys(ocComponents).length === 0)
+        return "";
+    const raw = addressFormatter.format(ocComponents, {
+        abbreviate: opts.abbreviate ?? false,
+        appendCountry: opts.appendCountry ?? false,
+    });
+    const trimmed = raw.replace(/\s+$/g, "");
+    return opts.separator !== undefined ? trimmed.replace(/\n+/g, opts.separator) : trimmed;
+}
+/**
+ * Translate a `ComponentTag` dict to the OpenCage vocabulary that `@fragaria/address-formatter`
+ * expects. Exported for testing and for adapters that want to pre-build the dict for batch
+ * formatting.
+ */
+export function toOpenCageComponents(components, country) {
+    const out = {};
+    const road = composeRoad(components);
+    if (road)
+        out.road = road;
+    if (components.house_number)
+        out.house_number = components.house_number;
+    if (components.venue)
+        out.house = components.venue;
+    if (components.locality)
+        out.city = components.locality;
+    if (components.dependent_locality)
+        out.suburb = components.dependent_locality;
+    if (components.subregion)
+        out.county = components.subregion;
+    if (components.region)
+        out.state = components.region;
+    const postcode = composePostcode(components);
+    if (postcode)
+        out.postcode = postcode;
+    if (components.po_box)
+        out.po_box = components.po_box;
+    if (components.attention)
+        out.attention = components.attention;
+    if (components.country)
+        out.country = components.country;
+    // country_code drives template selection, not output. Only emit it if at least one other
+    // component is present — otherwise the FR/US templates render the bare code as a fallback
+    // line ("US"), which is never what a corpus consumer wants.
+    const cc = country.trim().toLowerCase();
+    if (cc && Object.keys(out).length > 0)
+        out.country_code = cc;
+    return out;
+}
+/**
+ * Build the `road` line from prefix / particle / street / suffix / unit / intersection components.
+ * Order:
+ *
+ * ```
+ * [intersection_a & intersection_b]
+ * OR
+ * [street_prefix] [street_prefix_particle] [street] [street_suffix] [unit]
+ * ```
+ */
+function composeRoad(components) {
+    if (components.intersection_a && components.intersection_b) {
+        return `${components.intersection_a} & ${components.intersection_b}`;
+    }
+    const parts = [];
+    if (components.street_prefix)
+        parts.push(components.street_prefix);
+    if (components.street_prefix_particle)
+        parts.push(components.street_prefix_particle);
+    if (components.street)
+        parts.push(components.street);
+    if (components.street_suffix)
+        parts.push(components.street_suffix);
+    if (components.unit)
+        parts.push(components.unit);
+    return parts.join(" ").replace(/\s+/g, " ").trim();
+}
+/**
+ * Fold CEDEX into postcode for FR-style output: `"75008"` + cedex `"CEDEX 08"` → `"75008 CEDEX
+ * 08"`. If only one is present, return it. If neither, return empty.
+ */
+function composePostcode(components) {
+    const base = components.postcode?.trim() ?? "";
+    const cedex = components.cedex?.trim() ?? "";
+    if (base && cedex)
+        return `${base} ${cedex}`.replace(/\s+/g, " ");
+    return base || cedex;
+}
+//# sourceMappingURL=format.js.map

package/out/src/format.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"format.js","sourceRoot":"","sources":["../../src/format.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAEH,OAAO,gBAAgB,MAAM,6BAA6B,CAAA;AA2B1D;;;;;;;;;;;;;;;GAeG;AACH,MAAM,UAAU,mBAAmB,CAAC,UAAyB,EAAE,GAAW;IACzE,MAAM,QAAQ,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;IACvD,MAAM,GAAG,GAAkB,EAAE,CAAA;IAC7B,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC;QACjD,IAAI,CAAC,CAAC;YAAE,SAAQ;QAChB,MAAM,MAAM,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;QACnD,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;YAC/B,GAAG,CAAC,CAAiB,CAAC,GAAG,CAAC,CAAA;QAC3B,CAAC;IACF,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,aAAa,CAAC,UAAyB,EAAE,OAAe,EAAE,OAA6B,EAAE;IACxG,MAAM,YAAY,GAAG,oBAAoB,CAAC,UAAU,EAAE,OAAO,CAAC,CAAA;IAC9D,IAAI,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAA;IAErD,MAAM,GAAG,GAAG,gBAAgB,CAAC,MAAM,CAAC,YAAY,EAAE;QACjD,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,KAAK;QACpC,aAAa,EAAE,IAAI,CAAC,aAAa,IAAI,KAAK;KAC1C,CAAC,CAAA;IAEF,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IACxC,OAAO,IAAI,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,OAAO,CAAA;AACxF,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,oBAAoB,CAAC,UAAyB,EAAE,OAAe;IAC9E,MAAM,GAAG,GAA2B,EAAE,CAAA;IAEtC,MAAM,IAAI,GAAG,WAAW,CAAC,UAAU,CAAC,CAAA;IACpC,IAAI,IAAI;QAAE,GAAG,CAAC,IAAI,GAAG,IAAI,CAAA;IAEzB,IAAI,UAAU,CAAC,YAAY;QAAE,GAAG,CAAC,YAAY,GAAG,UAAU,CAAC,YAAY,CAAA;IAEvE,IAAI,UAAU,CAAC,KAAK;QAAE,GAAG,CAAC,KAAK,GAAG,UAAU,CAAC,KAAK,CAAA;IAElD,IAAI,UAAU,CAAC,QAAQ;QAAE,GAAG,CAAC,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAA;IACvD,IAAI,UAAU,CAAC,kBAAkB;QAAE,GAAG,CAAC,MAAM,GAAG,UAAU,CAAC,kBAAkB,CAAA;IAC7E,IAAI,UAAU,CAAC,SAAS;QAAE,GAAG,CAAC,MAAM,GAAG,UAAU,CAAC,SAAS,CAAA;IAC3D,IAAI,UAAU,CAAC,MAAM;QAAE,GAAG,CAAC,KAAK,GAAG,UAAU,CAAC,MAAM,CAAA;IAEpD,MAAM,QAAQ,GAAG,eAAe,CAAC,UAAU,CAAC,CAAA;IAC5C,IAAI,QAAQ;QAAE,GAAG,CAAC,QAAQ,GAAG,QAAQ,CAAA;IAErC,IAAI,UAAU,CAAC,MAAM;QAAE,GAAG,CAAC,MAAM,GAAG,UAAU,CAAC,MAAM,CAAA;IACrD,IAAI,UAAU,CAAC,SAAS;QAAE,GAAG,CAAC,SAAS,GAAG,UAAU,CAAC,SAAS,CAAA;IAE9D,IAAI,UAAU,CAAC,OAAO;QAAE,GAAG,CAAC,OAAO,GAAG,UAAU,CAAC,OAAO,CAAA;IAExD,yFAAyF;IACzF,0FAA0F;IAC1F,4DAA4D;IAC5D,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAA;IACvC,IAAI,EAAE,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC;QAAE,GAAG,CAAC,YAAY,GAAG,EAAE,CAAA;IAE5D,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;;;;;GASG;AACH,SAAS,WAAW,CAAC,UAAyB;IAC7C,IAAI,UAAU,CAAC,cAAc,IAAI,UAAU,CAAC,cAAc,EAAE,CAAC;QAC5D,OAAO,GAAG,UAAU,CAAC,cAAc,MAAM,UAAU,CAAC,cAAc,EAAE,CAAA;IACrE,CAAC;IAED,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,IAAI,UAAU,CAAC,aAAa;QAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,CAAA;IAClE,IAAI,UAAU,CAAC,sBAAsB;QAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,sBAAsB,CAAC,CAAA;IACpF,IAAI,UAAU,CAAC,MAAM;QAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAA;IACpD,IAAI,UAAU,CAAC,aAAa;QAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,CAAA;IAClE,IAAI,UAAU,CAAC,IAAI;QAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAA;IAEhD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACnD,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CAAC,UAAyB;IACjD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;IAC9C,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;IAC5C,IAAI,IAAI,IAAI,KAAK;QAAE,OAAO,GAAG,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;IACjE,OAAO,IAAI,IAAI,KAAK,CAAA;AACrB,CAAC"}

package/out/src/golden.d.ts

@@ -0,0 +1,50 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Golden eval-set validator (Phase 1 task #9 in the plan).
+ *
+ *   The golden set is hand-labeled ground truth for the neural classifier. Each entry must carry
+ *   components whose surface forms actually occur in `raw` — otherwise the entry will silently rot
+ *   the eval signal. This module:
+ *
+ *   - Defines `GoldenEntry` (schema check).
+ *   - Loads `.jsonl` files (one entry per line).
+ *   - Validates every entry: schema shape, ComponentTag membership, reachability of each component in
+ *       `raw` via the same `reconcileComponents` helper alignment uses.
+ *   - Returns a structured report of per-entry errors so the CLI / CI surface can act on it.
+ *
+ *   The 1000-entry target (500 US + 500 FR) is a human task. This module catches the regressions that
+ *   creep in over time as new entries land.
+ */
+import { type ComponentTag } from "@mailwoman/core/types";
+/** One entry in a golden `.jsonl` file. */
+export interface GoldenEntry {
+    raw: string;
+    components: Partial<Record<ComponentTag, string>>;
+    country: string;
+    source: "golden";
+    notes?: string;
+}
+/** Per-entry validation failure. */
+export interface GoldenIssue {
+    file: string;
+    line: number;
+    reason: string;
+}
+/** Aggregate report from `validateGoldenDir`. */
+export interface GoldenReport {
+    entries: number;
+    files: number;
+    issues: GoldenIssue[];
+}
+/** Parse a single JSONL line into a `GoldenEntry`. Throws on schema violations. */
+export declare function parseGoldenLine(line: string): GoldenEntry;
+/** Check that every component in `entry` appears in `entry.raw` (reconciliation-equivalent). */
+export declare function unreachableComponents(entry: GoldenEntry): ComponentTag[];
+/** Validate one `.jsonl` file end-to-end, returning a list of issues. */
+export declare function validateGoldenFile(path: string): Promise<GoldenIssue[]>;
+/** Validate every `.jsonl` in a golden directory. */
+export declare function validateGoldenDir(dir: string): Promise<GoldenReport>;
+//# sourceMappingURL=golden.d.ts.map

package/out/src/golden.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"golden.d.ts","sourceRoot":"","sources":["../../src/golden.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAkB,KAAK,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAOzE,2CAA2C;AAC3C,MAAM,WAAW,WAAW;IAC3B,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAA;IACjD,OAAO,EAAE,MAAM,CAAA;IACf,MAAM,EAAE,QAAQ,CAAA;IAChB,KAAK,CAAC,EAAE,MAAM,CAAA;CACd;AAED,oCAAoC;AACpC,MAAM,WAAW,WAAW;IAC3B,IAAI,EAAE,MAAM,CAAA;IACZ,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,MAAM,CAAA;CACd;AAED,iDAAiD;AACjD,MAAM,WAAW,YAAY;IAC5B,OAAO,EAAE,MAAM,CAAA;IACf,KAAK,EAAE,MAAM,CAAA;IACb,MAAM,EAAE,WAAW,EAAE,CAAA;CACrB;AAED,mFAAmF;AACnF,wBAAgB,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,CAyBzD;AAED,gGAAgG;AAChG,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,WAAW,GAAG,YAAY,EAAE,CAOxE;AAED,yEAAyE;AACzE,wBAAsB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,CAsB7E;AAED,qDAAqD;AACrD,wBAAsB,iBAAiB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAY1E"}

package/out/src/golden.js

@@ -0,0 +1,104 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Golden eval-set validator (Phase 1 task #9 in the plan).
+ *
+ *   The golden set is hand-labeled ground truth for the neural classifier. Each entry must carry
+ *   components whose surface forms actually occur in `raw` — otherwise the entry will silently rot
+ *   the eval signal. This module:
+ *
+ *   - Defines `GoldenEntry` (schema check).
+ *   - Loads `.jsonl` files (one entry per line).
+ *   - Validates every entry: schema shape, ComponentTag membership, reachability of each component in
+ *       `raw` via the same `reconcileComponents` helper alignment uses.
+ *   - Returns a structured report of per-entry errors so the CLI / CI surface can act on it.
+ *
+ *   The 1000-entry target (500 US + 500 FR) is a human task. This module catches the regressions that
+ *   creep in over time as new entries land.
+ */
+import { COMPONENT_TAGS } from "@mailwoman/core/types";
+import { readdir, readFile } from "node:fs/promises";
+import { extname, join } from "node:path";
+import { reconcileComponents } from "./format.js";
+const TAG_SET = new Set(COMPONENT_TAGS);
+/** Parse a single JSONL line into a `GoldenEntry`. Throws on schema violations. */
+export function parseGoldenLine(line) {
+    const obj = JSON.parse(line);
+    if (typeof obj.raw !== "string" || obj.raw.length === 0) {
+        throw new Error("missing/empty raw");
+    }
+    if (typeof obj.country !== "string" || !/^[A-Z]{2}$/u.test(obj.country)) {
+        throw new Error(`country must be ISO 3166-1 alpha-2 (got ${JSON.stringify(obj.country)})`);
+    }
+    if (obj.source !== "golden") {
+        throw new Error(`source must be "golden" (got ${JSON.stringify(obj.source)})`);
+    }
+    const components = (obj.components ?? {});
+    for (const [k, v] of Object.entries(components)) {
+        if (!TAG_SET.has(k))
+            throw new Error(`unknown ComponentTag: ${k}`);
+        if (typeof v !== "string" || v.length === 0) {
+            throw new Error(`components.${k} must be a non-empty string`);
+        }
+    }
+    return {
+        raw: obj.raw,
+        components: components,
+        country: obj.country,
+        source: "golden",
+        notes: typeof obj.notes === "string" ? obj.notes : undefined,
+    };
+}
+/** Check that every component in `entry` appears in `entry.raw` (reconciliation-equivalent). */
+export function unreachableComponents(entry) {
+    const reconciled = reconcileComponents(entry.components, entry.raw);
+    const missing = [];
+    for (const tag of Object.keys(entry.components)) {
+        if (!(tag in reconciled))
+            missing.push(tag);
+    }
+    return missing;
+}
+/** Validate one `.jsonl` file end-to-end, returning a list of issues. */
+export async function validateGoldenFile(path) {
+    const text = await readFile(path, "utf8");
+    const lines = text.split("\n");
+    const issues = [];
+    for (let i = 0; i < lines.length; i++) {
+        const line = lines[i].trim();
+        if (!line)
+            continue;
+        try {
+            const entry = parseGoldenLine(line);
+            const unreachable = unreachableComponents(entry);
+            if (unreachable.length > 0) {
+                issues.push({
+                    file: path,
+                    line: i + 1,
+                    reason: `components not reachable in raw: ${unreachable.join(", ")}`,
+                });
+            }
+        }
+        catch (err) {
+            issues.push({ file: path, line: i + 1, reason: err.message });
+        }
+    }
+    return issues;
+}
+/** Validate every `.jsonl` in a golden directory. */
+export async function validateGoldenDir(dir) {
+    const files = (await readdir(dir)).filter((n) => extname(n) === ".jsonl").sort();
+    const issues = [];
+    let entries = 0;
+    for (const name of files) {
+        const fullPath = join(dir, name);
+        const fileIssues = await validateGoldenFile(fullPath);
+        issues.push(...fileIssues);
+        const text = await readFile(fullPath, "utf8");
+        entries += text.split("\n").filter((l) => l.trim()).length;
+    }
+    return { entries, files: files.length, issues };
+}
+//# sourceMappingURL=golden.js.map

package/out/src/golden.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"golden.js","sourceRoot":"","sources":["../../src/golden.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,cAAc,EAAqB,MAAM,uBAAuB,CAAA;AACzE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AACpD,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AACzC,OAAO,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAA;AAEjD,MAAM,OAAO,GAAG,IAAI,GAAG,CAAS,cAAmC,CAAC,CAAA;AAyBpE,mFAAmF;AACnF,MAAM,UAAU,eAAe,CAAC,IAAY;IAC3C,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAmD,CAAA;IAC9E,IAAI,OAAO,GAAG,CAAC,GAAG,KAAK,QAAQ,IAAI,GAAG,CAAC,GAAG,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACzD,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAA;IACrC,CAAC;IACD,IAAI,OAAO,GAAG,CAAC,OAAO,KAAK,QAAQ,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;QACzE,MAAM,IAAI,KAAK,CAAC,2CAA2C,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;IAC3F,CAAC;IACD,IAAI,GAAG,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,gCAAgC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAA;IAC/E,CAAC;IACD,MAAM,UAAU,GAAG,CAAC,GAAG,CAAC,UAAU,IAAI,EAAE,CAA4B,CAAA;IACpE,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC;QACjD,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,yBAAyB,CAAC,EAAE,CAAC,CAAA;QAClE,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC7C,MAAM,IAAI,KAAK,CAAC,cAAc,CAAC,6BAA6B,CAAC,CAAA;QAC9D,CAAC;IACF,CAAC;IACD,OAAO;QACN,GAAG,EAAE,GAAG,CAAC,GAAG;QACZ,UAAU,EAAE,UAAuC;QACnD,OAAO,EAAE,GAAG,CAAC,OAAO;QACpB,MAAM,EAAE,QAAQ;QAChB,KAAK,EAAE,OAAO,GAAG,CAAC,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS;KAC5D,CAAA;AACF,CAAC;AAED,gGAAgG;AAChG,MAAM,UAAU,qBAAqB,CAAC,KAAkB;IACvD,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,GAAG,CAAC,CAAA;IACnE,MAAM,OAAO,GAAmB,EAAE,CAAA;IAClC,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAmB,EAAE,CAAC;QACnE,IAAI,CAAC,CAAC,GAAG,IAAI,UAAU,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IAC5C,CAAC;IACD,OAAO,OAAO,CAAA;AACf,CAAC;AAED,yEAAyE;AACzE,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,IAAY;IACpD,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IACzC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IAC9B,MAAM,MAAM,GAAkB,EAAE,CAAA;IAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAA;QAC7B,IAAI,CAAC,IAAI;YAAE,SAAQ;QACnB,IAAI,CAAC;YACJ,MAAM,KAAK,GAAG,eAAe,CAAC,IAAI,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,qBAAqB,CAAC,KAAK,CAAC,CAAA;YAChD,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5B,MAAM,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,IAAI;oBACV,IAAI,EAAE,CAAC,GAAG,CAAC;oBACX,MAAM,EAAE,oCAAoC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE;iBACpE,CAAC,CAAA;YACH,CAAC;QACF,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACd,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,GAAG,CAAC,EAAE,MAAM,EAAG,GAAa,CAAC,OAAO,EAAE,CAAC,CAAA;QACzE,CAAC;IACF,CAAC;IACD,OAAO,MAAM,CAAA;AACd,CAAC;AAED,qDAAqD;AACrD,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,GAAW;IAClD,MAAM,KAAK,GAAG,CAAC,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;IAChF,MAAM,MAAM,GAAkB,EAAE,CAAA;IAChC,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QAC1B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;QAChC,MAAM,UAAU,GAAG,MAAM,kBAAkB,CAAC,QAAQ,CAAC,CAAA;QACrD,MAAM,CAAC,IAAI,CAAC,GAAG,UAAU,CAAC,CAAA;QAC1B,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;QAC7C,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAA;IAC3D,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,MAAM,EAAE,MAAM,EAAE,CAAA;AAChD,CAAC"}

package/out/src/index.d.ts

@@ -0,0 +1,18 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export * from "./adapter.js";
+export * from "./adapters/index.js";
+export * from "./align.js";
+export * from "./build.js";
+export * from "./format.js";
+export * from "./golden.js";
+export * from "./parquet.js";
+export * from "./runner.js";
+export * from "./split.js";
+export * from "./synthesize.js";
+export * from "./tokenize.js";
+export * from "./types.js";
+//# sourceMappingURL=index.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,YAAY,CAAA;AAC1B,cAAc,YAAY,CAAA;AAC1B,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA;AAC3B,cAAc,cAAc,CAAA;AAC5B,cAAc,aAAa,CAAA;AAC3B,cAAc,YAAY,CAAA;AAC1B,cAAc,iBAAiB,CAAA;AAC/B,cAAc,eAAe,CAAA;AAC7B,cAAc,YAAY,CAAA"}

package/out/src/index.js

@@ -0,0 +1,18 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export * from "./adapter.js";
+export * from "./adapters/index.js";
+export * from "./align.js";
+export * from "./build.js";
+export * from "./format.js";
+export * from "./golden.js";
+export * from "./parquet.js";
+export * from "./runner.js";
+export * from "./split.js";
+export * from "./synthesize.js";
+export * from "./tokenize.js";
+export * from "./types.js";
+//# sourceMappingURL=index.js.map

package/out/src/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,cAAc,CAAA;AAC5B,cAAc,qBAAqB,CAAA;AACnC,cAAc,YAAY,CAAA;AAC1B,cAAc,YAAY,CAAA;AAC1B,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA;AAC3B,cAAc,cAAc,CAAA;AAC5B,cAAc,aAAa,CAAA;AAC3B,cAAc,YAAY,CAAA;AAC1B,cAAc,iBAAiB,CAAA;AAC/B,cAAc,eAAe,CAAA;AAC7B,cAAc,YAAY,CAAA"}

package/out/src/parquet-wrapper/index.d.ts

@@ -0,0 +1,12 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0).
+ */
+export * from "./reader.js";
+export * from "./schema.js";
+export * from "./writer.js";
+//# sourceMappingURL=index.d.ts.map

package/out/src/parquet-wrapper/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/parquet-wrapper/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA"}

package/out/src/parquet-wrapper/index.js

@@ -0,0 +1,12 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0).
+ */
+export * from "./reader.js";
+export * from "./schema.js";
+export * from "./writer.js";
+//# sourceMappingURL=index.js.map

package/out/src/parquet-wrapper/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/parquet-wrapper/index.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA;AAC3B,cAAc,aAAa,CAAA"}

package/out/src/parquet-wrapper/reader.d.ts

@@ -0,0 +1,31 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Typed wrapper around `@dsnp/parquetjs`'s `ParquetReader` that narrows the row-iterator generic to
+ *   a user-supplied record type and adds `AsyncDisposable` support so `await using` cleans up the
+ *   envelope reader without an explicit `close()`.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/reader.ts` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0). Two
+ *   trims relative to the original: (a) removed the
+ *   `@isp.nexus/core/polyfills/promises/withResolvers` import — Node 22 has it native; (b) replaced
+ *   the `PathBuilderLike` (path-ts) type on `openFile` with the plain `string | URL` the
+ *   `@dsnp/parquetjs` envelope reader accepts directly.
+ */
+import { ParquetReader as BaseParquetReader } from "@dsnp/parquetjs";
+import type { BufferReaderOptions } from "@dsnp/parquetjs/dist/lib/bufferReader.js";
+import { ParquetEnvelopeReader } from "@dsnp/parquetjs/dist/lib/reader.js";
+import { type ParquetRecordLike, ParquetSchema } from "./schema.js";
+/** A typed Parquet reader, wrapping the base Parquet reader. */
+export declare class ParquetReader<T extends ParquetRecordLike> extends BaseParquetReader implements AsyncDisposable {
+    schema: ParquetSchema<T>;
+    static openFile<T extends ParquetRecordLike>(filePath: string | URL, options?: BufferReaderOptions): Promise<ParquetReader<T>>;
+    static openBuffer<T extends ParquetRecordLike>(buffer: Buffer, options?: BufferReaderOptions): Promise<ParquetReader<T>>;
+    static openEnvelopeReader<T extends ParquetRecordLike>(envelopeReader: ParquetEnvelopeReader, opts?: BufferReaderOptions): Promise<ParquetReader<T>>;
+    [Symbol.asyncIterator](): AsyncGenerator<T, void, unknown>;
+    [Symbol.asyncDispose](): Promise<void>;
+    dispose(): Promise<void>;
+}
+//# sourceMappingURL=reader.d.ts.map

package/out/src/parquet-wrapper/reader.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"reader.d.ts","sourceRoot":"","sources":["../../../src/parquet-wrapper/reader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,iBAAiB,CAAA;AACpE,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,0CAA0C,CAAA;AACnF,OAAO,EAAE,qBAAqB,EAAE,MAAM,oCAAoC,CAAA;AAC1E,OAAO,EAAE,KAAK,iBAAiB,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AAEnE,gEAAgE;AAChE,qBAAa,aAAa,CAAC,CAAC,SAAS,iBAAiB,CAAE,SAAQ,iBAAkB,YAAW,eAAe;IACnG,MAAM,EAAE,aAAa,CAAC,CAAC,CAAC,CAAA;WAEV,QAAQ,CAAC,CAAC,SAAS,iBAAiB,EACzD,QAAQ,EAAE,MAAM,GAAG,GAAG,EACtB,OAAO,CAAC,EAAE,mBAAmB,GAC3B,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC;WAMN,UAAU,CAAC,CAAC,SAAS,iBAAiB,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,mBAAmB;WAMrF,kBAAkB,CAAC,CAAC,SAAS,iBAAiB,EACnE,cAAc,EAAE,qBAAqB,EACrC,IAAI,CAAC,EAAE,mBAAmB;IAkBX,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,cAAc,CAAC,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC;IAI7D,CAAC,MAAM,CAAC,YAAY,CAAC;IAIrB,OAAO;CAGpB"}

package/out/src/parquet-wrapper/reader.js

@@ -0,0 +1,54 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Typed wrapper around `@dsnp/parquetjs`'s `ParquetReader` that narrows the row-iterator generic to
+ *   a user-supplied record type and adds `AsyncDisposable` support so `await using` cleans up the
+ *   envelope reader without an explicit `close()`.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/reader.ts` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0). Two
+ *   trims relative to the original: (a) removed the
+ *   `@isp.nexus/core/polyfills/promises/withResolvers` import — Node 22 has it native; (b) replaced
+ *   the `PathBuilderLike` (path-ts) type on `openFile` with the plain `string | URL` the
+ *   `@dsnp/parquetjs` envelope reader accepts directly.
+ */
+import { ParquetReader as BaseParquetReader } from "@dsnp/parquetjs";
+import { ParquetEnvelopeReader } from "@dsnp/parquetjs/dist/lib/reader.js";
+import { ParquetSchema } from "./schema.js";
+/** A typed Parquet reader, wrapping the base Parquet reader. */
+export class ParquetReader extends BaseParquetReader {
+    static async openFile(filePath, options) {
+        const envelopeReader = await ParquetEnvelopeReader.openFile(filePath.toString(), options);
+        return ParquetReader.openEnvelopeReader(envelopeReader, options);
+    }
+    static async openBuffer(buffer, options) {
+        const envelopeReader = await ParquetEnvelopeReader.openBuffer(buffer, options);
+        return this.openEnvelopeReader(envelopeReader, options);
+    }
+    static async openEnvelopeReader(envelopeReader, opts) {
+        if (opts?.metadata) {
+            return new ParquetReader(opts.metadata, envelopeReader, opts);
+        }
+        try {
+            await envelopeReader.readHeader();
+            const metadata = await envelopeReader.readFooter();
+            return new ParquetReader(metadata, envelopeReader, opts);
+        }
+        catch (err) {
+            await envelopeReader.close();
+            throw err;
+        }
+    }
+    [Symbol.asyncIterator]() {
+        return super[Symbol.asyncIterator]();
+    }
+    async [Symbol.asyncDispose]() {
+        return this.close();
+    }
+    async dispose() {
+        return this[Symbol.asyncDispose]();
+    }
+}
+//# sourceMappingURL=reader.js.map

package/out/src/parquet-wrapper/reader.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reader.js","sourceRoot":"","sources":["../../../src/parquet-wrapper/reader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,iBAAiB,CAAA;AAEpE,OAAO,EAAE,qBAAqB,EAAE,MAAM,oCAAoC,CAAA;AAC1E,OAAO,EAA0B,aAAa,EAAE,MAAM,aAAa,CAAA;AAEnE,gEAAgE;AAChE,MAAM,OAAO,aAA2C,SAAQ,iBAAiB;IAGhF,MAAM,CAAU,KAAK,CAAC,QAAQ,CAC7B,QAAsB,EACtB,OAA6B;QAE7B,MAAM,cAAc,GAAG,MAAM,qBAAqB,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,EAAE,EAAE,OAAO,CAAC,CAAA;QAEzF,OAAO,aAAa,CAAC,kBAAkB,CAAI,cAAc,EAAE,OAAO,CAAC,CAAA;IACpE,CAAC;IAED,MAAM,CAAU,KAAK,CAAC,UAAU,CAA8B,MAAc,EAAE,OAA6B;QAC1G,MAAM,cAAc,GAAG,MAAM,qBAAqB,CAAC,UAAU,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;QAE9E,OAAO,IAAI,CAAC,kBAAkB,CAAI,cAAc,EAAE,OAAO,CAAC,CAAA;IAC3D,CAAC;IAED,MAAM,CAAU,KAAK,CAAC,kBAAkB,CACvC,cAAqC,EACrC,IAA0B;QAE1B,IAAI,IAAI,EAAE,QAAQ,EAAE,CAAC;YACpB,OAAO,IAAI,aAAa,CAAI,IAAI,CAAC,QAAQ,EAAE,cAAc,EAAE,IAAI,CAAC,CAAA;QACjE,CAAC;QAED,IAAI,CAAC;YACJ,MAAM,cAAc,CAAC,UAAU,EAAE,CAAA;YAEjC,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAC,UAAU,EAAE,CAAA;YAElD,OAAO,IAAI,aAAa,CAAI,QAAQ,EAAE,cAAc,EAAE,IAAI,CAAC,CAAA;QAC5D,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACd,MAAM,cAAc,CAAC,KAAK,EAAE,CAAA;YAC5B,MAAM,GAAG,CAAA;QACV,CAAC;IACF,CAAC;IAEe,CAAC,MAAM,CAAC,aAAa,CAAC;QACrC,OAAO,KAAK,CAAC,MAAM,CAAC,aAAa,CAAC,EAAsC,CAAA;IACzE,CAAC;IAEM,KAAK,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC;QACjC,OAAO,IAAI,CAAC,KAAK,EAAE,CAAA;IACpB,CAAC;IAEM,KAAK,CAAC,OAAO;QACnB,OAAO,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAA;IACnC,CAAC;CACD"}

package/out/src/parquet-wrapper/schema.d.ts

@@ -0,0 +1,45 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Typed wrapper around `@dsnp/parquetjs`'s schema definition. Adds:
+ *
+ *   - `ParquetSchema<T>`: a generic class narrowing the base schema's `schema` property to a
+ *       field-by-field typed dict.
+ *   - `ParquetSchemaDefinitionCache`: an LRU lookup so hot paths that compute the same schema
+ *       repeatedly pay the cost once. Implements `Disposable` so `using` works.
+ *   - `createBloomFilters<T>`: helper that takes a schema and a list of columns and returns the
+ *       `@dsnp/parquetjs`-shaped bloom-filter spec array.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/schema.ts` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0). Two
+ *   trims relative to the original: (a) `ParquetSchemaDefinitionCache`'s LRU generics use
+ *   `ParquetRecordLike` instead of `any` to satisfy this project's eslint config; (b)
+ *   `Symbol.dispose` is sync (the original was async, but `Disposable`'s contract is sync — async
+ *   cleanup belongs on `AsyncDisposable`).
+ */
+import { ParquetSchema as BaseParquetSchema } from "@dsnp/parquetjs";
+import type { createSBBFParams as BloomFilterCreation } from "@dsnp/parquetjs/dist/lib/bloomFilterIO/bloomFilterWriter.js";
+import type { FieldDefinition } from "@dsnp/parquetjs/dist/lib/declare.js";
+import { LRUCache } from "lru-cache";
+/** A Parquet record-like object, i.e. a record with string keys and JSON-serializable values. */
+export type ParquetRecordLike = {
+    [key: string]: unknown | undefined;
+};
+/** Typed Parquet schema definition. */
+export type ParquetSchemaDefinition<T = ParquetRecordLike> = {
+    [field in Extract<keyof T, string>]: FieldDefinition;
+};
+/** Typed Parquet schema. */
+export declare class ParquetSchema<T> extends BaseParquetSchema {
+    schema: ParquetSchemaDefinition<T>;
+}
+/** Given a Parquet schema and a list of columns, create a list of Bloom filters for those columns. */
+export declare function createBloomFilters<T>(parquetSchemaDef: ParquetSchemaDefinition<T>, columns: Extract<keyof T, string>[]): BloomFilterCreation[];
+export declare class ParquetSchemaDefinitionCache extends LRUCache<ParquetSchemaDefinition<ParquetRecordLike>, ParquetSchema<ParquetRecordLike>> implements Disposable {
+    constructor(max?: number);
+    findOrCreateSchema<T extends ParquetRecordLike>(schemaDef: ParquetSchemaDefinition<T>): ParquetSchema<T>;
+    [Symbol.dispose](): void;
+}
+//# sourceMappingURL=schema.d.ts.map

package/out/src/parquet-wrapper/schema.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"schema.d.ts","sourceRoot":"","sources":["../../../src/parquet-wrapper/schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,iBAAiB,CAAA;AACpE,OAAO,KAAK,EAAE,gBAAgB,IAAI,mBAAmB,EAAE,MAAM,6DAA6D,CAAA;AAC1H,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,qCAAqC,CAAA;AAC1E,OAAO,EAAE,QAAQ,EAAE,MAAM,WAAW,CAAA;AAEpC,iGAAiG;AACjG,MAAM,MAAM,iBAAiB,GAAG;IAC/B,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,GAAG,SAAS,CAAA;CAClC,CAAA;AAED,uCAAuC;AACvC,MAAM,MAAM,uBAAuB,CAAC,CAAC,GAAG,iBAAiB,IAAI;KAC3D,KAAK,IAAI,OAAO,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC,GAAG,eAAe;CACpD,CAAA;AAED,4BAA4B;AAC5B,qBAAa,aAAa,CAAC,CAAC,CAAE,SAAQ,iBAAiB;IAC9C,MAAM,EAAE,uBAAuB,CAAC,CAAC,CAAC,CAAA;CAC1C;AAED,sGAAsG;AACtG,wBAAgB,kBAAkB,CAAC,CAAC,EACnC,gBAAgB,EAAE,uBAAuB,CAAC,CAAC,CAAC,EAC5C,OAAO,EAAE,OAAO,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC,EAAE,yBAanC;AAED,qBAAa,4BACZ,SAAQ,QAAQ,CAAC,uBAAuB,CAAC,iBAAiB,CAAC,EAAE,aAAa,CAAC,iBAAiB,CAAC,CAC7F,YAAW,UAAU;gBAET,GAAG,SAAO;IAIf,kBAAkB,CAAC,CAAC,SAAS,iBAAiB,EAAE,SAAS,EAAE,uBAAuB,CAAC,CAAC,CAAC,GAAG,aAAa,CAAC,CAAC,CAAC;IAYxG,CAAC,MAAM,CAAC,OAAO,CAAC;CAGvB"}

package/out/src/parquet-wrapper/schema.js

@@ -0,0 +1,55 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Typed wrapper around `@dsnp/parquetjs`'s schema definition. Adds:
+ *
+ *   - `ParquetSchema<T>`: a generic class narrowing the base schema's `schema` property to a
+ *       field-by-field typed dict.
+ *   - `ParquetSchemaDefinitionCache`: an LRU lookup so hot paths that compute the same schema
+ *       repeatedly pay the cost once. Implements `Disposable` so `using` works.
+ *   - `createBloomFilters<T>`: helper that takes a schema and a list of columns and returns the
+ *       `@dsnp/parquetjs`-shaped bloom-filter spec array.
+ *
+ *   Salvaged 2026-05-17 from `isp-nexus/universe@6eeb7bd99643a6d62a8b8abbd50968a1e492b90b`
+ *   `sdk/parquet/schema.ts` (originally copyright OpenISP, Inc.; both projects are AGPL-3.0). Two
+ *   trims relative to the original: (a) `ParquetSchemaDefinitionCache`'s LRU generics use
+ *   `ParquetRecordLike` instead of `any` to satisfy this project's eslint config; (b)
+ *   `Symbol.dispose` is sync (the original was async, but `Disposable`'s contract is sync — async
+ *   cleanup belongs on `AsyncDisposable`).
+ */
+import { ParquetSchema as BaseParquetSchema } from "@dsnp/parquetjs";
+import { LRUCache } from "lru-cache";
+/** Typed Parquet schema. */
+export class ParquetSchema extends BaseParquetSchema {
+}
+/** Given a Parquet schema and a list of columns, create a list of Bloom filters for those columns. */
+export function createBloomFilters(parquetSchemaDef, columns) {
+    const bloomFilters = [];
+    for (const column of columns) {
+        if (!parquetSchemaDef[column]) {
+            throw new Error(`Bloom filter column ${column} not found in Parquet schema`);
+        }
+        bloomFilters.push({ column });
+    }
+    return bloomFilters;
+}
+export class ParquetSchemaDefinitionCache extends LRUCache {
+    constructor(max = 1000) {
+        super({ max });
+    }
+    findOrCreateSchema(schemaDef) {
+        const key = schemaDef;
+        let schema = this.get(key);
+        if (!schema) {
+            schema = new ParquetSchema(schemaDef);
+            this.set(key, schema);
+        }
+        return schema;
+    }
+    [Symbol.dispose]() {
+        this.clear();
+    }
+}
+//# sourceMappingURL=schema.js.map

package/out/src/parquet-wrapper/schema.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"schema.js","sourceRoot":"","sources":["../../../src/parquet-wrapper/schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,iBAAiB,CAAA;AAGpE,OAAO,EAAE,QAAQ,EAAE,MAAM,WAAW,CAAA;AAYpC,4BAA4B;AAC5B,MAAM,OAAO,aAAiB,SAAQ,iBAAiB;CAEtD;AAED,sGAAsG;AACtG,MAAM,UAAU,kBAAkB,CACjC,gBAA4C,EAC5C,OAAmC;IAEnC,MAAM,YAAY,GAA0B,EAAE,CAAA;IAE9C,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,IAAI,CAAC,gBAAgB,CAAC,MAAM,CAAC,EAAE,CAAC;YAC/B,MAAM,IAAI,KAAK,CAAC,uBAAuB,MAAM,8BAA8B,CAAC,CAAA;QAC7E,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,CAAA;IAC9B,CAAC;IAED,OAAO,YAAY,CAAA;AACpB,CAAC;AAED,MAAM,OAAO,4BACZ,SAAQ,QAAsF;IAG9F,YAAY,GAAG,GAAG,IAAI;QACrB,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,CAAA;IACf,CAAC;IAEM,kBAAkB,CAA8B,SAAqC;QAC3F,MAAM,GAAG,GAAG,SAAuD,CAAA;QACnE,IAAI,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAiC,CAAA;QAE1D,IAAI,CAAC,MAAM,EAAE,CAAC;YACb,MAAM,GAAG,IAAI,aAAa,CAAI,SAAS,CAAC,CAAA;YACxC,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,MAA0C,CAAC,CAAA;QAC1D,CAAC;QAED,OAAO,MAAM,CAAA;IACd,CAAC;IAEM,CAAC,MAAM,CAAC,OAAO,CAAC;QACtB,IAAI,CAAC,KAAK,EAAE,CAAA;IACb,CAAC;CACD"}