dsh-data-quality 0.1.0

package/lib/types/events.js

@@ -0,0 +1,41 @@
+/**
+ * The `data-quality/*` session-event vocabulary and its adaptive append gate.
+ * The vocabulary is declared on `SessionEventMap` for type-checked payloads;
+ * the runtime gate decides per host whether appending is safe:
+ *
+ * - Hosts whose known-type set already covers the vocabulary (a future harness
+ *   that adopts these events) append plainly.
+ * - Hosts with an `ignorable` append option (the master-build
+ *   `Session.append(type, data, { ignorable: true })` contract) append with the
+ *   marker, so builds that do not know the type skip the event on restore.
+ * - 0.1.0-rc.6 hosts have neither a registration surface nor the `ignorable`
+ *   append flag; appending an unknown type there would make the persistence
+ *   coordinator refuse the session log on restore, so the append is skipped
+ *   and the storage-domain report remains the durable copy.
+ * @module dsh-data-quality/events
+ */
+import { KNOWN_SESSION_EVENT_TYPES } from '@deepseek-ai/dsh-session';
+/** The event type vocabulary this plugin appends. */
+export const DATA_QUALITY_EVENT_TYPES = ['data-quality/profile', 'data-quality/clean', 'data-quality/verify'];
+/**
+ * Append one `data-quality/*` event when the host can carry it safely; skip
+ * silently otherwise (the storage-domain report is always the durable copy).
+ * The `ignorable` probe reads the UNBOUND method's source (a `.bind()` result
+ * reports `[native code]`): the rc.6 build contains no `ignorable` handling
+ * while the master build references the flag by name; property names survive
+ * minification, so the probe fails safe (skips) rather than corrupting a log.
+ * @param session - the calling session.
+ * @param type - the event type.
+ * @param data - the payload.
+ */
+export function appendDataQualityEvent(session, type, data) {
+    if (KNOWN_SESSION_EVENT_TYPES.has(type)) {
+        session.append(type, data);
+        return;
+    }
+    const append = session.append;
+    if (Function.prototype.toString.call(append).includes('ignorable')) {
+        append.call(session, type, data, { ignorable: true });
+    }
+}
+//# sourceMappingURL=events.js.map

package/lib/types/events.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"events.js","sourceRoot":"","sources":["../../src/events.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,yBAAyB,EAAgB,MAAM,0BAA0B,CAAA;AA2ClF,qDAAqD;AACrD,MAAM,CAAC,MAAM,wBAAwB,GAAG,CAAC,sBAAsB,EAAE,oBAAoB,EAAE,qBAAqB,CAAU,CAAA;AAQtH;;;;;;;;;;GAUG;AACH,MAAM,UAAU,sBAAsB,CAAC,OAAgB,EAAE,IAA0B,EAAE,IAA0B;IAC7G,IAAI,yBAAyB,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;QACxC,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QAC1B,OAAM;IACR,CAAC;IACD,MAAM,MAAM,GAAG,OAAO,CAAC,MAAqB,CAAA;IAC5C,IAAI,QAAQ,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;QACnE,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;IACvD,CAAC;AACH,CAAC"}

package/lib/types/index.d.ts

@@ -0,0 +1,45 @@
+/**
+ * `dsh-data-quality` — deterministic data profiling, cleaning, and
+ * verification for DeepSeek Harness. Mounts the `ctx.dataQuality` capability
+ * seam (Service Definition in `service.ts`, local Provider in
+ * `provider-local.ts`), registers the `data_profile` / `data_clean` /
+ * `data_verify` model tools (Consumers — all computation goes through the
+ * service layer), persists run reports to the `data_quality` storage domain,
+ * and appends adaptive `data-quality/*` session events on hosts that can
+ * carry them.
+ *
+ * Function plugin — no default export (the Loader unwraps
+ * `exports.default ?? exports`, and a stray default would discard
+ * `name`/`inject`/`Config`/`apply`).
+ * @module dsh-data-quality
+ */
+import type { Context } from '@deepseek-ai/cordis';
+import { Config } from './config.js';
+export declare const name = "data-quality";
+/** The three model tools and the durable report domain. */
+export declare const inject: string[];
+export { Config, resolveConfig } from './config.js';
+export type { Config as DataQualityConfig, ResolvedConfig } from './config.js';
+export { VERSION } from './version.js';
+export { DataQualityService, type CitationCheckRequest, type CitationCheckResult, type ProfileRequest, type CleanRequest, type CleanRunReport, type VerifyRequest, } from './service.js';
+export { LocalDataQualityService, type ProviderDeps } from './provider-local.js';
+export type { CleanRule, CleanRuleLog, CleanResult } from './clean.js';
+export type { ProfileReport, ColumnProfile, NumericProfile } from './profile.js';
+export type { VerifyRule, VerifyReport, VerifyRuleResult, VerifyEvidenceRow } from './verify.js';
+export { parseLocator, checkCitations } from './verify.js';
+export { profileTable, renderProfileText } from './profile.js';
+export { applyCleanRules, serializeDelimited } from './clean.js';
+export { verifyTable, renderVerifyText } from './verify.js';
+export { DatasetError, loadTable, loadDocument, parseDelimited, parseJsonTable, resolveWorkspacePath, sampleRows, isMissing, parseNumeric, parseDate, parseBoolean, type Table, type Row, type Cell, type DocumentRoot, } from './dataset.js';
+export { dataQualityDomainSpec, reportKeyOf, reportRecordSchema, type ReportRecord, type ReportStore } from './store.js';
+export { DATA_QUALITY_EVENT_TYPES, appendDataQualityEvent, type DataQualityEventData, type DataQualityEventType } from './events.js';
+export { MAX_CELL_TEXT, truncateCell, truncateRow } from './present.js';
+/**
+ * Mount the seam: resolve config (fail loud), open the report domain, publish
+ * `ctx.dataQuality`, and register the three tools. With `enabled: false` the
+ * plugin registers nothing and stays inert.
+ * @param ctx - the plugin context (host).
+ * @param config - raw plugin config.
+ */
+export declare function apply(ctx: Context, config?: Config): Promise<void>;
+//# sourceMappingURL=index.d.ts.map

package/lib/types/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAA;AAGlD,OAAO,EAAE,MAAM,EAAiB,MAAM,aAAa,CAAA;AAQnD,eAAO,MAAM,IAAI,iBAAiB,CAAA;AAClC,2DAA2D;AAC3D,eAAO,MAAM,MAAM,UAA6B,CAAA;AAEhD,OAAO,EAAE,MAAM,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AACnD,YAAY,EAAE,MAAM,IAAI,iBAAiB,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC9E,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AACtC,OAAO,EACL,kBAAkB,EAClB,KAAK,oBAAoB,EACzB,KAAK,mBAAmB,EACxB,KAAK,cAAc,EACnB,KAAK,YAAY,EACjB,KAAK,cAAc,EACnB,KAAK,aAAa,GACnB,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,uBAAuB,EAAE,KAAK,YAAY,EAAE,MAAM,qBAAqB,CAAA;AAChF,YAAY,EAAE,SAAS,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AACtE,YAAY,EAAE,aAAa,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,cAAc,CAAA;AAChF,YAAY,EAAE,UAAU,EAAE,YAAY,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAA;AAChG,OAAO,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC1D,OAAO,EAAE,YAAY,EAAE,iBAAiB,EAAE,MAAM,cAAc,CAAA;AAC9D,OAAO,EAAE,eAAe,EAAE,kBAAkB,EAAE,MAAM,YAAY,CAAA;AAChE,OAAO,EAAE,WAAW,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAA;AAC3D,OAAO,EACL,YAAY,EACZ,SAAS,EACT,YAAY,EACZ,cAAc,EACd,cAAc,EACd,oBAAoB,EACpB,UAAU,EACV,SAAS,EACT,YAAY,EACZ,SAAS,EACT,YAAY,EACZ,KAAK,KAAK,EACV,KAAK,GAAG,EACR,KAAK,IAAI,EACT,KAAK,YAAY,GAClB,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,qBAAqB,EAAE,WAAW,EAAE,kBAAkB,EAAE,KAAK,YAAY,EAAE,KAAK,WAAW,EAAE,MAAM,YAAY,CAAA;AACxH,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAE,KAAK,oBAAoB,EAAE,KAAK,oBAAoB,EAAE,MAAM,aAAa,CAAA;AACpI,OAAO,EAAE,aAAa,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,cAAc,CAAA;AAEvE;;;;;;GAMG;AACH,wBAAsB,KAAK,CAAC,GAAG,EAAE,OAAO,EAAE,MAAM,GAAE,MAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAmC5E"}

package/lib/types/index.js

@@ -0,0 +1,78 @@
+/**
+ * `dsh-data-quality` — deterministic data profiling, cleaning, and
+ * verification for DeepSeek Harness. Mounts the `ctx.dataQuality` capability
+ * seam (Service Definition in `service.ts`, local Provider in
+ * `provider-local.ts`), registers the `data_profile` / `data_clean` /
+ * `data_verify` model tools (Consumers — all computation goes through the
+ * service layer), persists run reports to the `data_quality` storage domain,
+ * and appends adaptive `data-quality/*` session events on hosts that can
+ * carry them.
+ *
+ * Function plugin — no default export (the Loader unwraps
+ * `exports.default ?? exports`, and a stray default would discard
+ * `name`/`inject`/`Config`/`apply`).
+ * @module dsh-data-quality
+ */
+import { resolveConfig } from "./config.js";
+import { dataQualityDomainSpec, reportKeyOf } from "./store.js";
+import { LocalDataQualityService } from "./provider-local.js";
+import { defineProfileTool } from "./tools/profile.js";
+import { defineCleanTool } from "./tools/clean.js";
+import { defineVerifyTool } from "./tools/verify.js";
+import { VERSION } from "./version.js";
+export const name = 'data-quality';
+/** The three model tools and the durable report domain. */
+export const inject = ['tools', 'storageDomain'];
+export { Config, resolveConfig } from "./config.js";
+export { VERSION } from "./version.js";
+export { DataQualityService, } from "./service.js";
+export { LocalDataQualityService } from "./provider-local.js";
+export { parseLocator, checkCitations } from "./verify.js";
+export { profileTable, renderProfileText } from "./profile.js";
+export { applyCleanRules, serializeDelimited } from "./clean.js";
+export { verifyTable, renderVerifyText } from "./verify.js";
+export { DatasetError, loadTable, loadDocument, parseDelimited, parseJsonTable, resolveWorkspacePath, sampleRows, isMissing, parseNumeric, parseDate, parseBoolean, } from "./dataset.js";
+export { dataQualityDomainSpec, reportKeyOf, reportRecordSchema } from "./store.js";
+export { DATA_QUALITY_EVENT_TYPES, appendDataQualityEvent } from "./events.js";
+export { MAX_CELL_TEXT, truncateCell, truncateRow } from "./present.js";
+/**
+ * Mount the seam: resolve config (fail loud), open the report domain, publish
+ * `ctx.dataQuality`, and register the three tools. With `enabled: false` the
+ * plugin registers nothing and stays inert.
+ * @param ctx - the plugin context (host).
+ * @param config - raw plugin config.
+ */
+export async function apply(ctx, config = {}) {
+    const resolved = resolveConfig(config);
+    const logger = ctx.logger('data-quality');
+    if (!resolved.enabled) {
+        logger.info('disabled: enabled is false — no service or tools are mounted');
+        return;
+    }
+    let store;
+    let domain;
+    if (resolved.storeReports) {
+        domain = await ctx.storageDomain.open(dataQualityDomainSpec);
+        const reports = domain.table('reports');
+        store = {
+            put: async (record) => {
+                const key = reportKeyOf(record);
+                await reports.put(key, record);
+                return key;
+            },
+            get: (key) => reports.get(key),
+        };
+    }
+    const service = new LocalDataQualityService(ctx, resolved, { store, now: Date.now });
+    ctx.tools.register(defineProfileTool(service));
+    ctx.tools.register(defineCleanTool(service));
+    ctx.tools.register(defineVerifyTool(service));
+    logger.info(`dsh-data-quality ${VERSION} mounted: ctx.dataQuality + data_profile/data_clean/data_verify`);
+    if (domain !== undefined) {
+        const handle = domain;
+        ctx.effect(() => async () => {
+            await handle.close();
+        });
+    }
+}
+//# sourceMappingURL=index.js.map

package/lib/types/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAKH,OAAO,EAAU,aAAa,EAAE,MAAM,aAAa,CAAA;AACnD,OAAO,EAAE,qBAAqB,EAAE,WAAW,EAAoB,MAAM,YAAY,CAAA;AACjF,OAAO,EAAE,uBAAuB,EAAE,MAAM,qBAAqB,CAAA;AAC7D,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAA;AACtD,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAClD,OAAO,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAA;AACpD,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AAEtC,MAAM,CAAC,MAAM,IAAI,GAAG,cAAc,CAAA;AAClC,2DAA2D;AAC3D,MAAM,CAAC,MAAM,MAAM,GAAG,CAAC,OAAO,EAAE,eAAe,CAAC,CAAA;AAEhD,OAAO,EAAE,MAAM,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AAEnD,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AACtC,OAAO,EACL,kBAAkB,GAOnB,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,uBAAuB,EAAqB,MAAM,qBAAqB,CAAA;AAIhF,OAAO,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC1D,OAAO,EAAE,YAAY,EAAE,iBAAiB,EAAE,MAAM,cAAc,CAAA;AAC9D,OAAO,EAAE,eAAe,EAAE,kBAAkB,EAAE,MAAM,YAAY,CAAA;AAChE,OAAO,EAAE,WAAW,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAA;AAC3D,OAAO,EACL,YAAY,EACZ,SAAS,EACT,YAAY,EACZ,cAAc,EACd,cAAc,EACd,oBAAoB,EACpB,UAAU,EACV,SAAS,EACT,YAAY,EACZ,SAAS,EACT,YAAY,GAKb,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,qBAAqB,EAAE,WAAW,EAAE,kBAAkB,EAAuC,MAAM,YAAY,CAAA;AACxH,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAwD,MAAM,aAAa,CAAA;AACpI,OAAO,EAAE,aAAa,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,cAAc,CAAA;AAEvE;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,KAAK,CAAC,GAAY,EAAE,SAAiB,EAAE;IAC3D,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,CAAA;IACtC,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,cAAc,CAAC,CAAA;IACzC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC;QACtB,MAAM,CAAC,IAAI,CAAC,8DAA8D,CAAC,CAAA;QAC3E,OAAM;IACR,CAAC;IAED,IAAI,KAA8B,CAAA;IAClC,IAAI,MAAwD,CAAA;IAC5D,IAAI,QAAQ,CAAC,YAAY,EAAE,CAAC;QAC1B,MAAM,GAAG,MAAM,GAAG,CAAC,aAAa,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAA;QAC5D,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CAAA;QACvC,KAAK,GAAG;YACN,GAAG,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE;gBACpB,MAAM,GAAG,GAAG,WAAW,CAAC,MAAM,CAAC,CAAA;gBAC/B,MAAM,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,CAAA;gBAC9B,OAAO,GAAG,CAAA;YACZ,CAAC;YACD,GAAG,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC;SAC/B,CAAA;IACH,CAAC;IAED,MAAM,OAAO,GAAG,IAAI,uBAAuB,CAAC,GAAG,EAAE,QAAQ,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,CAAA;IACpF,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,iBAAiB,CAAC,OAAO,CAAC,CAAC,CAAA;IAC9C,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,eAAe,CAAC,OAAO,CAAC,CAAC,CAAA;IAC5C,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,gBAAgB,CAAC,OAAO,CAAC,CAAC,CAAA;IAC7C,MAAM,CAAC,IAAI,CAAC,oBAAoB,OAAO,iEAAiE,CAAC,CAAA;IAEzG,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,MAAM,CAAA;QACrB,GAAG,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC,KAAK,IAAI,EAAE;YAC1B,MAAM,MAAM,CAAC,KAAK,EAAE,CAAA;QACtB,CAAC,CAAC,CAAA;IACJ,CAAC;AACH,CAAC"}

package/lib/types/present.d.ts

@@ -0,0 +1,24 @@
+/**
+ * Presentation bounds for tool-facing canonical values: cell text is truncated
+ * so one oversized cell cannot flood the model context. This is a display
+ * safety invariant, not a deployment tunable — the durable storage-domain
+ * report always keeps full fidelity.
+ * @module dsh-data-quality/present
+ */
+import type { JsonValue } from '@deepseek-ai/dsh-session';
+/** Maximum characters one cell contributes to a tool-facing row payload. */
+export declare const MAX_CELL_TEXT = 120;
+/**
+ * Truncate one cell for display: long strings are cut with an ellipsis
+ * marker; non-strings pass through.
+ * @param cell - the cell to bound.
+ * @returns the display-safe cell.
+ */
+export declare function truncateCell(cell: JsonValue): JsonValue;
+/**
+ * Truncate every string cell of one row.
+ * @param row - the row to bound.
+ * @returns a display-safe copy.
+ */
+export declare function truncateRow(row: Record<string, JsonValue>): Record<string, JsonValue>;
+//# sourceMappingURL=present.d.ts.map

package/lib/types/present.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"present.d.ts","sourceRoot":"","sources":["../../src/present.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,0BAA0B,CAAA;AAEzD,4EAA4E;AAC5E,eAAO,MAAM,aAAa,MAAM,CAAA;AAEhC;;;;;GAKG;AACH,wBAAgB,YAAY,CAAC,IAAI,EAAE,SAAS,GAAG,SAAS,CAKvD;AAED;;;;GAIG;AACH,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAMrF"}

package/lib/types/present.js

@@ -0,0 +1,34 @@
+/**
+ * Presentation bounds for tool-facing canonical values: cell text is truncated
+ * so one oversized cell cannot flood the model context. This is a display
+ * safety invariant, not a deployment tunable — the durable storage-domain
+ * report always keeps full fidelity.
+ * @module dsh-data-quality/present
+ */
+/** Maximum characters one cell contributes to a tool-facing row payload. */
+export const MAX_CELL_TEXT = 120;
+/**
+ * Truncate one cell for display: long strings are cut with an ellipsis
+ * marker; non-strings pass through.
+ * @param cell - the cell to bound.
+ * @returns the display-safe cell.
+ */
+export function truncateCell(cell) {
+    if (typeof cell === 'string' && cell.length > MAX_CELL_TEXT) {
+        return `${cell.slice(0, MAX_CELL_TEXT)}…`;
+    }
+    return cell;
+}
+/**
+ * Truncate every string cell of one row.
+ * @param row - the row to bound.
+ * @returns a display-safe copy.
+ */
+export function truncateRow(row) {
+    const out = {};
+    for (const [key, value] of Object.entries(row)) {
+        out[key] = truncateCell(value);
+    }
+    return out;
+}
+//# sourceMappingURL=present.js.map

package/lib/types/present.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"present.js","sourceRoot":"","sources":["../../src/present.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,4EAA4E;AAC5E,MAAM,CAAC,MAAM,aAAa,GAAG,GAAG,CAAA;AAEhC;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAAC,IAAe;IAC1C,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,IAAI,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QAC5D,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,aAAa,CAAC,GAAG,CAAA;IAC3C,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,WAAW,CAAC,GAA8B;IACxD,MAAM,GAAG,GAA8B,EAAE,CAAA;IACzC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/C,GAAG,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,KAAK,CAAC,CAAA;IAChC,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}

package/lib/types/profile.d.ts

@@ -0,0 +1,79 @@
+/**
+ * Deterministic dataset profiling: per-column type inference, missingness,
+ * cardinality, numeric distribution, IQR outlier counts, and duplicate-row
+ * detection. Pure functions over a parsed {@link Table} — no clock, no RNG,
+ * no I/O; `generatedAt` is injected by the caller.
+ * @module dsh-data-quality/profile
+ */
+import { type Table } from './dataset.js';
+/** Inferred column type from parsed cell classes. */
+export type InferredType = 'number' | 'date' | 'boolean' | 'string' | 'empty' | 'mixed';
+/** Numeric distribution of one numeric column. */
+export interface NumericProfile {
+    readonly min: number;
+    readonly max: number;
+    readonly mean: number;
+    readonly median: number;
+    readonly p25: number;
+    readonly p75: number;
+    /** Values outside `[p25 - 1.5*IQR, p75 + 1.5*IQR]`. */
+    readonly outliers: number;
+}
+/** One column's profile card. */
+export interface ColumnProfile {
+    readonly name: string;
+    readonly inferredType: InferredType;
+    /** Missing cells over the profiled rows. */
+    readonly missing: number;
+    /** `missing / profiledRows` (0 when the column has no rows). */
+    readonly missingRate: number;
+    /** Distinct non-missing values over the profiled rows. */
+    readonly unique: number;
+    /** Numeric distribution; present only for `number` columns with values. */
+    readonly numeric?: NumericProfile;
+    /** Up to 5 most frequent values for low-cardinality (<= 10 distinct) string/boolean columns. */
+    topValues?: Array<{
+        readonly value: string;
+        readonly count: number;
+    }>;
+    /** Suspicion notes, e.g. mixed-type composition. */
+    notes: string[];
+}
+/** The full profile report (also the persisted and tool-returned value). */
+export interface ProfileReport {
+    readonly dataset: string;
+    readonly rowCount: number;
+    /** Whether column cards describe a systematic sample rather than every row. */
+    readonly sampled: boolean;
+    /** Rows the column cards were computed over. */
+    readonly profiledRows: number;
+    readonly columnCount: number;
+    /** Rows whose full content duplicates an earlier row (over ALL rows). */
+    readonly duplicateRows: number;
+    columns: ColumnProfile[];
+    /** Storage-domain key of the persisted report, when persistence is on (set by the provider). */
+    readonly reportKey?: string;
+    /** Injected generation timestamp (epoch ms). */
+    readonly generatedAt: number;
+}
+/** Compute the numeric distribution of already-parsed values. */
+export declare function numericProfile(values: readonly number[]): NumericProfile | undefined;
+/** Count rows whose full content duplicates an earlier row (first occurrence is not counted). */
+export declare function countDuplicateRows(table: Table, signal?: AbortSignal): number;
+/**
+ * Profile a parsed table. Column cards are computed over a deterministic
+ * systematic sample when `sample` is given; row counts and duplicate counts
+ * always cover the full table.
+ * @param table - the parsed dataset.
+ * @param options - dataset label, optional sample size, injected timestamp, abort signal.
+ * @returns the profile report.
+ */
+export declare function profileTable(table: Table, options: {
+    dataset: string;
+    sample?: number | undefined;
+    generatedAt: number;
+    signal?: AbortSignal | undefined;
+}): ProfileReport;
+/** Human-readable profile summary for the tool's Native render. */
+export declare function renderProfileText(report: ProfileReport): string;
+//# sourceMappingURL=profile.d.ts.map

package/lib/types/profile.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"profile.d.ts","sourceRoot":"","sources":["../../src/profile.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAA0F,KAAK,KAAK,EAAE,MAAM,cAAc,CAAA;AAEjI,qDAAqD;AACrD,MAAM,MAAM,YAAY,GAAG,QAAQ,GAAG,MAAM,GAAG,SAAS,GAAG,QAAQ,GAAG,OAAO,GAAG,OAAO,CAAA;AAEvF,kDAAkD;AAClD,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;IACpB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;IACpB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;IACrB,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;IACvB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;IACpB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;IACpB,uDAAuD;IACvD,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAA;CAC1B;AAED,iCAAiC;AACjC,MAAM,WAAW,aAAa;IAC5B,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;IACrB,QAAQ,CAAC,YAAY,EAAE,YAAY,CAAA;IACnC,4CAA4C;IAC5C,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;IACxB,gEAAgE;IAChE,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;IAC5B,0DAA0D;IAC1D,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;IACvB,2EAA2E;IAC3E,QAAQ,CAAC,OAAO,CAAC,EAAE,cAAc,CAAA;IACjC,gGAAgG;IAChG,SAAS,CAAC,EAAE,KAAK,CAAC;QAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;QAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAA;IACrE,oDAAoD;IACpD,KAAK,EAAE,MAAM,EAAE,CAAA;CAChB;AAED,4EAA4E;AAC5E,MAAM,WAAW,aAAa;IAC5B,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;IACxB,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAA;IACzB,+EAA+E;IAC/E,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAA;IACzB,gDAAgD;IAChD,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAA;IAC7B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;IAC5B,yEAAyE;IACzE,QAAQ,CAAC,aAAa,EAAE,MAAM,CAAA;IAC9B,OAAO,EAAE,aAAa,EAAE,CAAA;IACxB,gGAAgG;IAChG,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAA;IAC3B,gDAAgD;IAChD,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;CAC7B;AAiBD,iEAAiE;AACjE,wBAAgB,cAAc,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,cAAc,GAAG,SAAS,CAmBpF;AAED,iGAAiG;AACjG,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,EAAE,WAAW,GAAG,MAAM,CAa7E;AAsFD;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,KAAK,EAAE,KAAK,EACZ,OAAO,EAAE;IAAE,OAAO,EAAE,MAAM,CAAC;IAAC,MAAM,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC;IAAC,WAAW,EAAE,MAAM,CAAC;IAAC,MAAM,CAAC,EAAE,WAAW,GAAG,SAAS,CAAA;CAAE,GAC/G,aAAa,CAef;AAED,mEAAmE;AACnE,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,aAAa,GAAG,MAAM,CAsB/D"}

package/lib/types/profile.js

@@ -0,0 +1,196 @@
+/**
+ * Deterministic dataset profiling: per-column type inference, missingness,
+ * cardinality, numeric distribution, IQR outlier counts, and duplicate-row
+ * detection. Pure functions over a parsed {@link Table} — no clock, no RNG,
+ * no I/O; `generatedAt` is injected by the caller.
+ * @module dsh-data-quality/profile
+ */
+import { isMissing, parseBoolean, parseDate, parseNumeric, sampleRows, throwIfAborted } from "./dataset.js";
+/** Round to 6 significant digits for stable, readable report numbers. */
+function round6(value) {
+    return Number(value.toPrecision(6));
+}
+/** Linear-interpolation quantile over an ascending-sorted array. */
+function quantile(sorted, q) {
+    const index = (sorted.length - 1) * q;
+    const low = Math.floor(index);
+    const high = Math.ceil(index);
+    const lower = sorted[low];
+    const upper = sorted[high];
+    return lower + (upper - lower) * (index - low);
+}
+/** Compute the numeric distribution of already-parsed values. */
+export function numericProfile(values) {
+    if (values.length === 0)
+        return undefined;
+    const sorted = [...values].sort((a, b) => a - b);
+    const sum = sorted.reduce((acc, value) => acc + value, 0);
+    const p25 = quantile(sorted, 0.25);
+    const p75 = quantile(sorted, 0.75);
+    const iqr = p75 - p25;
+    const lowFence = p25 - 1.5 * iqr;
+    const highFence = p75 + 1.5 * iqr;
+    const outliers = iqr === 0 ? 0 : sorted.filter((value) => value < lowFence || value > highFence).length;
+    return {
+        min: round6(sorted[0]),
+        max: round6(sorted[sorted.length - 1]),
+        mean: round6(sum / sorted.length),
+        median: round6(quantile(sorted, 0.5)),
+        p25: round6(p25),
+        p75: round6(p75),
+        outliers,
+    };
+}
+/** Count rows whose full content duplicates an earlier row (first occurrence is not counted). */
+export function countDuplicateRows(table, signal) {
+    const seen = new Set();
+    let duplicates = 0;
+    for (const [index, row] of table.rows.entries()) {
+        if (index % 1024 === 0)
+            throwIfAborted(signal);
+        const key = JSON.stringify(table.columns.map((column) => row[column] ?? null));
+        if (seen.has(key)) {
+            duplicates += 1;
+        }
+        else {
+            seen.add(key);
+        }
+    }
+    return duplicates;
+}
+/** Profile one column over the given rows. */
+function profileColumn(rows, column, signal) {
+    let missing = 0;
+    let numbers = 0;
+    let dates = 0;
+    let booleans = 0;
+    let strings = 0;
+    const numericValues = [];
+    const distinct = new Set();
+    const frequencies = new Map();
+    for (const [index, row] of rows.entries()) {
+        if (index % 1024 === 0)
+            throwIfAborted(signal);
+        const cell = row[column];
+        if (isMissing(cell)) {
+            missing += 1;
+            continue;
+        }
+        const text = typeof cell === 'string' ? cell : JSON.stringify(cell);
+        distinct.add(text);
+        frequencies.set(text, (frequencies.get(text) ?? 0) + 1);
+        const numeric = parseNumeric(cell);
+        if (numeric !== undefined) {
+            numbers += 1;
+            numericValues.push(numeric);
+            continue;
+        }
+        if (parseDate(cell) !== undefined) {
+            dates += 1;
+            continue;
+        }
+        if (parseBoolean(cell) !== undefined) {
+            booleans += 1;
+            continue;
+        }
+        strings += 1;
+    }
+    const present = rows.length - missing;
+    const notes = [];
+    let inferredType;
+    if (present === 0) {
+        inferredType = 'empty';
+    }
+    else if (numbers === present) {
+        inferredType = 'number';
+    }
+    else if (dates === present) {
+        inferredType = 'date';
+    }
+    else if (booleans === present) {
+        inferredType = 'boolean';
+    }
+    else if (strings === present) {
+        inferredType = 'string';
+    }
+    else {
+        inferredType = 'mixed';
+        const parts = [
+            numbers > 0 ? `${numbers} numeric` : undefined,
+            dates > 0 ? `${dates} date` : undefined,
+            booleans > 0 ? `${booleans} boolean` : undefined,
+            strings > 0 ? `${strings} string` : undefined,
+        ].filter((part) => part !== undefined);
+        notes.push(`mixed types among ${present} present values: ${parts.join(', ')}`);
+    }
+    const profile = {
+        name: column,
+        inferredType,
+        missing,
+        missingRate: rows.length === 0 ? 0 : round6(missing / rows.length),
+        unique: distinct.size,
+        notes,
+    };
+    const numeric = inferredType === 'number' ? numericProfile(numericValues) : undefined;
+    const topValues = (inferredType === 'string' || inferredType === 'boolean') && distinct.size > 0 && distinct.size <= 10
+        ? [...frequencies.entries()]
+            .sort((a, b) => b[1] - a[1] || (a[0] < b[0] ? -1 : a[0] > b[0] ? 1 : 0))
+            .slice(0, 5)
+            .map(([value, count]) => ({ value, count }))
+        : undefined;
+    return {
+        ...profile,
+        ...(numeric !== undefined ? { numeric } : {}),
+        ...(topValues !== undefined ? { topValues } : {}),
+    };
+}
+/**
+ * Profile a parsed table. Column cards are computed over a deterministic
+ * systematic sample when `sample` is given; row counts and duplicate counts
+ * always cover the full table.
+ * @param table - the parsed dataset.
+ * @param options - dataset label, optional sample size, injected timestamp, abort signal.
+ * @returns the profile report.
+ */
+export function profileTable(table, options) {
+    throwIfAborted(options.signal);
+    const profiled = options.sample === undefined ? table.rows : sampleRows(table.rows, options.sample);
+    const columns = table.columns.map((column) => profileColumn(profiled, column, options.signal));
+    const duplicateRows = countDuplicateRows(table, options.signal);
+    return {
+        dataset: options.dataset,
+        rowCount: table.rows.length,
+        sampled: profiled.length !== table.rows.length,
+        profiledRows: profiled.length,
+        columnCount: table.columns.length,
+        duplicateRows,
+        columns,
+        generatedAt: options.generatedAt,
+    };
+}
+/** Human-readable profile summary for the tool's Native render. */
+export function renderProfileText(report) {
+    const lines = [];
+    lines.push(`Profile of ${report.dataset}: ${report.rowCount} rows x ${report.columnCount} columns` +
+        (report.sampled ? ` (column cards over a systematic sample of ${report.profiledRows} rows)` : ''));
+    if (report.duplicateRows > 0)
+        lines.push(`Duplicate rows: ${report.duplicateRows}`);
+    for (const column of report.columns) {
+        const parts = [`${column.name}: ${column.inferredType}`];
+        if (column.missing > 0)
+            parts.push(`missing ${column.missing} (${(column.missingRate * 100).toFixed(1)}%)`);
+        parts.push(`unique ${column.unique}`);
+        if (column.numeric !== undefined) {
+            parts.push(`min ${column.numeric.min}, p25 ${column.numeric.p25}, median ${column.numeric.median}, p75 ${column.numeric.p75}, max ${column.numeric.max}, mean ${column.numeric.mean}` +
+                (column.numeric.outliers > 0 ? `, ${column.numeric.outliers} IQR outliers` : ''));
+        }
+        if (column.topValues !== undefined) {
+            parts.push(`top: ${column.topValues.map((entry) => `${JSON.stringify(entry.value)} x${entry.count}`).join(', ')}`);
+        }
+        for (const note of column.notes)
+            parts.push(`note: ${note}`);
+        lines.push(`- ${parts.join('; ')}`);
+    }
+    return lines.join('\n');
+}
+//# sourceMappingURL=profile.js.map

package/lib/types/profile.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"profile.js","sourceRoot":"","sources":["../../src/profile.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,SAAS,EAAE,YAAY,EAAE,SAAS,EAAE,YAAY,EAAE,UAAU,EAAE,cAAc,EAAwB,MAAM,cAAc,CAAA;AAqDjI,yEAAyE;AACzE,SAAS,MAAM,CAAC,KAAa;IAC3B,OAAO,MAAM,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAA;AACrC,CAAC;AAED,oEAAoE;AACpE,SAAS,QAAQ,CAAC,MAAyB,EAAE,CAAS;IACpD,MAAM,KAAK,GAAG,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;IACrC,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAC7B,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IAC7B,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAW,CAAA;IACnC,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAW,CAAA;IACpC,OAAO,KAAK,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,GAAG,GAAG,CAAC,CAAA;AAChD,CAAC;AAED,iEAAiE;AACjE,MAAM,UAAU,cAAc,CAAC,MAAyB;IACtD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAA;IACzC,MAAM,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAChD,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,GAAG,KAAK,EAAE,CAAC,CAAC,CAAA;IACzD,MAAM,GAAG,GAAG,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;IAClC,MAAM,GAAG,GAAG,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;IAClC,MAAM,GAAG,GAAG,GAAG,GAAG,GAAG,CAAA;IACrB,MAAM,QAAQ,GAAG,GAAG,GAAG,GAAG,GAAG,GAAG,CAAA;IAChC,MAAM,SAAS,GAAG,GAAG,GAAG,GAAG,GAAG,GAAG,CAAA;IACjC,MAAM,QAAQ,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,QAAQ,IAAI,KAAK,GAAG,SAAS,CAAC,CAAC,MAAM,CAAA;IACvG,OAAO;QACL,GAAG,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAW,CAAC;QAChC,GAAG,EAAE,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAW,CAAC;QAChD,IAAI,EAAE,MAAM,CAAC,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC;QACjC,MAAM,EAAE,MAAM,CAAC,QAAQ,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;QACrC,GAAG,EAAE,MAAM,CAAC,GAAG,CAAC;QAChB,GAAG,EAAE,MAAM,CAAC,GAAG,CAAC;QAChB,QAAQ;KACT,CAAA;AACH,CAAC;AAED,iGAAiG;AACjG,MAAM,UAAU,kBAAkB,CAAC,KAAY,EAAE,MAAoB;IACnE,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,IAAI,UAAU,GAAG,CAAC,CAAA;IAClB,KAAK,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QAChD,IAAI,KAAK,GAAG,IAAI,KAAK,CAAC;YAAE,cAAc,CAAC,MAAM,CAAC,CAAA;QAC9C,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,IAAI,CAAC,CAAC,CAAA;QAC9E,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAClB,UAAU,IAAI,CAAC,CAAA;QACjB,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QACf,CAAC;IACH,CAAC;IACD,OAAO,UAAU,CAAA;AACnB,CAAC;AAED,8CAA8C;AAC9C,SAAS,aAAa,CAAC,IAAoB,EAAE,MAAc,EAAE,MAAoB;IAC/E,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,MAAM,aAAa,GAAa,EAAE,CAAA;IAClC,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAA;IAClC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAkB,CAAA;IAC7C,KAAK,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QAC1C,IAAI,KAAK,GAAG,IAAI,KAAK,CAAC;YAAE,cAAc,CAAC,MAAM,CAAC,CAAA;QAC9C,MAAM,IAAI,GAAG,GAAG,CAAC,MAAM,CAAC,CAAA;QACxB,IAAI,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,OAAO,IAAI,CAAC,CAAA;YACZ,SAAQ;QACV,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAA;QACnE,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAClB,WAAW,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;QACvD,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QAClC,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,IAAI,CAAC,CAAA;YACZ,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;YAC3B,SAAQ;QACV,CAAC;QACD,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClC,KAAK,IAAI,CAAC,CAAA;YACV,SAAQ;QACV,CAAC;QACD,IAAI,YAAY,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACrC,QAAQ,IAAI,CAAC,CAAA;YACb,SAAQ;QACV,CAAC;QACD,OAAO,IAAI,CAAC,CAAA;IACd,CAAC;IAED,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,GAAG,OAAO,CAAA;IACrC,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,IAAI,YAA0B,CAAA;IAC9B,IAAI,OAAO,KAAK,CAAC,EAAE,CAAC;QAClB,YAAY,GAAG,OAAO,CAAA;IACxB,CAAC;SAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;QAC/B,YAAY,GAAG,QAAQ,CAAA;IACzB,CAAC;SAAM,IAAI,KAAK,KAAK,OAAO,EAAE,CAAC;QAC7B,YAAY,GAAG,MAAM,CAAA;IACvB,CAAC;SAAM,IAAI,QAAQ,KAAK,OAAO,EAAE,CAAC;QAChC,YAAY,GAAG,SAAS,CAAA;IAC1B,CAAC;SAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;QAC/B,YAAY,GAAG,QAAQ,CAAA;IACzB,CAAC;SAAM,CAAC;QACN,YAAY,GAAG,OAAO,CAAA;QACtB,MAAM,KAAK,GAAG;YACZ,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,UAAU,CAAC,CAAC,CAAC,SAAS;YAC9C,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,OAAO,CAAC,CAAC,CAAC,SAAS;YACvC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,UAAU,CAAC,CAAC,CAAC,SAAS;YAChD,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,SAAS,CAAC,CAAC,CAAC,SAAS;SAC9C,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,CAAC,CAAA;QACtC,KAAK,CAAC,IAAI,CAAC,qBAAqB,OAAO,oBAAoB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAChF,CAAC;IAED,MAAM,OAAO,GAAkB;QAC7B,IAAI,EAAE,MAAM;QACZ,YAAY;QACZ,OAAO;QACP,WAAW,EAAE,IAAI,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC;QAClE,MAAM,EAAE,QAAQ,CAAC,IAAI;QACrB,KAAK;KACN,CAAA;IACD,MAAM,OAAO,GAAG,YAAY,KAAK,QAAQ,CAAC,CAAC,CAAC,cAAc,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,SAAS,CAAA;IACrF,MAAM,SAAS,GACb,CAAC,YAAY,KAAK,QAAQ,IAAI,YAAY,KAAK,SAAS,CAAC,IAAI,QAAQ,CAAC,IAAI,GAAG,CAAC,IAAI,QAAQ,CAAC,IAAI,IAAI,EAAE;QACnG,CAAC,CAAC,CAAC,GAAG,WAAW,CAAC,OAAO,EAAE,CAAC;aACvB,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;aACvE,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;aACX,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC;QAChD,CAAC,CAAC,SAAS,CAAA;IACf,OAAO;QACL,GAAG,OAAO;QACV,GAAG,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7C,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClD,CAAA;AACH,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAY,EACZ,OAAgH;IAEhH,cAAc,CAAC,OAAO,CAAC,MAAM,CAAC,CAAA;IAC9B,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;IACnG,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAA;IAC9F,MAAM,aAAa,GAAG,kBAAkB,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;IAC/D,OAAO;QACL,OAAO,EAAE,OAAO,CAAC,OAAO;QACxB,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM;QAC3B,OAAO,EAAE,QAAQ,CAAC,MAAM,KAAK,KAAK,CAAC,IAAI,CAAC,MAAM;QAC9C,YAAY,EAAE,QAAQ,CAAC,MAAM;QAC7B,WAAW,EAAE,KAAK,CAAC,OAAO,CAAC,MAAM;QACjC,aAAa;QACb,OAAO;QACP,WAAW,EAAE,OAAO,CAAC,WAAW;KACjC,CAAA;AACH,CAAC;AAED,mEAAmE;AACnE,MAAM,UAAU,iBAAiB,CAAC,MAAqB;IACrD,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,KAAK,CAAC,IAAI,CAAC,cAAc,MAAM,CAAC,OAAO,KAAK,MAAM,CAAC,QAAQ,WAAW,MAAM,CAAC,WAAW,UAAU;QAChG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,8CAA8C,MAAM,CAAC,YAAY,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;IACpG,IAAI,MAAM,CAAC,aAAa,GAAG,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,mBAAmB,MAAM,CAAC,aAAa,EAAE,CAAC,CAAA;IACnF,KAAK,MAAM,MAAM,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;QACpC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC,YAAY,EAAE,CAAC,CAAA;QACxD,IAAI,MAAM,CAAC,OAAO,GAAG,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,WAAW,MAAM,CAAC,OAAO,KAAK,CAAC,MAAM,CAAC,WAAW,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAC3G,KAAK,CAAC,IAAI,CAAC,UAAU,MAAM,CAAC,MAAM,EAAE,CAAC,CAAA;QACrC,IAAI,MAAM,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;YACjC,KAAK,CAAC,IAAI,CACR,OAAO,MAAM,CAAC,OAAO,CAAC,GAAG,SAAS,MAAM,CAAC,OAAO,CAAC,GAAG,YAAY,MAAM,CAAC,OAAO,CAAC,MAAM,SAAS,MAAM,CAAC,OAAO,CAAC,GAAG,SAAS,MAAM,CAAC,OAAO,CAAC,GAAG,UAAU,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE;gBACxK,CAAC,MAAM,CAAC,OAAO,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,MAAM,CAAC,OAAO,CAAC,QAAQ,eAAe,CAAC,CAAC,CAAC,EAAE,CAAC,CACnF,CAAA;QACH,CAAC;QACD,IAAI,MAAM,CAAC,SAAS,KAAK,SAAS,EAAE,CAAC;YACnC,KAAK,CAAC,IAAI,CAAC,QAAQ,MAAM,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;QACpH,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,KAAK;YAAE,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;QAC5D,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IACrC,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC"}

package/lib/types/provider-local.d.ts

@@ -0,0 +1,56 @@
+/**
+ * The local deterministic Provider of the `ctx.dataQuality` seam: orchestrates
+ * dataset loading, the pure engines, durable report persistence, and the
+ * adaptive `data-quality/*` session events. All computation is TypeScript in
+ * this process — no model arithmetic, no external processes.
+ * @module dsh-data-quality/provider-local
+ */
+import type { Context } from '@deepseek-ai/cordis';
+import type { ResolvedConfig } from './config.js';
+import { type ReportStore } from './store.js';
+import { DataQualityService, type CitationCheckRequest, type CitationCheckResult, type CleanRequest, type CleanRunReport, type ProfileRequest, type VerifyRequest } from './service.js';
+import type { ProfileReport } from './profile.js';
+import type { VerifyReport } from './verify.js';
+import type { Session } from '@deepseek-ai/dsh-session';
+/** Provider construction dependencies. */
+export interface ProviderDeps {
+    /** Report persistence; absent when `storeReports` is off. */
+    readonly store?: ReportStore | undefined;
+    /** Injected clock — the single time source for reports and freshness defaults. */
+    readonly now: () => number;
+}
+/**
+ * The local deterministic `dataQuality` implementation. Mounted by the plugin
+ * entry; third-party plugins consume it through `ctx.dataQuality`.
+ */
+export declare class LocalDataQualityService extends DataQualityService {
+    private readonly config;
+    private readonly deps;
+    /**
+     * @param ctx - the plugin context.
+     * @param config - the resolved plugin config.
+     * @param deps - store handle plus the injected clock.
+     */
+    constructor(ctx: Context, config: ResolvedConfig, deps: ProviderDeps);
+    /** The absolute root for service-level calls without a session workspace. */
+    serviceRoot(): string;
+    /** Persist one report when persistence is on; returns the storage key. */
+    private persist;
+    /** Emit the adaptive session event when the call carries a session. */
+    emitEvent(session: Session | undefined, kind: 'profile' | 'clean' | 'verify', dataset: string, reportKey: string | undefined, summary: {
+        rows: number;
+        columns?: number;
+        rules?: number;
+        failedRules?: number;
+        passed?: boolean;
+    }): void;
+    /** @inheritdoc DataQualityService.profileDataset */
+    profileDataset(request: ProfileRequest): Promise<ProfileReport>;
+    /** @inheritdoc DataQualityService.cleanDataset */
+    cleanDataset(request: CleanRequest): Promise<CleanRunReport>;
+    /** @inheritdoc DataQualityService.verifyDataset */
+    verifyDataset(request: VerifyRequest): Promise<VerifyReport>;
+    /** @inheritdoc DataQualityService.verifyCitations */
+    verifyCitations(request: CitationCheckRequest): Promise<CitationCheckResult>;
+}
+//# sourceMappingURL=provider-local.d.ts.map

package/lib/types/provider-local.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"provider-local.d.ts","sourceRoot":"","sources":["../../src/provider-local.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAA;AAClD,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAMjD,OAAO,EAAkC,KAAK,WAAW,EAAE,MAAM,YAAY,CAAA;AAE7E,OAAO,EACL,kBAAkB,EAClB,KAAK,oBAAoB,EACzB,KAAK,mBAAmB,EACxB,KAAK,YAAY,EACjB,KAAK,cAAc,EACnB,KAAK,cAAc,EACnB,KAAK,aAAa,EACnB,MAAM,cAAc,CAAA;AACrB,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,cAAc,CAAA;AACjD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC/C,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,0BAA0B,CAAA;AAEvD,0CAA0C;AAC1C,MAAM,WAAW,YAAY;IAC3B,6DAA6D;IAC7D,QAAQ,CAAC,KAAK,CAAC,EAAE,WAAW,GAAG,SAAS,CAAA;IACxC,kFAAkF;IAClF,QAAQ,CAAC,GAAG,EAAE,MAAM,MAAM,CAAA;CAC3B;AAED;;;GAGG;AACH,qBAAa,uBAAwB,SAAQ,kBAAkB;IAC7D,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAgB;IACvC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAc;IAEnC;;;;OAIG;gBACS,GAAG,EAAE,OAAO,EAAE,MAAM,EAAE,cAAc,EAAE,IAAI,EAAE,YAAY;IAMpE,6EAA6E;IAC7E,WAAW,IAAI,MAAM;IAIrB,0EAA0E;YAC5D,OAAO;IASrB,uEAAuE;IACvE,SAAS,CAAC,OAAO,EAAE,OAAO,GAAG,SAAS,EAAE,IAAI,EAAE,SAAS,GAAG,OAAO,GAAG,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,GAAG,SAAS,EAAE,OAAO,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,CAAC,EAAE,MAAM,CAAC;QAAC,KAAK,CAAC,EAAE,MAAM,CAAC;QAAC,WAAW,CAAC,EAAE,MAAM,CAAC;QAAC,MAAM,CAAC,EAAE,OAAO,CAAA;KAAE,GAAG,IAAI;IAUxO,oDAAoD;IACrC,cAAc,CAAC,OAAO,EAAE,cAAc,GAAG,OAAO,CAAC,aAAa,CAAC;IAe9E,kDAAkD;IACnC,YAAY,CAAC,OAAO,EAAE,YAAY,GAAG,OAAO,CAAC,cAAc,CAAC;IAgD3E,mDAAmD;IACpC,aAAa,CAAC,OAAO,EAAE,aAAa,GAAG,OAAO,CAAC,YAAY,CAAC;IAqB3E,qDAAqD;IACtC,eAAe,CAAC,OAAO,EAAE,oBAAoB,GAAG,OAAO,CAAC,mBAAmB,CAAC;CAoB5F"}