@aceshooting/lyra-ui 3.9.0 → 4.1.0

package/dist/components/eval-result/eval-result.class.js

@@ -0,0 +1,165 @@
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+import { html, nothing } from 'lit';
+import { property } from 'lit/decorators.js';
+import { LyraElement } from '../../internal/lyra-element.js';
+import { styles } from './eval-result.styles.js';
+import '../rubric-form/rubric-form.class.js';
+import '../data-grid/data-grid.class.js';
+import '../diff-view/diff-view.class.js';
+const EMPTY_RUNS = [];
+const EMPTY_COLUMNS = [];
+const EMPTY_KEYS = [];
+const EMPTY_VALUE = {};
+/**
+ * `<lr-eval-result>` — rubric scoring, human review, and comparison across a single evaluation
+ * example's runs (one per model or prompt version), LangSmith/Arize-eval-result style.
+ *
+ * Composes three existing primitives directly rather than re-deriving any of their behavior:
+ * `<lr-data-grid>` renders the `runs` comparison table (`columns` is a plain pass-through to its
+ * own `DataGridColumn[]` shape, the same way `rubricKeys` is a pass-through to
+ * `<lr-rubric-form>`'s own `keys` -- neither is re-derived here); `<lr-rubric-form>` is the
+ * human-review scoring surface for whichever run is currently selected, reading/writing that
+ * run's own `review` value and re-emitting its `lr-input`/`lr-validity-change`/`lr-submit`/
+ * `lr-skip` events with the run id attached; `<lr-diff-view>` compares the selected run's output
+ * against `baselineRunId`'s output -- `layout="split"` once they resolve to two distinct runs,
+ * `layout="unified"` (an all-equal diff, i.e. a plain read of the one run's output) once they
+ * resolve to the same run or no baseline resolves at all -- so there is no separate un-diffed
+ * "just show the output" code path to keep in sync with the comparison one.
+ *
+ * `selectedRunId`/`baselineRunId` are both fully controlled: this component never mutates either
+ * property itself. Each one falls back to `runs[0]?.id` purely for *rendering* whenever the
+ * property is unset, so the component renders something useful with zero configuration beyond
+ * `runs` -- but moving the selection for real requires the host to set `selectedRunId` in
+ * response to `lr-run-select`, the same shape `<lr-rubric-form>`'s own `itemId` already uses. A
+ * `selectedRunId`/`baselineRunId` that doesn't match any entry in `runs` degrades gracefully: the
+ * comparison grid still renders, and the review/diff sections simply don't (no error, no crash).
+ *
+ * @customElement lr-eval-result
+ * @event lr-run-select - A comparison-grid row was activated. `detail: { runId }`.
+ * @event lr-review-input - The selected run's rubric value changed. `detail: { runId, value }`.
+ * @event lr-review-validity-change - The selected run's rubric validity changed. `detail: { runId, valid, errors }`.
+ * @event lr-review-submit - The selected run's rubric form was submitted. `detail: { runId, value }`.
+ * @event lr-review-skip - The selected run's rubric form was skipped (`reviewSkippable` only). `detail: { runId }`.
+ * @csspart base - The outer wrapper.
+ * @csspart empty - The message shown when `runs` has no entries.
+ * @csspart grid - The `<lr-data-grid>` comparison table.
+ * @csspart review - The `<lr-rubric-form>` scoring the selected run.
+ * @csspart diff - The wrapper around the diff caption and `<lr-diff-view>`.
+ * @csspart diff-labels - The caption row naming the two compared runs (only rendered while comparing two distinct runs).
+ * @csspart diff-label-old - The baseline run's caption.
+ * @csspart diff-label-new - The selected run's caption.
+ * @csspart diff-view - The `<lr-diff-view>` comparing the baseline and selected runs' output.
+ */
+export class LyraEvalResult extends LyraElement {
+    constructor() {
+        super(...arguments);
+        /** The runs (one per model or prompt version) being compared for this evaluation example. */
+        this.runs = EMPTY_RUNS;
+        /** Column definitions for the comparison grid -- forwarded verbatim to `<lr-data-grid>`'s own `columns`. */
+        this.columns = EMPTY_COLUMNS;
+        /** Rubric field definitions for the review form -- forwarded verbatim to `<lr-rubric-form>`'s own `keys`. */
+        this.rubricKeys = EMPTY_KEYS;
+        /** The run currently open for review, and the diff's "new" side. Falls back to `runs[0]?.id` when unset. */
+        this.selectedRunId = '';
+        /** The run compared against, and the diff's "old" side. Falls back to `runs[0]?.id` when unset. */
+        this.baselineRunId = '';
+        /** Shows a Skip control on the review form (forwarded to `<lr-rubric-form>`'s own `skippable`). */
+        this.reviewSkippable = false;
+        /** Disables the review form's controls. The comparison grid stays interactive (selecting a run to inspect is not a mutation). */
+        this.disabled = false;
+    }
+    static { this.styles = [LyraElement.styles, styles]; }
+    get effectiveSelectedRunId() {
+        return this.selectedRunId || this.runs[0]?.id || '';
+    }
+    get selectedRun() {
+        const id = this.effectiveSelectedRunId;
+        return this.runs.find((run) => run.id === id);
+    }
+    get effectiveBaselineRunId() {
+        return this.baselineRunId || this.runs[0]?.id || '';
+    }
+    get baselineRun() {
+        const id = this.effectiveBaselineRunId;
+        return this.runs.find((run) => run.id === id);
+    }
+    renderReview(selected) {
+        return html `<lr-rubric-form
+      part="review"
+      .keys=${this.rubricKeys}
+      .value=${selected.review ?? EMPTY_VALUE}
+      item-id=${selected.id}
+      ?skippable=${this.reviewSkippable}
+      ?disabled=${this.disabled}
+      @lr-input=${(e) => this.emit('lr-review-input', { runId: selected.id, value: e.detail.value })}
+      @lr-validity-change=${(e) => this.emit('lr-review-validity-change', { runId: selected.id, valid: e.detail.valid, errors: e.detail.errors })}
+      @lr-submit=${(e) => this.emit('lr-review-submit', { runId: e.detail.itemId, value: e.detail.value })}
+      @lr-skip=${(e) => this.emit('lr-review-skip', { runId: e.detail.itemId })}
+    ></lr-rubric-form>`;
+    }
+    renderDiff(selected, baseline) {
+        const comparing = Boolean(baseline && baseline.id !== selected.id);
+        return html `<div part="diff">
+      ${comparing
+            ? html `<div part="diff-labels">
+            <span part="diff-label-old">${baseline.label}</span>
+            <span part="diff-label-new">${selected.label}</span>
+          </div>`
+            : nothing}
+      <lr-diff-view
+        part="diff-view"
+        layout=${comparing ? 'split' : 'unified'}
+        .oldText=${baseline?.output ?? selected.output}
+        .newText=${selected.output}
+      ></lr-diff-view>
+    </div>`;
+    }
+    render() {
+        if (this.runs.length === 0) {
+            return html `<div part="base"><p part="empty">${this.localize('noData')}</p></div>`;
+        }
+        const selected = this.selectedRun;
+        return html `
+      <div part="base">
+        <lr-data-grid
+          part="grid"
+          .columns=${this.columns}
+          .rows=${this.runs}
+          .rowKey=${(row) => row.id}
+          .selectedKey=${this.effectiveSelectedRunId}
+          aria-label=${this.getAttribute('aria-label') || nothing}
+          @lr-row-click=${(e) => this.emit('lr-run-select', { runId: e.detail.row.id })}
+        ></lr-data-grid>
+        ${selected ? this.renderReview(selected) : nothing}
+        ${selected ? this.renderDiff(selected, this.baselineRun) : nothing}
+      </div>
+    `;
+    }
+}
+__decorate([
+    property({ attribute: false })
+], LyraEvalResult.prototype, "runs", void 0);
+__decorate([
+    property({ attribute: false })
+], LyraEvalResult.prototype, "columns", void 0);
+__decorate([
+    property({ attribute: false })
+], LyraEvalResult.prototype, "rubricKeys", void 0);
+__decorate([
+    property({ attribute: 'selected-run-id' })
+], LyraEvalResult.prototype, "selectedRunId", void 0);
+__decorate([
+    property({ attribute: 'baseline-run-id' })
+], LyraEvalResult.prototype, "baselineRunId", void 0);
+__decorate([
+    property({ type: Boolean, attribute: 'review-skippable' })
+], LyraEvalResult.prototype, "reviewSkippable", void 0);
+__decorate([
+    property({ type: Boolean, reflect: true })
+], LyraEvalResult.prototype, "disabled", void 0);
+//# sourceMappingURL=eval-result.class.js.map

package/dist/components/eval-result/eval-result.class.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-result.class.js","sourceRoot":"","sources":["../../../src/components/eval-result/eval-result.class.ts"],"names":[],"mappings":";;;;;;AAAA,OAAO,EAAE,IAAI,EAAE,OAAO,EAAuB,MAAM,KAAK,CAAC;AACzD,OAAO,EAAE,QAAQ,EAAE,MAAM,mBAAmB,CAAC;AAC7C,OAAO,EAAE,WAAW,EAAE,MAAM,gCAAgC,CAAC;AAC7D,OAAO,EAAE,MAAM,EAAE,MAAM,yBAAyB,CAAC;AAGjD,OAAO,qCAAqC,CAAC;AAC7C,OAAO,iCAAiC,CAAC;AACzC,OAAO,iCAAiC,CAAC;AAEzC,MAAM,UAAU,GAAoB,EAAE,CAAC;AACvC,MAAM,aAAa,GAAoC,EAAE,CAAC;AAC1D,MAAM,UAAU,GAAgB,EAAE,CAAC;AACnC,MAAM,WAAW,GAAgB,EAAE,CAAC;AA4BpC;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AACH,MAAM,OAAO,cAAe,SAAQ,WAAmC;IAAvE;;QAGE,6FAA6F;QAC7D,SAAI,GAAoB,UAAU,CAAC;QAEnE,4GAA4G;QAC5E,YAAO,GAAoC,aAAa,CAAC;QAEzF,6GAA6G;QAC7E,eAAU,GAAgB,UAAU,CAAC;QAErE,4GAA4G;QAChE,kBAAa,GAAG,EAAE,CAAC;QAE/D,mGAAmG;QACvD,kBAAa,GAAG,EAAE,CAAC;QAE/D,mGAAmG;QACvC,oBAAe,GAAG,KAAK,CAAC;QAEpF,iIAAiI;QACrF,aAAQ,GAAG,KAAK,CAAC;IA6E/D,CAAC;aAlGQ,WAAM,GAAG,CAAC,WAAW,CAAC,MAAM,EAAE,MAAM,CAAC,AAA/B,CAAgC;IAuB7C,IAAY,sBAAsB;QAChC,OAAO,IAAI,CAAC,aAAa,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC;IACtD,CAAC;IAED,IAAY,WAAW;QACrB,MAAM,EAAE,GAAG,IAAI,CAAC,sBAAsB,CAAC;QACvC,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC;IAChD,CAAC;IAED,IAAY,sBAAsB;QAChC,OAAO,IAAI,CAAC,aAAa,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC;IACtD,CAAC;IAED,IAAY,WAAW;QACrB,MAAM,EAAE,GAAG,IAAI,CAAC,sBAAsB,CAAC;QACvC,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC;IAChD,CAAC;IAEO,YAAY,CAAC,QAAuB;QAC1C,OAAO,IAAI,CAAA;;cAED,IAAI,CAAC,UAAU;eACd,QAAQ,CAAC,MAAM,IAAI,WAAW;gBAC7B,QAAQ,CAAC,EAAE;mBACR,IAAI,CAAC,eAAe;kBACrB,IAAI,CAAC,QAAQ;kBACb,CAAC,CAAsC,EAAE,EAAE,CACrD,IAAI,CAAC,IAAI,CAAC,iBAAiB,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;4BACvD,CAAC,CAAkE,EAAE,EAAE,CAC3F,IAAI,CAAC,IAAI,CAAC,2BAA2B,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;mBACnG,CAAC,CAAsD,EAAE,EAAE,CACtE,IAAI,CAAC,IAAI,CAAC,kBAAkB,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;iBACvE,CAAC,CAAkC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;uBACzF,CAAC;IACtB,CAAC;IAEO,UAAU,CAAC,QAAuB,EAAE,QAAmC;QAC7E,MAAM,SAAS,GAAG,OAAO,CAAC,QAAQ,IAAI,QAAQ,CAAC,EAAE,KAAK,QAAQ,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,IAAI,CAAA;QACP,SAAS;YACT,CAAC,CAAC,IAAI,CAAA;0CAC4B,QAAS,CAAC,KAAK;0CACf,QAAQ,CAAC,KAAK;iBACvC;YACT,CAAC,CAAC,OAAO;;;iBAGA,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS;mBAC7B,QAAQ,EAAE,MAAM,IAAI,QAAQ,CAAC,MAAM;mBACnC,QAAQ,CAAC,MAAM;;WAEvB,CAAC;IACV,CAAC;IAED,MAAM;QACJ,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC3B,OAAO,IAAI,CAAA,oCAAoC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC;QACrF,CAAC;QACD,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,CAAC;QAClC,OAAO,IAAI,CAAA;;;;qBAIM,IAAI,CAAC,OAAO;kBACf,IAAI,CAAC,IAAI;oBACP,CAAC,GAAkB,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE;yBACzB,IAAI,CAAC,sBAAsB;uBAC7B,IAAI,CAAC,YAAY,CAAC,YAAY,CAAC,IAAI,OAAO;0BACvC,CAAC,CAAsC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;;UAElH,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,OAAO;UAChD,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,OAAO;;KAErE,CAAC;IACJ,CAAC;CACF;AA/FiC;IAA/B,QAAQ,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;4CAAoC;AAGnC;IAA/B,QAAQ,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;+CAA0D;AAGzD;IAA/B,QAAQ,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;kDAAsC;AAGzB;IAA3C,QAAQ,CAAC,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAC;qDAAoB;AAGnB;IAA3C,QAAQ,CAAC,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAC;qDAAoB;AAGH;IAA3D,QAAQ,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,SAAS,EAAE,kBAAkB,EAAE,CAAC;uDAAyB;AAGxC;IAA3C,QAAQ,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;gDAAkB"}

package/dist/components/eval-result/eval-result.d.ts

@@ -0,0 +1,5 @@
+export * from './eval-result.class.js';
+import '../rubric-form/rubric-form.js';
+import '../data-grid/data-grid.js';
+import '../diff-view/diff-view.js';
+//# sourceMappingURL=eval-result.d.ts.map

package/dist/components/eval-result/eval-result.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-result.d.ts","sourceRoot":"","sources":["../../../src/components/eval-result/eval-result.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAC;AAGvC,OAAO,+BAA+B,CAAC;AACvC,OAAO,2BAA2B,CAAC;AACnC,OAAO,2BAA2B,CAAC"}

package/dist/components/eval-result/eval-result.js

@@ -0,0 +1,8 @@
+export * from './eval-result.class.js';
+import { LyraEvalResult } from './eval-result.class.js';
+import { defineElement } from '../../internal/prefix.js';
+import '../rubric-form/rubric-form.js';
+import '../data-grid/data-grid.js';
+import '../diff-view/diff-view.js';
+defineElement('eval-result', LyraEvalResult);
+//# sourceMappingURL=eval-result.js.map

package/dist/components/eval-result/eval-result.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-result.js","sourceRoot":"","sources":["../../../src/components/eval-result/eval-result.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAC;AACvC,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,aAAa,EAAE,MAAM,0BAA0B,CAAC;AACzD,OAAO,+BAA+B,CAAC;AACvC,OAAO,2BAA2B,CAAC;AACnC,OAAO,2BAA2B,CAAC;AAEnC,aAAa,CAAC,aAAa,EAAE,cAAc,CAAC,CAAC"}

package/dist/components/eval-result/eval-result.styles.d.ts

@@ -0,0 +1,2 @@
+export declare const styles: import("lit").CSSResult;
+//# sourceMappingURL=eval-result.styles.d.ts.map

package/dist/components/eval-result/eval-result.styles.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-result.styles.d.ts","sourceRoot":"","sources":["../../../src/components/eval-result/eval-result.styles.ts"],"names":[],"mappings":"AAEA,eAAO,MAAM,MAAM,yBA4ClB,CAAC"}

package/dist/components/eval-result/eval-result.styles.js

@@ -0,0 +1,47 @@
+import { css } from 'lit';
+export const styles = css `
+  :host {
+    display: block;
+    container-type: inline-size;
+  }
+
+  [part='base'] {
+    display: flex;
+    flex-direction: column;
+    gap: var(--lr-space-l);
+    min-inline-size: 0;
+    color: var(--lr-color-text);
+  }
+
+  [part='empty'] {
+    margin: 0;
+    color: var(--lr-color-text-quiet);
+    font-size: var(--lr-font-size-sm);
+  }
+
+  [part='grid'],
+  [part='review'],
+  [part='diff'],
+  [part='diff-view'] {
+    min-inline-size: 0;
+  }
+
+  [part='diff'] {
+    display: flex;
+    flex-direction: column;
+    gap: var(--lr-space-2xs);
+  }
+
+  [part='diff-labels'] {
+    display: flex;
+    justify-content: space-between;
+    flex-wrap: wrap;
+    gap: var(--lr-space-s);
+    font-weight: var(--lr-font-weight-semibold);
+    font-size: var(--lr-font-size-xs);
+    color: var(--lr-color-text-quiet);
+    text-transform: uppercase;
+    letter-spacing: var(--lr-size-0-02em);
+  }
+`;
+//# sourceMappingURL=eval-result.styles.js.map

package/dist/components/eval-result/eval-result.styles.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-result.styles.js","sourceRoot":"","sources":["../../../src/components/eval-result/eval-result.styles.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,GAAG,EAAE,MAAM,KAAK,CAAC;AAE1B,MAAM,CAAC,MAAM,MAAM,GAAG,GAAG,CAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA4CxB,CAAC"}

package/dist/components/evaluation-run/evaluation-run.class.d.ts

@@ -0,0 +1,168 @@
+import { type TemplateResult, type PropertyValues } from 'lit';
+import { LyraElement } from '../../internal/lyra-element.js';
+import type { AgentStatus, Citation, GroundingAssessment } from '../../ai/types.js';
+import type { ToolTimelineEntry, ToolTimelineApprovalDetail } from '../tool-timeline/tool-timeline.class.js';
+import '../live-region/live-region.js';
+import '../progress/progress-bar.js';
+import '../empty/empty.js';
+import '../details/details.js';
+import '../badge/badge.js';
+import '../markdown/markdown.js';
+import '../code-block/code-block.js';
+import '../grounding-summary/grounding-summary.js';
+import '../tool-timeline/tool-timeline.js';
+/** How an example's `input`/`output` text is rendered -- `'markdown'` (the default) through
+ *  `<lr-markdown>`, `'code'` through `<lr-code-block>` (consulting the matching `*Language`
+ *  property for shiki highlighting). */
+export type EvaluationContentFormat = 'markdown' | 'code';
+/**
+ * One example's result within an evaluation batch. `status` reuses the shared `AgentStatus`
+ * contract from `src/ai/types.ts` -- the same run-lifecycle vocabulary an agent step already
+ * uses -- rather than inventing a parallel pass/fail enum; rubric-driven scoring (did the output
+ * actually pass) is a separate concern owned by the sibling result-review component, not this
+ * one's batch-progress/trace-display job. `grounding`/`citations` compose directly into
+ * `<lr-grounding-summary>` (`assessment`/`citations`) and `toolTrace` directly into
+ * `<lr-tool-timeline>` (`entries`) with no adapters -- this component owns no grounding-scoring or
+ * tool-call rendering logic of its own.
+ */
+export interface EvaluationExampleResult {
+    id: string;
+    /** Falls back to a localized "Example {index}" (1-based, in array order) when unset. */
+    label?: string;
+    status: AgentStatus;
+    input: string;
+    /** `'markdown'` (the default, when unset) renders via `<lr-markdown>`. */
+    inputFormat?: EvaluationContentFormat;
+    /** A shiki-recognized language id, consulted only when `inputFormat` is `'code'`. */
+    inputLanguage?: string;
+    output: string;
+    outputFormat?: EvaluationContentFormat;
+    outputLanguage?: string;
+    /** This example's grounding/citation-support assessment, when the run computed one. Omitted
+     *  entirely means no grounding section renders for this example. */
+    grounding?: GroundingAssessment;
+    /** Evidence citations backing `grounding`, forwarded verbatim to `<lr-grounding-summary>`'s own
+     *  `citations`. Only consulted while `grounding` is also set. */
+    citations?: Citation[];
+    /** This example's own tool-call trace. Omitted or empty means no tool-trace section renders for
+     *  this example. */
+    toolTrace?: ToolTimelineEntry[];
+}
+/** `detail` for `lr-example-toggle`. */
+export interface EvaluationExampleToggleDetail {
+    id: string;
+    expanded: boolean;
+}
+/** `detail` for `lr-example-citation-select` -- the nested per-example `<lr-grounding-summary>`'s
+ *  own `lr-citation-select` detail (`{ citation }`), correlated with the example it came from so a
+ *  host doesn't need to walk the DOM to find out which example's evidence was activated. */
+export interface EvaluationCitationSelectDetail {
+    exampleId: string;
+    citation: Citation;
+}
+/** `detail` for `lr-example-tool-approval-decide` -- the nested per-example `<lr-tool-timeline>`'s
+ *  own `lr-tool-approval-decide` detail (`ToolTimelineApprovalDetail`), correlated with the
+ *  example it came from. */
+export interface EvaluationToolApprovalDetail extends ToolTimelineApprovalDetail {
+    exampleId: string;
+}
+export interface LyraEvaluationRunEventMap {
+    'lr-example-toggle': CustomEvent<EvaluationExampleToggleDetail>;
+    'lr-example-citation-select': CustomEvent<EvaluationCitationSelectDetail>;
+    'lr-example-tool-approval-decide': CustomEvent<EvaluationToolApprovalDetail>;
+}
+/**
+ * `<lr-evaluation-run>` — an evaluation batch's live progress: an overall `<lr-progress-bar>`
+ * counting terminal (done/error/cancelled) examples against the batch total, plus one
+ * `<lr-details>` disclosure per example showing its input/output (`<lr-markdown>` or
+ * `<lr-code-block>`, per `inputFormat`/`outputFormat`), a `<lr-grounding-summary>` when the
+ * example carries a `GroundingAssessment`, and a `<lr-tool-timeline>` when it carries
+ * `toolTrace` entries. Controlled: `examples` mirrors this package's other data-driven
+ * components' own convention (a plain prop the host replaces wholesale to update, never mutated
+ * in place).
+ *
+ * Nested-component events that need per-example correlation are intercepted and re-emitted under
+ * this component's own name with the originating example's `id` folded into `detail` (matching
+ * `<lr-tool-timeline>`'s own precedent of extending a shared `*EventDetail` type from
+ * `src/ai/types.ts` rather than inventing a divergent shape) -- a host listening at this
+ * component's boundary never needs to walk the DOM to find out which example a nested selection
+ * or approval decision came from.
+ *
+ * @customElement lr-evaluation-run
+ * @event lr-example-toggle - An example's disclosure was expanded or collapsed. `detail: { id,
+ *   expanded }`.
+ * @event lr-example-citation-select - An evidence citation in a nested `<lr-grounding-summary>`
+ *   was activated. `detail: { exampleId, citation }`.
+ * @event lr-example-tool-approval-decide - A pending tool call in a nested `<lr-tool-timeline>`
+ *   was approved or denied. `detail: { exampleId, invocationId, approved, args? }`.
+ * @csspart base - The root wrapper.
+ * @csspart header - The batch-progress header row.
+ * @csspart header-label - The run's label, defaulting to a localized "Evaluation run".
+ * @csspart progress - The batch `<lr-progress-bar>`.
+ * @csspart summary - The "N of M examples complete" text.
+ * @csspart counts - Wrapper around the running/failed count badges.
+ * @csspart count - One count badge; carries `data-kind="running"|"error"`.
+ * @csspart examples - Wrapper around the per-example `<lr-details>` rows.
+ * @csspart example - One example's `<lr-details>` row; carries `data-status` (the example's
+ *   `status.kind`).
+ * @csspart example-summary - The wrapper around an example's label and status badge, in the
+ *   `<lr-details>` `summary` slot.
+ * @csspart example-label - An example's label text.
+ * @csspart example-status - An example's status badge.
+ * @csspart input-section - Wrapper around an example's input heading + rendered content.
+ * @csspart output-section - Wrapper around an example's output heading + rendered content.
+ * @csspart grounding-section - Wrapper around an example's `<lr-grounding-summary>`, only
+ *   rendered when the example carries a `grounding` assessment.
+ * @csspart tool-trace-section - Wrapper around an example's `<lr-tool-timeline>`, only rendered
+ *   when the example carries non-empty `toolTrace` entries.
+ * @csspart section-heading - The heading text inside any of the four sections above.
+ * @csspart input - The rendered `<lr-markdown>`/`<lr-code-block>` for an example's input.
+ * @csspart output - The rendered `<lr-markdown>`/`<lr-code-block>` for an example's output.
+ * @csspart grounding-summary - The nested `<lr-grounding-summary>` for an example's grounding
+ *   assessment.
+ * @csspart tool-trace - The nested `<lr-tool-timeline>` for an example's tool calls.
+ * @csspart empty - The empty-state message shown when `examples` is empty.
+ * @csspart live-region - The internal status-announcement live region.
+ */
+export declare class LyraEvaluationRun extends LyraElement<LyraEvaluationRunEventMap> {
+    static styles: import("lit").CSSResultGroup[];
+    /** The batch's examples so far. Controlled -- never mutated by this component; pass a new array
+     *  to update it (e.g. as each example finishes, or as the whole batch streams in). */
+    examples: EvaluationExampleResult[];
+    /** The batch's expected total example count. `null` (the default) derives it from
+     *  `examples.length` instead -- the common case once every result has already arrived; set this
+     *  explicitly while a batch is still streaming in and the eventual total is already known ahead
+     *  of every example actually completing. */
+    total: number | null;
+    /** Header label and accessible-name source. Falls back to a localized "Evaluation run" when
+     *  unset. */
+    label: string;
+    private expandedIds;
+    private liveRegion?;
+    /** `true` until the first completed update -- gates the status-change announcements below so a
+     *  freshly-mounted run never announces whatever statuses its very first `examples` happens to
+     *  carry (mirrors `<lr-task-list>`'s identical `isMounting` gate). */
+    private isMounting;
+    /** Last-seen `status.kind` per example id, diffed against the incoming `examples` on every
+     *  update to decide what to announce. */
+    private previousStatusById;
+    protected updated(changed: PropertyValues): void;
+    private statusCounts;
+    private exampleLabel;
+    private statusLabel;
+    private diffAndAnnounce;
+    private onExampleToggle;
+    private onCitationSelect;
+    private onToolApprovalDecide;
+    private renderContent;
+    private renderGrounding;
+    private renderToolTrace;
+    private renderExample;
+    render(): TemplateResult;
+}
+declare global {
+    interface HTMLElementTagNameMap {
+        'lr-evaluation-run': LyraEvaluationRun;
+    }
+}
+//# sourceMappingURL=evaluation-run.class.d.ts.map

package/dist/components/evaluation-run/evaluation-run.class.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"evaluation-run.class.d.ts","sourceRoot":"","sources":["../../../src/components/evaluation-run/evaluation-run.class.ts"],"names":[],"mappings":"AAAA,OAAO,EAAiB,KAAK,cAAc,EAAE,KAAK,cAAc,EAAE,MAAM,KAAK,CAAC;AAE9E,OAAO,EAAE,WAAW,EAAE,MAAM,gCAAgC,CAAC;AAE7D,OAAO,KAAK,EAAE,WAAW,EAAmB,QAAQ,EAAE,mBAAmB,EAAE,MAAM,mBAAmB,CAAC;AAKrG,OAAO,KAAK,EAAE,iBAAiB,EAAE,0BAA0B,EAA4B,MAAM,yCAAyC,CAAC;AACvI,OAAO,+BAA+B,CAAC;AACvC,OAAO,6BAA6B,CAAC;AACrC,OAAO,mBAAmB,CAAC;AAC3B,OAAO,uBAAuB,CAAC;AAC/B,OAAO,mBAAmB,CAAC;AAC3B,OAAO,yBAAyB,CAAC;AACjC,OAAO,6BAA6B,CAAC;AACrC,OAAO,2CAA2C,CAAC;AACnD,OAAO,mCAAmC,CAAC;AAG3C;;wCAEwC;AACxC,MAAM,MAAM,uBAAuB,GAAG,UAAU,GAAG,MAAM,CAAC;AAE1D;;;;;;;;;GASG;AACH,MAAM,WAAW,uBAAuB;IACtC,EAAE,EAAE,MAAM,CAAC;IACX,wFAAwF;IACxF,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,WAAW,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,0EAA0E;IAC1E,WAAW,CAAC,EAAE,uBAAuB,CAAC;IACtC,qFAAqF;IACrF,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,uBAAuB,CAAC;IACvC,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB;wEACoE;IACpE,SAAS,CAAC,EAAE,mBAAmB,CAAC;IAChC;qEACiE;IACjE,SAAS,CAAC,EAAE,QAAQ,EAAE,CAAC;IACvB;wBACoB;IACpB,SAAS,CAAC,EAAE,iBAAiB,EAAE,CAAC;CACjC;AAED,wCAAwC;AACxC,MAAM,WAAW,6BAA6B;IAC5C,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,OAAO,CAAC;CACnB;AAED;;4FAE4F;AAC5F,MAAM,WAAW,8BAA8B;IAC7C,SAAS,EAAE,MAAM,CAAC;IAClB,QAAQ,EAAE,QAAQ,CAAC;CACpB;AAED;;4BAE4B;AAC5B,MAAM,WAAW,4BAA6B,SAAQ,0BAA0B;IAC9E,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,yBAAyB;IACxC,mBAAmB,EAAE,WAAW,CAAC,6BAA6B,CAAC,CAAC;IAChE,4BAA4B,EAAE,WAAW,CAAC,8BAA8B,CAAC,CAAC;IAC1E,iCAAiC,EAAE,WAAW,CAAC,4BAA4B,CAAC,CAAC;CAC9E;AAyBD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAoDG;AACH,qBAAa,iBAAkB,SAAQ,WAAW,CAAC,yBAAyB,CAAC;IAC3E,MAAM,CAAC,MAAM,iCAAgC;IAE7C;0FACsF;IACtD,QAAQ,EAAE,uBAAuB,EAAE,CAAM;IAEzE;;;gDAG4C;IAChB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAQ;IAExD;iBACa;IACD,KAAK,SAAM;IAEd,OAAO,CAAC,WAAW,CAAqB;IAExB,OAAO,CAAC,UAAU,CAAC,CAAiB;IAE7D;;0EAEsE;IACtE,OAAO,CAAC,UAAU,CAAQ;IAE1B;6CACyC;IACzC,OAAO,CAAC,kBAAkB,CAAsC;IAEhE,SAAS,CAAC,OAAO,CAAC,OAAO,EAAE,cAAc,GAAG,IAAI,CAI/C;IAED,OAAO,CAAC,YAAY;IAMpB,OAAO,CAAC,YAAY;IAIpB,OAAO,CAAC,WAAW;IAmBnB,OAAO,CAAC,eAAe;IA4CvB,OAAO,CAAC,eAAe;IAUvB,OAAO,CAAC,gBAAgB;IAKxB,OAAO,CAAC,oBAAoB;IAK5B,OAAO,CAAC,aAAa;IAYrB,OAAO,CAAC,eAAe;IAevB,OAAO,CAAC,eAAe;IAcvB,OAAO,CAAC,aAAa;IA2BrB,MAAM,IAAI,cAAc,CA0CvB;CACF;AAED,OAAO,CAAC,MAAM,CAAC;IACb,UAAU,qBAAqB;QAC7B,mBAAmB,EAAE,iBAAiB,CAAC;KACxC;CACF"}