🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
import { Button } from "@/src/components/design-system/Button/Button";
import { fn } from "storybook/test";
import { type ComponentProps } from "react";
import preview from "../../../../.storybook/preview";
import { AnnotationQueueDetails } from "./AnnotationQueueDetails";
const scoreConfigs = [
{
id: "accuracy",
name: "Accuracy",
dataType: "CATEGORICAL",
description: "How closely does the answer match the expected result?",
categories: [
{ label: "Correct", value: 1 },
{ label: "Incorrect", value: 0 },
],
isArchived: false,
},
{
id: "confidence",
name: "Confidence",
dataType: "NUMERIC",
minValue: 0,
maxValue: 1,
isArchived: false,
},
{
id: "feedback",
name: "Feedback",
dataType: "TEXT",
description: "Explain anything the reviewer should follow up on.",
isArchived: false,
},
{
id: "toxicity",
name: "Toxicity",
dataType: "BOOLEAN",
isArchived: false,
},
{
id: "legacy",
name: "Legacy rating",
dataType: "NUMERIC",
minValue: 1,
maxValue: 5,
isArchived: true,
},
] satisfies ComponentProps<typeof AnnotationQueueDetails>["scoreConfigs"];
const meta = preview.meta({
component: AnnotationQueueDetails,
args: {
editAction: (
<Button
text="Edit score configs"
onClick={fn()}
variant="secondary"
size="sm"
/>
),
},
render: (args) => (
<div className="h-[600px] w-80">
<AnnotationQueueDetails {...args} />
</div>
),
});
export const Default = meta.story({
args: {
description: "Review answer quality and leave feedback for the team.",
scoreConfigs,
},
});
export const Empty = meta.story({
args: { description: null, scoreConfigs: [] },
});
export const WithLongText = meta.story({
args: {
description:
"Review answers that require additional context.\nInclude the reason for any score that needs follow-up.",
scoreConfigs: [
{
...scoreConfigs[0],
name: "Answer quality for complex requests",
description:
"Consider whether the answer is complete, supported by the provided context, and clear enough for someone unfamiliar with the conversation.",
},
],
},
});