chore: ai, public course and other ui improvements

2026-05-09 06:05:03 +01:00
parent 5717dd8004
commit c31ad63955
318 changed files with 20668 additions and 4788 deletions
@@ -0,0 +1,187 @@
+import type {
+  PublicCourseCalloutData,
+  PublicCourseOrgData,
+  PublicCourseSidebarSection,
+  PublicExerciseViewData,
+  PublicLessonViewData
+} from '@cio/ui/custom/public-course';
+import type { ExerciseQuestionModel } from '@cio/question-types';
+
+export const CALLOUT_FIXTURE: PublicCourseCalloutData = {
+  title: 'Want to go deeper on this?',
+  description: 'Join the full "AI for Builders" cohort where we ship a production-grade RAG pipeline in 4 weeks.',
+  buttonLabel: 'Read more on my site',
+  buttonUrl: 'https://example.com/ai-for-builders'
+};
+
+export const ORG_FIXTURE: PublicCourseOrgData = {
+  id: 'org-1',
+  name: 'ClassroomIO',
+  siteName: 'classroomio',
+  avatarUrl: 'https://avatars.githubusercontent.com/u/119528464?s=200&v=4'
+};
+
+export const COURSE_TITLE_FIXTURE = 'AI for Builders — Crash Course';
+
+export const SIDEBAR_FIXTURE: PublicCourseSidebarSection[] = [
+  {
+    id: 'section-intro',
+    title: 'Getting started',
+    items: [
+      {
+        kind: 'lesson',
+        id: 'lesson-welcome',
+        slug: 'welcome',
+        title: 'Welcome to the course',
+        isUnlocked: true
+      },
+      {
+        kind: 'lesson',
+        id: 'lesson-setup',
+        slug: 'setting-up-your-environment',
+        title: 'Setting up your environment',
+        isUnlocked: true
+      }
+    ]
+  },
+  {
+    id: 'section-foundations',
+    title: 'Foundations',
+    items: [
+      {
+        kind: 'lesson',
+        id: 'lesson-intro-llms',
+        slug: 'intro-to-large-language-models',
+        title: 'Intro to large language models',
+        isUnlocked: true
+      },
+      {
+        kind: 'lesson',
+        id: 'lesson-hallucinations',
+        slug: 'hallucination-and-limitations',
+        title: 'Hallucination & limitations',
+        isUnlocked: true
+      },
+      {
+        kind: 'exercise',
+        id: 'exercise-foundations-quiz',
+        slug: 'foundations-quiz',
+        title: 'Foundations quiz',
+        isUnlocked: true
+      }
+    ]
+  },
+  {
+    id: 'section-advanced',
+    title: 'Advanced',
+    items: [
+      {
+        kind: 'lesson',
+        id: 'lesson-rag',
+        slug: 'retrieval-augmented-generation',
+        title: 'Retrieval-augmented generation',
+        isUnlocked: false
+      },
+      {
+        kind: 'lesson',
+        id: 'lesson-agents',
+        slug: 'building-agents',
+        title: 'Building agents',
+        isUnlocked: false
+      }
+    ]
+  }
+];
+
+export const LESSON_FIXTURE: PublicLessonViewData = {
+  kind: 'lesson',
+  title: 'Hallucination & limitations',
+  sectionTitle: 'Foundations',
+  body: `
+    <p>Large language models are incredibly capable, but they also confidently produce incorrect answers — what the field calls <em>hallucination</em>. Before we design systems around them, it helps to build an intuition for where they fail.</p>
+    <h2>When LLMs hallucinate</h2>
+    <ul>
+      <li>When asked about things outside of their training distribution.</li>
+      <li>When prompted with ambiguous or leading questions.</li>
+      <li>When asked to cite sources without retrieval grounding.</li>
+    </ul>
+    <p>In the next lessons we'll cover concrete mitigations, including retrieval, tool use, and structured outputs.</p>
+  `,
+  video: {
+    type: 'youtube',
+    link: 'https://www.youtube.com/watch?v=dQw4w9WgXcQ'
+  },
+  isUnlocked: true
+};
+
+export const LESSON_LOCKED_FIXTURE: PublicLessonViewData = {
+  ...LESSON_FIXTURE,
+  title: 'Building agents',
+  isUnlocked: false,
+  video: null,
+  body: ''
+};
+
+const EXERCISE_QUESTION_FIXTURES: ExerciseQuestionModel[] = [
+  {
+    id: 'q-radio-1',
+    title: 'Which of these is the clearest sign of hallucination?',
+    questionType: 'RADIO',
+    points: 1,
+    options: [
+      { id: 'a', label: 'The model refuses to answer a question it doesn’t know.', isCorrect: false },
+      { id: 'b', label: 'The model confidently invents a fact that doesn’t exist.', isCorrect: true },
+      { id: 'c', label: 'The model cites a source verbatim.', isCorrect: false }
+    ]
+  },
+  {
+    id: 'q-checkbox-1',
+    title: 'Pick every mitigation that grounds an LLM in real data.',
+    questionType: 'CHECKBOX',
+    points: 2,
+    options: [
+      { id: 'a', label: 'Retrieval-augmented generation', isCorrect: true },
+      { id: 'b', label: 'Turning up temperature', isCorrect: false },
+      { id: 'c', label: 'Calling a tool / function', isCorrect: true },
+      { id: 'd', label: 'Asking the model to be more confident', isCorrect: false }
+    ]
+  },
+  {
+    id: 'q-tf-1',
+    title: 'LLMs always know when they’re wrong.',
+    questionType: 'TRUE_FALSE',
+    points: 1,
+    options: [
+      { id: 'true-option', label: 'True', value: 'true', isCorrect: false },
+      { id: 'false-option', label: 'False', value: 'false', isCorrect: true }
+    ]
+  },
+  {
+    id: 'q-short-1',
+    title: 'What’s the shorthand term for "retrieval-augmented generation"?',
+    questionType: 'SHORT_ANSWER',
+    points: 1,
+    options: [
+      { id: 'rag-1', label: 'rag', isCorrect: true },
+      { id: 'rag-2', label: 'RAG', isCorrect: true }
+    ],
+    settings: { caseSensitive: false }
+  },
+  {
+    id: 'q-numeric-1',
+    title: 'How many mitigations did we cover in this lesson? (A number is fine.)',
+    questionType: 'NUMERIC',
+    points: 1,
+    options: [{ id: 'three', label: '3', value: '3', isCorrect: true }],
+    settings: { tolerance: 0 }
+  }
+];
+
+export const EXERCISE_FIXTURE: PublicExerciseViewData = {
+  kind: 'exercise',
+  title: 'Foundations quiz',
+  description: 'A short check-in on the foundations we just covered. Graded instantly in your browser.',
+  sectionTitle: 'Foundations',
+  isUnlocked: true,
+  questions: EXERCISE_QUESTION_FIXTURES
+};