Implement the following plan:

Plan: TextSubmissionBoard — AI Discussion Question Generator

Context

During share-out stages, the instructor has difficulty drawing students into conversation. The idea: after students submit text responses via <TextSubmission>, the instructor presses a button that uses AI to analyze all submissions and generate targeted discussion questions. Some questions address individual students ("Ask Alice: Why did you say X?"), others surface themes across submissions. Each question is rated by interestingness and lists relevant students.

This is instructor-device-only — the projector display shows nothing for this component.

Follows the existing StrategyEntry → StrategyBoard pattern exactly.

Files to Create

1. packages/db/src/schema/text-aggregations.ts — New DB table

Follow strategyAggregations pattern from strategy-submissions.ts:67-92.

export interface DiscussionQuestion {
  question: string;
  interestingness: number; // 1-5
  type: "individual" | "thematic";
  relevantStudents: Array<{ name: string; reason: string }>;
}

export interface ProcessedDiscussionQuestions {
  questions: DiscussionQuestion[];
  submissionCount: number;
}

Table textAggregations: id, submissionId, semester, meetingSlug, rawSubmissions (jsonb — Array<{ content, userName }>), processedQuestions (jsonb — ProcessedDiscussionQuestions), processedAt, createdAt. Unique on (submissionId, semester).

2. apps/web/components/mdx/text-submission-board.tsx — Server component

Same structure as strategy-board.tsx: check role, render login prompt for guests, pass isInstructor to client component. Students see "Discussion in progress..."

3. apps/web/components/mdx/text-submission-board-client.tsx — Client component

Follow strategy-board-client.tsx pattern:

• SWR fetch from admin text-submissions endpoint (which will now return aggregation)
• Socket listeners: textSubmission:submitted (update count) + textAggregation:processed (refresh results)
• "Generate Discussion Questions" button → POST to /api/admin/text-submissions/:id/process
• Shows results sorted by interestingness descending
• Individual questions: blue-bordered, thematic: purple-bordered
• Each question shows: type badge, star rating, question text, student name pills (hover for reason)
• "Regenerate" button replaces "Generate" after first run

Files to Modify

4. packages/db/src/schema/index.ts

Add: export * from "./text-aggregations";

5. apps/api/src/socket.ts + apps/web/lib/socket.ts

Add "textAggregation:processed" event to ServerToClientEvents in both files. Add emitTextAggregationProcessed() emitter function in API socket.ts.

6. apps/api/src/routes/admin/text-submissions.ts

• Add imports: textAggregations, ProcessedDiscussionQuestions, nanoid, chatJSON, emitTextAggregationProcessed
• Modify GET endpoint: Add aggregation lookup (query textAggregations), include in response as aggregation: aggregation ?? null
• Add POST /:submissionId/process: Collect all latest text submissions with user names, call chatJSON<ProcessedDiscussionQuestions>(), upsert textAggregations, emit socket event

7. apps/web/mdx-components.tsx

Import TextSubmissionBoard, add to components object (after StrategyBoard line 189).

8. apps/web/components/display/display-mdx-provider.tsx

Add TextSubmissionBoard: Null to the null-components section.

LLM Prompt (for the process endpoint)

The system prompt instructs the AI to:

• Generate 5-10 discussion questions from student text submissions
• Individual questions target a specific student ("Ask [Name]: ...") — draws out deeper thinking
• Thematic questions surface patterns, tensions, or disagreements across submissions
• Rate each question's interestingness 1-5
• List relevant students with brief reasons
• Temperature: 0.4 (slightly creative but reliable for JSON)

Database Migration

After creating the schema file:

1. bun run db:generate
2. bun run db:migrate

Verification

1. Run bun run tsc to verify no type errors
2. Start dev server, navigate to a meeting page with <TextSubmissionBoard id="...">
3. As instructor: confirm submission count appears and updates in real-time
4. With submissions present: click "Generate Discussion Questions"
5. Verify questions appear sorted by interestingness with type badges, stars, and student name pills
6. Click "Regenerate" to confirm reprocessing works
7. As student: confirm "Discussion in progress..." message appears
8. On display page: confirm nothing renders for TextSubmissionBoard

If you need specific details from before exiting plan mode (like exact code snippets, error messages, or content you generated), read the full transcript at: ~/.claude/projects/-Users-challen-www-usingandunderstanding-ai/06f0330d-5995-47ee-97de-60bbd6c63677.jsonl

If this plan can be broken down into multiple independent tasks, consider using the TeamCreate tool to create a team and parallelize the work.

Your task

First, run git status, git diff, and git log --oneline -10 to understand the current state.

Then add and commit all modified files with a descriptive message. Break the changes into multiple commits if appropriate. If there are untracked files, ask if they should be added or added to the appropriate .gitignore file. The user may also want to take over at this point, so provide an abort option as well. Always check changes for unnecessary debugging statements or logging that might have been accidentally included. Prompt the user if you find any, and offer to remove each one by one. Do not use interactive git staging (git add -p) as it does not work.

Great! Do we need new tests for this functionality?

Didn't we write tests using mocking for the other component that uses LLM summarization? Can we use a similar approach? But yes, let's right that e2e test now, and then we'll write more when we create Thursday's activity.

Your task

First, run git status, git diff, and git log --oneline -10 to understand the current state.

Then add and commit all modified files with a descriptive message. Break the changes into multiple commits if appropriate. If there are untracked files, ask if they should be added or added to the appropriate .gitignore file. The user may also want to take over at this point, so provide an abort option as well. Always check changes for unnecessary debugging statements or logging that might have been accidentally included. Prompt the user if you find any, and offer to remove each one by one. Do not use interactive git staging (git add -p) as it does not work.

Your task

First, run git status, git diff, and git log --oneline -10 to understand the current state.

Then add and commit all modified files with a descriptive message. Break the changes into multiple commits if appropriate. If there are untracked files, ask if they should be added or added to the appropriate .gitignore file. The user may also want to take over at this point, so provide an abort option as well. Always check changes for unnecessary debugging statements or logging that might have been accidentally included. Prompt the user if you find any, and offer to remove each one by one. Do not use interactive git staging (git add -p) as it does not work.

Read the output file to retrieve the result: [tmp-path]

How do I open the display view for the activity?

Can we add a button to the sidebar that opens this in the instructor view?

OK, the display isn't quite what I had in mind yet. Rememember that this is for students to view, not for instructors. It should not have the facilitation notes. Ideally it can stay full screen and updates as the activity progresses, although maybe I'm just seeing the pre-start view now? In that case let's have a generic text about waiting for class to start.

But even with a session started it shows the overview. I'd like to be able to just put it up and then run the activity and not worry about having to manipulate it, so it should use the entire screen but not require scrolling.