Skip to content

Roadmap: Event Automation Pipeline

Jump to bottom
Zayan Hoodani edited this page Feb 9, 2026 · 7 revisions

Roadmap: Event Automation Pipeline

There's a lot of existing resources for students, but they're spread out across emails, group chats, and flyers. If this website is to become a unified source of information about events on campus, we have to make announcing events as easy as possible. So, we're creating a pipeline that allows for events to be created from a photo of a flyer or a forwarded email.

Feature Roadmap

Use this outline to track our progress. Sub-features are nested to indicate scope, and ordered by priority.

1. Create Event from Image (Flyer Parsing)

  • 1.1. Infrastructure & Research
    • Research & Select OCR Engine (Issue: Research OCR)
  • 1.2. Core Logic
    • Create Mock OCR Service (for testing)
    • Create Date Extraction Utility (Regex)
    • Create Location Extraction Utility
  • 1.3. UI Integration
    • Integrate Camera/Upload on Create Event Page
    • Auto-fill form fields from OCR result

2. Create Event from Email Forward

  • 2.1. Infrastructure & Research
    • Research & Select Email Webhook Provider (Issue: Research Email Provider)
    • Configure DNS & "Inbound Parse" settings
  • 2.2. Webhook Implementation
    • Create Basic Webhook API Route (POST /api/webhooks/email)
    • Define Email Payload Validation Schema

4. Handle Recurring Events

  • 4.1. Database Schema
    • Add recurrence_rule column to Events table
  • 4.2. Logic
    • Parse natural language ("every Friday") into recurrence rules

5. Image Ingestion & OCR (The Eyes)

  • 5.1. Infrastructure
    • Set up AWS S3 for Image Storage (Issue: S3 Client & Upload)
  • 5.2. OCR Implementation (Scribe.js)
    • Create OCR Helper Function (Issue: Implement Scribe OCR)
    • Decision: We are using Scribe.js (Local/Serverless) over Google Cloud Vision to keep costs zero.

6. AI Event Parsing (The Brain - DeepSeek R1)

  • 6.1. AI Integration
    • Set up DeepSeek API Client (Issue: DeepSeek Setup)
    • Create Prompt Engineering Template (Issue: Define Event Parsing Prompt)
  • 6.2. Parsing Logic
    • Implement parseEventFromText service (Connects OCR output to DeepSeek)
    • Map AI JSON response to Database Schema

7. User Interface Integration

  • 7.1. Create Event Page
    • Add "Upload Flyer" button to UI
    • Auto-fill form fields with data returned from AI

Clone this wiki locally