AWSBuilderFeed Twitter Bot

Automated pipeline to fetch AWS Builder articles and post them to Twitter/X.

Follow the bot: @uAWSBuilderFeed

What This Does

This bot automatically:

1. Fetches new articles from AWS Builder every hour
2. Formats tweets with title, hashtags, and URL
3. Posts one article per hour to Twitter via Make.com → Buffer
4. Prevents duplicate posts using SQLite database

Architecture

Prefect (Scheduler)
    ↓ Every hour
Fetch Articles → SQLite Queue
    ↓ Every hour
Post Tweet → Make.com Webhook → Buffer → Twitter

Key Features:

• ✅ LIFO queue ensures articles posted in order
• ✅ Duplicate detection (never posts same article twice)
• ✅ Hashtags from article tags (first 3)
• ✅ Automatic scheduling with Prefect
• ✅ Webhook delivery for instant posting

Setup

1. Install Dependencies

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Initialize Database

PYTHONPATH=. python -c "from src.database import init_db; init_db()"

3. Configure Make.com Webhook (Optional)

For automated posting to Twitter:

1. Create webhook in Make.com
2. Connect to Buffer → Twitter
3. Add to .env:

   MAKECOM_WEBHOOK_URL=https://hook.us2.make.com/your_webhook
   MAKECOM_API_KEY=your_api_key

See WEBHOOK_SETUP.md for detailed instructions.

Usage

Manual Testing

# Fetch articles
PYTHONPATH=. python -c "from src.fetcher import process_articles; print(process_articles())"

# Post one tweet
PYTHONPATH=. python -c "from src.twitter import post_tweet; result = post_tweet(); print(f'Mode: {result[\"mode\"]}')"

Run with Prefect

Terminal 1 - Start Prefect Server:

prefect server start

Terminal 2 - Deploy Flows:

source .venv/bin/activate
PYTHONPATH=. python deploy.py

Monitor: Open http://localhost:4200

Utility Scripts

./scripts/reset_db.sh   # Reset database and fetch articles
./scripts/delete_db.sh  # Delete database only

Project Structure

builderfeed/
├── src/
│   ├── database.py    # SQLite operations
│   ├── fetcher.py     # AWS Builder API
│   ├── twitter.py     # Tweet formatting & webhook
│   └── flows.py       # Prefect flows
├── data/
│   ├── builderfeed.db      # SQLite database
│   ├── tweets_queue.json   # JSON backup
│   └── mock_tweets.txt     # Human-readable log
├── scripts/
│   ├── reset_db.sh         # Reset database
│   └── delete_db.sh        # Delete database
├── deploy.py          # Prefect deployment
├── config.py          # Configuration
└── dev_log/           # MMDD development log

Schedules

• Fetch Flow: Every hour (0 * * * *)
• Tweet Flow: Every hour (0 * * * *)

Tweet Format

Article Title

#hashtag1 #hashtag2 #hashtag3

https://builder.aws.com/content/...

Development

Built using MMDD (Micromanaged Driven Development) methodology.

See dev_log/ for complete development history:

• Unit 01: Database setup
• Unit 02: Article fetching
• Unit 03: Tweet formatting
• Unit 04: Prefect orchestration
• Unit 05: Make.com integration

Tech Stack

• Python 3.12+
• Prefect 3.x - Workflow orchestration
• SQLite - Article queue and duplicate tracking
• httpx - HTTP client for API calls
• Make.com - Webhook automation
• Buffer - Social media scheduling

Links

• Twitter Bot: @uAWSBuilderFeed
• AWS Builder: https://builder.aws.com
• Make.com: https://make.com
• Buffer: https://buffer.com
• Prefect: https://prefect.io

License

MIT License - See LICENSE file