minzeyaphyo/burmddit
1
0
Fork 1
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
dddb86ea9424f7ae5912563742e47c6bef713516
burmddit/DEPLOYMENT-GUIDE.md

9.4 KiB
Raw Blame History

Burmddit - Complete Deployment Guide

From Zero to Production in 30 Minutes

Target URL: burmddit.vercel.app
Cost: $0-$60/month (mostly Claude API)

📋 PRE-REQUISITES

Accounts Needed (All Free to Start):

  1. ✅ GitHub - github.com (code hosting)
  2. ✅ Vercel - vercel.com (frontend hosting)
  3. ✅ Railway - railway.app (backend + database)
  4. ✅ Anthropic - console.anthropic.com (Claude API for translation)

Time to create accounts: ~10 minutes

🚀 DEPLOYMENT STEPS

STEP 1: Push Code to GitHub (5 mins)

# On your local machine or server:
cd /home/ubuntu/.openclaw/workspace/burmddit

# Initialize git
git init
git add .
git commit -m "Initial Burmddit deployment"

# Create repository on GitHub (via website):
# 1. Go to github.com/new
# 2. Name: burmddit
# 3. Public or Private (your choice)
# 4. Create repository

# Push to GitHub
git remote add origin https://github.com/YOUR_USERNAME/burmddit.git
git branch -M main
git push -u origin main

✅ Done! Your code is now on GitHub

STEP 2: Deploy Backend to Railway (10 mins)

2.1: Create Railway Project

  1. Go to railway.app
  2. Click "New Project"
  3. Select "Deploy from GitHub repo"
  4. Choose your burmddit repository
  5. Railway will auto-detect it as a Python project

2.2: Add PostgreSQL Database

  1. In your Railway project, click "+ New"
  2. Select "Database" → "PostgreSQL"
  3. Railway creates database instantly
  4. Copy the DATABASE_URL (Click database → Connect → Copy connection string)

2.3: Set Environment Variables

In Railway project settings → Variables, add:

DATABASE_URL=<paste from step 2.2>
ANTHROPIC_API_KEY=<your Claude API key from console.anthropic.com>
ADMIN_PASSWORD=<choose a secure password>
PYTHONPATH=/app/backend

2.4: Configure Build

Railway → Settings → Build:

  • Root Directory: backend
  • Build Command: pip install -r requirements.txt
  • Start Command: python run_pipeline.py

2.5: Initialize Database

In Railway console (click database service):

python init_db.py init

✅ Done! Backend is live on Railway

STEP 3: Deploy Frontend to Vercel (5 mins)

3.1: Connect GitHub to Vercel

  1. Go to vercel.com/new
  2. Click "Import Git Repository"
  3. Select your burmddit repo
  4. Vercel auto-detects Next.js

3.2: Configure Settings

Framework Preset: Next.js
Root Directory: frontend
Build Command: (default next build)
Output Directory: (default .next)

3.3: Set Environment Variables

In Vercel project settings → Environment Variables:

DATABASE_URL=<same as Railway PostgreSQL URL>
NEXT_PUBLIC_SITE_URL=https://burmddit.vercel.app

3.4: Deploy

Click "Deploy"
Wait 2-3 minutes...

✅ Done! Frontend is live at burmddit.vercel.app!

STEP 4: Set Up Automation (5 mins)

Option A: GitHub Actions (Recommended)

Create file: .github/workflows/daily-publish.yml

name: Daily Content Pipeline

on:
  schedule:
    # Run at 6 AM UTC daily
    - cron: '0 6 * * *'
  workflow_dispatch: # Allow manual trigger

jobs:
  run-pipeline:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      
      - name: Set up Python
        uses: actions/setup-python@v4
        with:
          python-version: '3.11'
      
      - name: Install dependencies
        run: |
          cd backend
          pip install -r requirements.txt
      
      - name: Run pipeline
        env:
          DATABASE_URL: ${{ secrets.DATABASE_URL }}
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
        run: |
          cd backend
          python run_pipeline.py

Add secrets to GitHub:

  1. Repo → Settings → Secrets and variables → Actions
  2. Add: DATABASE_URL and ANTHROPIC_API_KEY

Option B: Railway Cron (Simpler but less flexible)

In Railway, use built-in cron:

  1. Project settings → Deployments
  2. Add cron trigger: 0 6 * * *
  3. Command: python backend/run_pipeline.py

✅ Done! Auto-publishes 10 articles daily at 6 AM UTC!

🧪 TESTING

Test the Pipeline Manually:

# SSH into Railway or run locally with env vars:
cd backend

# Test scraper
python scraper.py

# Test compiler
python compiler.py

# Test translator
python translator.py

# Test full pipeline
python run_pipeline.py

Check Database:

python init_db.py stats

Test Website:

  1. Visit burmddit.vercel.app
  2. Should see homepage with categories
  3. Once pipeline runs, articles will appear

💰 COST BREAKDOWN

Monthly Costs:

Free tier (Month 1-3):

  • Vercel: FREE (Hobby plan)
  • Railway: FREE ($5 credit/month) or $5/month
  • GitHub Actions: FREE (2,000 mins/month)
  • Total: $0-$5/month

With Claude API (Month 1+):

  • Claude API: ~$30-60/month
    • 10 articles/day × 30 days = 300 articles
    • ~1,500 words per article = 2,000 tokens
    • 300 × 2,000 = 600k tokens/month
    • Claude pricing: ~$0.015/1k tokens input, $0.075/1k tokens output
    • Estimated: $30-60/month
  • Total: $35-65/month

Optimization tips:

  • Use Claude 3 Haiku for translation (cheaper, still good quality)
  • Batch translations to reduce API calls
  • Cache common translations

📊 MONITORING

Check Pipeline Status:

Railway Dashboard:

  • View logs for each pipeline run
  • Check database size
  • Monitor CPU/memory usage

Vercel Dashboard:

  • Page views
  • Load times
  • Error rates

Database Stats:

python init_db.py stats

🐛 TROUBLESHOOTING

Pipeline Not Running

Check logs:

# Railway → Deployments → View logs
# Or locally:
tail -f burmddit_pipeline.log

Common issues:

  • API key not set → Check environment variables
  • Database connection failed → Verify DATABASE_URL
  • Scraping blocked → Check rate limits, use VPN if needed

No Articles Showing

Verify pipeline ran:

python init_db.py stats

Check articles table:

SELECT COUNT(*) FROM articles WHERE status = 'published';

Manual trigger:

python backend/run_pipeline.py

Translation Errors

Check API key:

echo $ANTHROPIC_API_KEY

Test translation:

python backend/translator.py

Rate limit hit:

  • Anthropic free tier: 50 requests/min
  • Paid tier: 1,000 requests/min
  • Add delays if needed

🔧 CUSTOMIZATION

Change Number of Articles:

Edit backend/config.py:

PIPELINE = {
    'articles_per_day': 20,  # Change from 10 to 20
    ...
}

Add New Content Sources:

Edit backend/config.py:

SOURCES = {
    'your_source': {
        'enabled': True,
        'url': 'https://example.com/feed/',
        ...
    }
}

Update backend/scraper.py to handle new source format.

Change Translation Quality:

Use Claude 3 Opus for best quality (more expensive):

TRANSLATION = {
    'model': 'claude-3-opus-20240229',
    ...
}

Or Claude 3 Haiku for lower cost (still good):

TRANSLATION = {
    'model': 'claude-3-haiku-20240307',
    ...
}

🎨 FRONTEND CUSTOMIZATION

Change Colors:

Edit frontend/tailwind.config.ts:

colors: {
  primary: {
    500: '#YOUR_COLOR',
    // ... other shades
  }
}

Replace text logo in frontend/components/Header.tsx:

<Image src="/logo.png" width={40} height={40} alt="Burmddit" />

Add logo.png to frontend/public/

Change Fonts:

Update frontend/app/layout.tsx with Google Fonts link

📈 SCALING

When Traffic Grows:

Vercel (Frontend):

  • Free tier: Unlimited bandwidth
  • Pro tier ($20/mo): Analytics, more team members

Railway (Backend + DB):

  • Free tier: $5 credit (good for 1-2 months)
  • Hobby tier: $5/mo (500 hours)
  • Pro tier: $20/mo (unlimited)

Database:

  • Railway PostgreSQL: 100 MB free → 8 GB on paid
  • For larger: Migrate to Supabase or AWS RDS

Claude API:

  • Pay-as-you-go scales automatically
  • Monitor costs in Anthropic console

✅ POST-DEPLOYMENT CHECKLIST

After deployment, verify:

  • Frontend loads at burmddit.vercel.app
  • Database tables created (run init_db.py stats)
  • Pipeline runs successfully (trigger manually first)
  • Articles appear on homepage
  • All 4 categories work
  • Mobile responsive (test on phone)
  • Search works (if implemented)
  • Burmese fonts display correctly
  • GitHub Actions/Railway cron scheduled
  • Environment variables secure (not in code)

🎉 SUCCESS!

You now have: ✅ Fully automated AI content platform
✅ 10 articles published daily
✅ Professional Burmese website
✅ Zero manual work needed
✅ Scalable infrastructure

Next steps:

  1. Monitor first week's content quality
  2. Adjust scraping sources if needed
  3. Promote on social media
  4. Apply for Google AdSense (after 3 months)
  5. Build email list
  6. Scale to 20 articles/day if demand grows

📞 SUPPORT

Issues? Check:

  1. Railway logs
  2. Vercel deployment logs
  3. GitHub Actions run history
  4. Database stats (init_db.py stats)

Still stuck? Review code comments in:

  • backend/run_pipeline.py (main orchestrator)
  • backend/scraper.py (if scraping issues)
  • backend/translator.py (if translation issues)

Built by Bob (OpenClaw AI) for Zeya Phyo
Deploy time: ~30 minutes
Maintenance: ~5 minutes/week
Passive income potential: $2,000-5,000/month 🚀

Let's make AI accessible to all Burmese speakers! 🇲🇲

Reference in New Issue View Git Blame Copy Permalink