4.7 KiB

Raw Blame History

Story 4.1: Deploy Content to Cloud - Quick Start Guide

Prerequisites

Ensure BUNNY_ACCOUNT_API_KEY is in your .env file (for creating zones):
```
BUNNY_ACCOUNT_API_KEY=your_account_api_key_here
```
Note: File uploads use per-zone storage_zone_password from the database, NOT an API key from .env. These passwords are set automatically when sites are created via provision-site or sync-sites commands.

Run the database migration:

uv run python scripts/migrate_add_deployment_fields.py

Usage

Automatic Deployment (Recommended)

Content deploys automatically after batch generation completes:

uv run python -m src.cli generate-batch \
  --job-file jobs/my_job.json \
  --username admin \
  --password mypass

Output will show deployment progress after all tiers complete:

  Deployment: Starting automatic deployment for project 123...
  Deployment: 48 articles, 6 pages deployed
  Deployment: Complete in 45.2s

Manual Deployment

Deploy (or re-deploy) a batch manually:

uv run python -m src.cli deploy-batch \
  --batch-id 123 \
  --admin-user admin \
  --admin-password mypass

Dry Run Mode

Preview what would be deployed without actually uploading:

uv run python -m src.cli deploy-batch \
  --batch-id 123 \
  --dry-run

What Gets Deployed

Articles: All generated articles with formatted_html
- Uploaded to: {slug}.html (e.g., how-to-fix-engines.html)
- URL logged to: deployment_logs/YYYY-MM-DD_tier1_urls.txt (Tier 1)
- URL logged to: deployment_logs/YYYY-MM-DD_other_tiers_urls.txt (Tier 2+)
Boilerplate Pages: About, contact, privacy (if they exist)
- Uploaded to: about.html, contact.html, privacy.html
- NOT logged to URL files

URL Logging

Deployed article URLs are automatically logged to tier-segregated files:

deployment_logs/
  2025-10-22_tier1_urls.txt
  2025-10-22_other_tiers_urls.txt

Each file contains one URL per line:

https://example.com/article-1.html
https://example.com/article-2.html
https://example.com/article-3.html

Duplicate URLs are automatically prevented (safe to re-run deployments).

Database Updates

After successful deployment, each article is updated with:

deployed_url: Public URL where content is live
deployed_at: Timestamp of deployment
status: Changed to 'deployed'

Query deployed content:

from src.database.session import db_manager
from src.database.repositories import GeneratedContentRepository

session = db_manager.get_session()
repo = GeneratedContentRepository(session)

deployed = repo.get_deployed_content(project_id=123)
for article in deployed:
    print(f"{article.title}: {article.deployed_url}")

Deployment Summary

After deployment completes, you'll see a summary:

======================================================================
Deployment Summary
======================================================================
Articles deployed:  48
Articles failed:    2
Pages deployed:     6
Pages failed:       0
Total time:         45.2s

Errors:
  Article 15 (Engine Maintenance Tips): Connection timeout
  Article 32 (Common Problems): Invalid HTML content
======================================================================

Error Handling

By default, deployment continues even if individual files fail. This ensures partial deployments complete successfully.

Failed files are:

Logged to console with error details
Listed in deployment summary
NOT marked as deployed in database

To stop on first error:

uv run python -m src.cli deploy-batch \
  --batch-id 123 \
  --continue-on-error false

Troubleshooting

"Authentication failed for zone"

Check that the storage_zone_password in your database is correct. This is set when sites are created via provision-site or sync-sites commands.

"Article has no formatted_html to deploy"

Ensure articles have templates applied. This happens automatically during batch processing in _post_process_tier().

"Site not found"

Ensure articles are assigned to sites. This happens automatically during batch processing via site assignment logic.

Manual Re-deployment

To re-deploy content after fixes:

Fix the issue (update HTML, fix credentials, etc.)

Run manual deployment:

uv run python -m src.cli deploy-batch --batch-id 123

Duplicate URLs are automatically prevented in log files

Integration with Other Stories

Story 3.1: Articles must be assigned to sites before deployment
Story 3.4: Boilerplate pages are deployed if they exist in site_pages table
Story 4.2: URL log files are consumed by post-deployment processes
Story 4.3: Database status updates enable deployment tracking

4.7 KiB Raw Blame History