felipevillegas.com

Personal professional portfolio site for Felipe Villegas — L&D practitioner and software engineer.

Built with Django 6.0 + Wagtail 7.3. Content is managed through the Wagtail admin; there are no hardcoded page routes.

Local setup

Requires Python 3.13 and uv.

git clone https://github.com/dr-rompecabezas/felipevillegas.git
cd felipevillegas

cp .env.example .env
# edit .env — at minimum set SECRET_KEY

uv sync
uv run python manage.py migrate
uv run python manage.py create_initial_pages
uv run python manage.py populate_home_page
uv run python manage.py populate_projects
uv run python manage.py populate_seo
uv run python manage.py createsuperuser
uv run python manage.py runserver

• Site: http://localhost:8000
• Wagtail admin: http://localhost:8000/admin/

Dev uses SQLite by default. Set DATABASE_URL in .env to use PostgreSQL instead.

Page tree

create_initial_pages builds the full tree and points the default Site at HomePage:

Root
└── HomePage                 → /
    ├── WorkIndexPage        → /work/
    ├── PhotographyIndexPage → /photography/
    └── ContactPage          → /contact/

Content seeding

Four management commands populate launch content after the page tree exists:

Command	What it does
populate_home_page	Seeds the HomePage StreamField (hero, about, skills, contact CTA)
populate_projects	Creates all TechTag snippets and ProjectPage entries under WorkIndexPage
populate_seo	Sets seo_title and search_description on the four static/index pages

All three are idempotent and accept --force to overwrite existing content. ProjectPage and GalleryPage can also be added and edited through the Wagtail admin at /admin/.

Environment variables

See .env.example for the full list. Production-only variables (S3) can be left blank in development.

Variable	Required	Notes
SECRET_KEY	Always	Any long random string in dev
DATABASE_URL	Production	Defaults to SQLite in dev
AWS_*	Production	S3 media storage
WAGTAILADMIN_BASE_URL	Production	Used in Wagtail email links
ANTHROPIC_API_KEY	Chat enabled	Without it, the chat view returns 503
ANTHROPIC_MODEL	Optional	Defaults to claude-haiku-4-5-20251001
CHAT_DAILY_INPUT_TOKEN_BUDGET	Optional	Per-IP daily input cap (default: 5000)
CHAT_DAILY_OUTPUT_TOKEN_BUDGET	Optional	Per-IP daily output cap (default: 2000)
CHAT_RPM	Optional	Per-IP requests per minute (default: 6)
CHAT_INPUT_MAX_CHARS	Optional	Hard cap on user message length (default: 1000)
CHAT_MAX_OUTPUT_TOKENS	Optional	max_tokens passed to Anthropic (default: 400)

Interactive page chat (Anthropic)

The /interactive/ page hosts a single-turn chat backed by Claude Haiku 4.5. The system prompt is editable in the Wagtail admin (InteractivePage → AI chat), and the page has a chat_enabled toggle.

Cost and abuse controls are layered, all keyed by client IP:

1. Hard input length cap (CHAT_INPUT_MAX_CHARS).
2. Per-IP requests-per-minute (CHAT_RPM).
3. Per-IP daily input + output token budgets (CHAT_DAILY_INPUT_TOKEN_BUDGET, CHAT_DAILY_OUTPUT_TOKEN_BUDGET).

These are the first line of defence; set a hard monthly spend cap in the Anthropic console ($10 is a sensible starting point) as the last line. The in-app caps are not enforceable below the API surface, so the console cap is the only thing that bounds total spend if a worker dies mid-request, the cache is wiped, or budget settings are misconfigured.

Local dev:

# In .env
ANTHROPIC_API_KEY=sk-ant-...

uv run python manage.py runserver
# Visit /interactive/ and pick a starter chip or type a question.

For media-performance tuning in production:

• AWS_S3_CACHE_CONTROL defaults to public, max-age=31536000, immutable for uploaded media objects.
• Set AWS_QUERYSTRING_AUTH=False only when media objects are publicly readable (for example via public bucket policy or CloudFront) so image URLs are stable and cacheable.

Deployment

Deployed on Railway. Push to main triggers an auto-deploy:

1. python manage.py migrate (release command)
2. gunicorn config.wsgi (start command)

Static files are served by WhiteNoise. Media files are stored in S3.

Set DJANGO_SETTINGS_MODULE=config.settings.production in the Railway environment along with all production variables from .env.example.