Erich Blume 5b52f18356 cv+docs: relocate content under ~/blumeops/{cv,docs}

Keeps blumeops-managed state grouped under a single namespace in the
home dir rather than scattered top-level dirs. Caddy block paths are
derived from cv_content_dir / docs_content_dir, so the role-defaults
edit propagates automatically.

Validated end-to-end on indri: tarballs extracted to the new paths,
sentinels written, second run is idempotent. Old ~/cv and ~/docs from
the earlier validation run were removed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-29 14:54:01 -07:00

2.6 KiB

Raw Permalink Blame History

title

modified

last-reviewed

CV (Resume)

Personal resume/CV served as a static HTML page with PDF download, built from YAML source via Jinja2 and WeasyPrint.

Quick Reference

Property	Value
Public URL	`cv.eblu.me` (via flyio-proxy)
Private URL	`cv.ops.eblu.me` (Caddy on indri)
Deployment	Ansible role `cv` on indri (no daemon — Caddy serves files directly)
Content dir	`~/blumeops/cv/content/` on indri
Source repo	`forge.eblu.me/eblume/cv` (private, not mirrored to GitHub)
Content packages	`forge.eblu.me/eblume/-/packages` (generic package `cv`)

Migrated from minikube to indri-native on 2026-04-29 (see cv-on-indri).

Architecture

Source: resume.yaml (content) + template.html (Jinja2) + style.css in the cv repo
Build: render.py (uv script runner) generates index.html; WeasyPrint generates resume.pdf
Release: Dagger build function packages index.html, style.css, resume.pdf into a tarball, uploaded to Forgejo generic packages
Deploy: ansible role downloads the tarball into ~/blumeops/cv/content/ on indri; Caddy serves the directory directly

Endpoints

Path	Description
`/`	Resume HTML page
`/resume.pdf`	PDF download (Caddy adds `Content-Disposition: attachment`)

Configuration

Key files (blumeops):

ansible/roles/cv/defaults/main.yml — pinned cv_version and tarball URL
ansible/roles/cv/tasks/main.yml — sentinel-gated download + extract
ansible/roles/caddy/defaults/main.yml — cv service entry (kind: static, download_paths for the PDF)

Key files (cv repo):

resume.yaml — Resume content (YAML)
template.html — Jinja2 HTML template
style.css — CSS with screen/print media queries
render.py — uv script runner (PEP 723) that renders YAML → HTML
src/cv_ci/main.py — Dagger pipeline (alpine + uv + WeasyPrint)
.forgejo/workflows/cv-release.yaml — Release workflow

Release flow

Release a new package from the cv repo (Release CV workflow)
Run the blumeops Deploy CV workflow → bumps cv_version in the ansible role and pushes
Run mise run provision-indri -- --tags cv from gilbert
Purge the Fly.io proxy cache so the new content is fetched

cv-on-indri — Operations how-to
docs — Similar architecture (Caddy serving a tarball-extracted dir)
flyio-proxy — Exposes cv.eblu.me publicly via Tailscale tunnel

2.6 KiB Raw Permalink Blame History