LeonsLab

A creative-space portfolio for Leon Rösch — built with Astro, deployed on Vercel.

Local development

npm install
npm run dev      # → http://localhost:4321
npm run build    # → static output to ./dist

Project layout

src/
  components/   Reusable UI (Header, Footer, Hero, Cartoon/Sticker blocks…)
  views/        Page-level views composed from components
  pages/        Astro file-based routes (DE at root, EN under /en/)
  layouts/      BaseLayout
  content/      Astro content collections (cartoons, stickers)
  i18n/         Bilingual UI dictionary + helpers
  styles/       Tailwind v4 + global tokens
public/         Static assets (optimised images live here)

Primer Design Tool

A sub-page under /primer-design ships an In-Fusion primer-design tool for P. aeruginosa cloning experiments. The tool covers three applications:

1. In-frame deletion on the suicide vector pEXG2 — 4 primers, 3-fragment In-Fusion assembly.
2. In-locus C-terminal tagging on pEXG2 — 4 primers, 3-fragment assembly with extended tag-junction homology.
3. Constitutive plasmid expression on pBBR1MCS2 — 2 primers, 2-fragment assembly.

The page itself lives in src/views/PrimerDesignView.astro. The Python backend (Vercel serverless function) sits at api/design/handler.py and delegates into lib/primer_design/. Genome FASTAs are streamed from a Cloudflare R2 bucket — they are not committed.

The architectural contract (scope, hard-validation rules, application shapes) lives in docs/primer_design/decisions_log.md.

Phase status

• Phase 1 (current): scaffolding only — folder tree, placeholder modules, navigation entry, deployment config. No algorithm yet.
• Phase 2: algorithm implementation, R2 genome upload, PDF writer, end-to-end wiring.

Elastase Assay Calculator

A sub-page under /elastase-assay calculates volumes for a DQ Elastin elastase assay from a single input (number of samples) and returns a PDF report.

Since the tool has only one integer input, every possible report (num_samples 1–100 — a 96-well plate fits at most 39 duplicate samples alongside the fixed standard curve and negative control, so 100 leaves headroom without being unbounded) is pre-rendered offline and stored in the same R2 bucket used for the primer-design genomes, under the elastase-assay/reports/ prefix — regenerate/upload with scripts/upload_elastase_reports.py. The runtime path (api/elastase-assay.js) is a small Node.js function that just streams the matching object back; there's no Python involved at request time, which avoids running two Python serverless functions in the same deployment (api/design is the only one). The generation logic itself lives in lib/elastase_assay/ for local regeneration.