ResuMeme

Live on: https://black-desert-031be090f.7.azurestaticapps.net/

A small project for learning the Azure stack end-to-end: Static Web Apps, managed Functions, Blob Storage, and Azure OpenAI. The app takes an uploaded CV (PDF or DOCX), parses it, runs the content through an LLM, and renders the result back to the user. It deploys via azd to a Free-tier SWA + Standard LRS storage account, and authenticates to Azure OpenAI with a key.

Stack

• Frontend: vanilla HTML/CSS/JS on Azure Static Web Apps
• Backend: Python 3.11 Azure Functions (managed by SWA, Python v2 programming model)
• Storage: Azure Blob Storage with a 30-day lifecycle policy
• AI: Azure OpenAI (model deployment configurable via env var)
• IaC: Bicep, deployed with the Azure Developer CLI (azd)

Local development

Prereqs: Python 3.11, Node 20+, Azure Functions Core Tools v4, Static Web Apps CLI, Azurite.

# Install global tools (one time)
npm i -g azure-functions-core-tools@4 --unsafe-perm true
npm i -g @azure/static-web-apps-cli azurite

# Set up backend
cd api
python -m venv .venv
.venv\Scripts\activate     # macOS/Linux: source .venv/bin/activate
pip install -r requirements-dev.txt

# Generate test fixtures (one time)
python tests/fixtures/generate_fixtures.py
cd ..

Run the full stack locally, three terminals from the repo root:

# Terminal 1: Storage emulator
azurite --silent --skipApiVersionCheck

# Terminal 2: Functions
cd api && func start

# Terminal 3: SWA proxy on :4280
swa start frontend --api-location api

Open http://localhost:4280.

Tests

cd api
$env:PYTHONPATH = "$PWD"   # PowerShell; bash: export PYTHONPATH=$PWD
.venv\Scripts\python.exe -m pytest tests/ -v

44 tests across 7 modules, including blob client and rate limiter integration with Azurite.

Manual smoke test checklist

• Upload a PDF, renders the result at /cv/<id>
• Upload a DOCX, renders the result at /cv/<id>
• Reload the result page, output is identical (deterministic per id)
• Copy the share link, opens correctly in incognito
• File over 5MB rejected with a friendly error
• Non-PDF/DOCX rejected with a friendly error
• Invalid id (/cv/zzzzzzzz) shows the 404 page
• Two uploads from the same IP within 30 seconds, second one returns 429
• /careers and /legal resolve correctly

Deployment

azd auth login
azd up                    # first time
azd deploy                # subsequent code-only deploys

After azd up, set the four AI env vars on the deployed SWA via the portal (Static Web App, Environment variables):

Name	Example value
AZURE_OPENAI_ENDPOINT	https://<your-resource>.cognitiveservices.azure.com/
AZURE_OPENAI_KEY	(paste the key)
AZURE_OPENAI_DEPLOYMENT	gpt-4o-mini (or whatever deployment name you created)
AZURE_OPENAI_API_VERSION	2025-04-01-preview

The SWA also expects a deployment token in the GitHub repo secret AZURE_STATIC_WEB_APPS_API_TOKEN so the GitHub Actions workflow can auto-deploy on main pushes. Get the token from the portal under "Manage deployment token" and add it under Settings → Secrets and variables → Actions.

Project structure

frontend/                       Static site (HTML/CSS/JS, no build step)
  staticwebapp.config.json      SWA routes and platform config
  index.html / cv.html / ...    Pages
  js/                           Vanilla JS
  css/                          Stylesheets
api/                            Azure Functions (Python v2 model)
  function_app.py               Routes: /upload, /cv/{id}, /health, /diag
  shared/                       Parsers, blob client, rate limiter, LLM client
  tests/                        pytest suite
infra/                          Bicep templates
azure.yaml                      azd project config
.github/workflows/              GitHub Actions deploy workflow