Radtracker

Personal productivity dashboard for teleradiology. Track revenue, goals and trends with Streamlit + SQLite. English-native UI with a Brazilian Portuguese option.

Features

• 📅 Today: KPI cards, distribution by modality, 7-day sparkline
• 📆 This Month: goal progress, daily revenue, pace alerts
• 📈 Analysis: moving averages (MA7/MA30), weekly comparison, mix evolution
• 🤖 AI Insights: GPT-OSS 120B via OpenRouter (rule-based fallback when offline)
• 🔒 Authentication: login + password (scrypt) with optional TOTP 2FA, configured via SSH
• ⚙️ Settings: price per exam, monthly goal, your name, API key, AI prompt

Prerequisites

• Python 3.12+
• uv (package manager)

Installation

git clone https://github.com/docg1701/radtracker.git
cd radtracker
uv sync                       # installs all dependencies

Usage

uv run streamlit run app.py   # http://localhost:8501

The authentication gate requires data/auth.json. For local use, create a scratch one:

python -c "from src.auth_store import create_bootstrap_auth; print(create_bootstrap_auth('dev', 'dev-password-123', 'data/auth.json', cookie_secure=False))"

(In production, deploy.yml creates this file automatically from the vault — see docs/deployment.md.)

AI (OpenRouter)

For AI-generated insights:

1. Create a free account at https://openrouter.ai
2. Generate an API key
3. In the Settings tab, paste your key into the "OpenRouter API key" field
4. Optionally, customize your name and the AI prompt in the same tab
5. In the AI Chat tab, click "Start analysis"
6. Without a key, the rule-based fallback is used automatically

The key is stored locally in the SQLite database — no .env files.

Running the tests

uv run pytest tests/ -v
uv run pytest tests/ -v --cov=src --cov-report=term-missing

Project structure

radtracker/
├── app.py                  # Streamlit entry point (auth gate after set_page_config)
├── src/
│   ├── auth_crypto.py      # scrypt + TOTP (RFC 6238) + HMAC session tokens
│   ├── auth_store.py       # data/auth.json + gate helpers
│   ├── db.py               # SQLite schema + CRUD (5 tables)
│   ├── calculations.py     # Business logic (earnings, MA, projections)
│   ├── charts.py           # Plotly charts (donut, gauge, line)
│   ├── charts_analysis.py  # Analysis charts (MA, WoW, mix evolution)
│   ├── chart_colors.py     # Shared color palette + hex_to_rgba
│   ├── formatting.py       # fmt_money, MONTHS, md_escape
│   ├── i18n.py             # EN/PT translation catalog + translate()/t()
│   ├── insights_rules.py   # Rule-based insights engine (bilingual)
│   ├── llm_client.py       # OpenRouter GPT-OSS 120B client
│   ├── text_sanitize.py    # Markdown/$ sanitization for LLM output
│   └── ui/
│       ├── login.py        # Auth gate (login, TOTP, logout, language selector)
│       ├── sidebar.py      # Data entry form
│       ├── today.py        # "Today" tab
│       ├── month.py        # "This Month" tab
│       ├── analysis.py     # "Analysis" tab
│       ├── chat.py         # "AI Chat" tab
│       └── settings.py     # "Settings" tab
├── scripts/
│   └── manage_auth.py      # SSH auth management CLI: 2FA, password, repair
├── tests/                  # Test suite (326 tests)
├── data/                   # SQLite DB + auth.json (gitignored)
├── docs/                   # Guides and design reference
├── pyproject.toml          # Project metadata + dependencies
├── uv.lock                 # Locked dependency versions
└── README.md

Stack

• Streamlit ≥1.54
• Pandas + NumPy
• Plotly
• SQLite (via SQLAlchemy)
• httpx (OpenRouter API)
• pytest + pytest-cov

Deploy

Radtracker can be deployed to any VPS with Docker + Ansible. Works on a local network (HTTP) or with your own domain (HTTPS + Let's Encrypt).

git clone https://github.com/docg1701/radtracker.git
cd radtracker
# follow the guide in docs/deployment.md

Production stack: Streamlit → Docker → Caddy (TLS) → fail2ban (sshd) → Ansible In-app authentication: login + TOTP 2FA (see docs/deployment.md)

See docs/deployment.md for the full guide.

License

MIT