README.md

---
title: Fist-aid assitant
emoji: 🚀
colorFrom: blue
colorTo: green
sdk: streamlit
sdk_version: 1.33.0  
app_file: app.py
pinned: false
---

# Snakebit Firstaid 🚀

**Short description:** Web application that provides quick, essential first-aid guidance for snakebite (poisoned wound) and normal wounds, and then offers an advanced ChatGPT-powered assistant to deliver tailored next-step guidance.

---

## Features

* Simple, two-panel quick-first-aid interface:

  * Left: Snakebite (poisoned wound) — short essential steps.
  * Right: Normal wound — short essential steps.
  * Users confirm completion and press **Next → Advanced** which forwards the wound type to the advanced assistant.
* Advanced assistant powered by OpenAI ChatGPT (via the `openai` package):

  * Confirms basic-first-aid steps completed.
  * Collects victim information (name, age, height, weight, notes).
  * Saves victim info to `./data/victims.json`.
  * Provides in-depth, concise first-aid & monitoring steps and supports follow-up questions.
* Safe defaults for deployment (server settings in `.streamlit/config.toml`).

---

## Files in the repo

* `app.py` — main Streamlit app (single-file app).
* `requirements.txt` — project dependencies.
* `.streamlit/config.toml` — UI theme + server defaults.
* `.streamlit/secrets.toml` — (empty by default; store `OPENAI_API_KEY` here locally if needed — **do not commit**).
* `data/victims.json` — auto-created and appended with victim records by the app (add `data/.gitkeep` if you want the empty folder tracked).
* `.gitignore` — prevents committing secrets, virtualenvs, and local data.
* `README.md` — this document.

---

## Quick start (local)

> Tested with Python 3.10+ (recommend 3.11). Adjust if your environment differs.

1. Create a virtual environment and activate it:

```bash
python -m venv .venv
# macOS / Linux
source .venv/bin/activate
# Windows (PowerShell)
.\.venv\Scripts\Activate.ps1
```

2. Install dependencies:

```bash
pip install -r requirements.txt
```

3. Provide your OpenAI API key securely. Recommended options:

* Add to environment (temporary):

```bash
export OPENAI_API_KEY="sk-..."
# Windows PowerShell
$env:OPENAI_API_KEY = "sk-..."
```

* OR (local only, not for public repos) add to `.streamlit/secrets.toml`:

```toml
OPENAI_API_KEY = "sk-REPLACE_WITH_YOUR_KEY"
# Optionally set the model
# OPENAI_MODEL = "gpt-4o-mini"
```

4. Run the app:

```bash
streamlit run app.py
```

Open the address printed by Streamlit (usually [http://localhost:8501](http://localhost:8501)).

---

## Deploying to Hugging Face Spaces (Streamlit)

1. Create a new Space on Hugging Face and choose **Streamlit** as the SDK.
2. Push this repository (all files except secrets) to the Space.
3. In the Space settings, add a secret named `OPENAI_API_KEY` with your OpenAI key. Do **not** commit `secrets.toml` to the repo.
4. Start the Space — it will use `app.py` automatically.

---

## Configuration & secrets

* `.streamlit/config.toml` contains the UI theme and safe server defaults (headless mode, address `0.0.0.0`, disabled XSRF for certain container environments).
* **Never commit** `.streamlit/secrets.toml` (contains API keys). Add it to `.gitignore` (already provided).

---

## Security & emergency notes

* This app is *not* a substitute for professional medical care. It gives short, essential guidance to assist non-professionals.
* If a victim is deteriorating, losing consciousness, has severe bleeding, or you suspect life-threatening poisoning, **call local emergency services immediately**.
* The app stores victim records in `./data/victims.json`. Treat this as potentially sensitive information; if you deploy publicly, review data retention policies and consider adding encryption or removing storage.

---

## Requirements

See `requirements.txt`. Minimal runtime packages include `streamlit` and `openai`. Optionally include `toml`, `python-dotenv`, and `Pillow`.

---

## Development notes

* The app accepts `OPENAI_MODEL` environment variable to override the default model used.
* The app supports both the modern `openai.OpenAI(api_key=...)` client shape and the classic `openai` package usage (the code tries both imports).
* `data/victims.json` is appended; if you want the directory tracked but empty, create `data/.gitkeep` and commit it.

---

## Contributing

If you want improvements (multi-page separation, richer UI, persistent DB, authentication, localization), open an issue or send a pull request. I can help implement features such as:

* Separate pages for the simple/advanced parts (actual `next.py` pages).
* Better UX: progress indicators, larger touch targets, i18n (Vietnamese + English).
* Replace local JSON file with small database (SQLite) or cloud storage for multi-user deployments.

---

## License

Add a license (e.g., MIT) — currently not set in this repo.