rainytrek / README.md
annanurov's picture
Updated ReadMe.MD
ffab7b5 verified
---
title: Weather Forecast
emoji: 🌀️
colorFrom: blue
colorTo: green
sdk: docker
pinned: false
app_port: 7860
---
# 🌀️ Weather Forecast β€” Powered by Open-Meteo & Mistral 7B
Ask anything in natural language and get a **7-day weather forecast** for every city you mention.
## How it works
```
User prompt β†’ Mistral-7B (HF Inference API) β†’ list of cities
↓
Open-Meteo Geocoding API β†’ lat/lon for each city
↓
Open-Meteo Forecast API β†’ 7-day daily forecast
↓
Streamlit UI
```
## Stack
| Layer | Technology |
|-------|-----------|
| UI | Streamlit |
| Containerisation | Docker |
| LLM (NLP parsing) | Mistral-7B-Instruct-v0.3 via HF Inference API |
| Geocoding | Open-Meteo Geocoding API (free, no key) |
| Weather data | Open-Meteo Forecast API (free, no key) |
## Setup
files :
weather-app/
β”œβ”€β”€ src
β”œβ”€β”€src
└──streamlit_app.py
β”œβ”€β”€ Dockerfile
β”œβ”€β”€ requirements.txt
└── README.md
```
Open https://annanurov-rainytrek.hf.space/
## Example prompts
- *"What's the weather like in Paris and Tokyo?"*
- *"I'm travelling from Berlin to Rome next week, what should I pack?"*
- *"Compare weather in London, New York and Sydney"*
## Notes
- Open-Meteo is fully **open-source and free** β€” no API key required.
- The LLM call uses the **Hugging Face serverless Inference API** (`mistralai/Mistral-7B-Instruct-v0.3`). A free HF account token is sufficient for moderate usage.
- Temperatures are in **Β°C**, wind speed in **km/h**, precipitation in **mm**.