Upload README.md with huggingface_hub

README.md

@@ -1,170 +1,48 @@
-# 🌍 International Trade Flow Predictor with AI Assistant 🚢📈🤖
-
-[![Deploy to Hugging Face Spaces](https://github.com/jomasego/international_trade_predictions/actions/workflows/deploy-space.yml/badge.svg)](https://github.com/jomasego/international_trade_predictions/actions/workflows/deploy-space.yml)
-
-Welcome to the **International Trade Flow Predictor**! 🎉
-
-A fun, interactive web app to explore, analyze, and predict international trade flows using real data and machine learning. Now with AI-powered trade assistance! Built with Flask, vanilla JS, and a dash of data science magic. 🧙‍♂️✨
-
-[![Hugging Face Space](https://img.shields.io/badge/🤗-Hugging%20Face%20Space-cyan.svg)](https://huggingface.co/spaces/jomasego/ai-trade-assistant)
-
 ---
-
-## 🚀 Features
-
-- **Tabbed UI Navigation** 🗂️
-  - Basics: Quick trade lookups
-  - Exports/Imports by Country 🌎
-  - Exports/Imports by Product 📦
-  - Rankings 🏆
-  - Bilateral Trade ↔️
-  - Data Download 💾
-  - ML Prediction 🤖
-
-- **Data Visualizations** 📊
-  - Tables, charts, and downloadable CSVs everywhere!
-  - Interactive country/product selection
-  - All powered by UN COMTRADE, WTO, and World Bank data
-
-- **Machine Learning Models** 🧠
-  - Linear Regression
-  - XGBoost
-  - LSTM (Deep Learning)
-  - Predict future trade flows and visualize both historical and predicted values!
-
-- **Modern, Accessible UI**
-  - Keyboard navigation, ARIA attributes, and a clean look
-  - No frontend frameworks, just pure JS!
-
-- **AI Trade Assistant** 🤖
-  - Powered by Google Gemma-2b
-  - Interactive chat interface for trade data queries
-  - HS code explanation and interpretation
-  - Trade pattern recommendations
-  - Context-aware responses to help navigate the application
-
+title: Trade Flow Predictor
+emoji: 🌐
+colorFrom: blue
+colorTo: green
+sdk: docker
+sdk_version: "1.0"
+app_file: app.py
+pinned: false
 ---
 
-## 🛠️ Deployment Options
-
-You have two options for deploying this application:
-
-### Option 1: Full Web Application with Integrated AI Assistant
-
-This option gives you the complete Trade Flow Predictor application with the AI Assistant directly integrated into the application interface.
-
-1. **Clone the repo:**
-   ```bash
-   git clone https://github.com/jomasego/international_trade_predictions.git
-   cd international_trade_predictions/comtrade_predictions
-   ```
-
-2. **Set up environment variable:**
-   Create a `.env` file in the root directory with your Hugging Face API token:
-   ```
-   HUGGINGFACE_API_TOKEN=your_token_here
-   ```
-   This token is required to access the Google Gemma-2b model via Hugging Face Inference API.
-
-3. **Install dependencies:**
-   ```bash
-   pip install -r requirements.txt
-   ```
-
-4. **Launch the app:**
-   ```bash
-   python app.py
-   ```
-
-5. **Open your browser:**
-   - Go to [http://127.0.0.1:5000](http://127.0.0.1:5000)
-   - Navigate to the AI Assistant tab to interact with the integrated assistant
+# International Trade Flow Predictor with AI Assistant
 
-### Option 2: Standalone AI Trade Assistant on Hugging Face Space
-
-This option deploys just the AI Assistant component as a separate application on Hugging Face Spaces. This is useful if you only want to provide the AI Assistant functionality without the full application.
-
-1. **Use our existing space:**
-   - Visit [AI Trade Assistant](https://huggingface.co/spaces/jomasego/ai-trade-assistant) to use the pre-deployed version
-
-2. **OR deploy your own space:**
-
-   a. **Create a new Space** on Hugging Face with the Flask SDK
-
-   b. **Upload the files from the `clean-space` directory:**
-      - `app.py`
-      - `requirements.txt`
-      - `templates/index.html`
-
-   c. **Set up environment variable:**
-      - Add `HUGGINGFACE_API_TOKEN` to your space under Settings > Repository Secrets
-
-   d. **Your space will automatically deploy** and be accessible through the provided URL
-
-### Option 3: Automatic GitHub Actions Deployment
-
-This project is configured to automatically deploy to Hugging Face Spaces using GitHub Actions:
-
-1. **One-time setup:**
-   - Add required secrets to your GitHub repository:
-     - `HF_TOKEN`: Your Hugging Face API token
-     - `SPACE_NAME`: Your desired Space name
-     - `HUGGINGFACE_API_TOKEN`: Same as your HF_TOKEN
-
-2. **Automatic deployment:**
-   - Every push to the `main` branch triggers a deployment
-   - Your app will be available at: `https://huggingface.co/spaces/jomasego/[SPACE_NAME]`
-
-3. **Manual triggering:**
-   - You can also manually trigger deployment from the GitHub Actions tab
-
-### Using the AI Assistant
-
-Regardless of which deployment option you choose:
+Welcome to the **International Trade Flow Predictor**! 🎉
 
-1. Navigate to the AI Assistant interface
-2. Type your trade-related question in the chat input
-3. Use specialized features:
-   - HS Code Lookup: Get detailed explanations of HS codes
-   - Trade Recommendations: Receive insights on trade patterns based on countries or product categories
+A fun, interactive web app to explore, analyze, and predict international trade flows using real data and machine learning. Now with AI-powered trade assistance! Built with Flask, vanilla JS, and a dash of data science magic.
 
----
+## Features
 
-## 🔍 Explore the Tabs
+- **Interactive Data Visualization**: Explore international trade data with various chart types and filters
+- **Machine Learning Predictions**: Forecast future trade values using multiple ML models
+- **AI Trade Assistant**: Get intelligent assistance with all your trade-related questions
+- **Comprehensive Data Access**: Retrieve and analyze data from UN COMTRADE, WTO, and World Bank sources
 
-- **Basics:**
-  - Quick lookups by country, year, commodity, and flow
-- **Exports/Imports by Country:**
-  - See trade values for all countries in a given year/commodity
-- **Exports/Imports by Product:**
-  - Explore top traded products for a country
-- **Rankings:**
-  - Who are the top exporters/importers? Find out!
-- **Bilateral Trade:**
-  - Analyze trade between any two countries
-- **Data Download:**
-  - Custom CSV downloads for power users
-- **Prediction (ML):**
-  - Select countries, commodity, and year, pick a model, and predict the future! See both historical and predicted values plotted together.
+## AI Assistant Capabilities
 
----
+The built-in AI Trade Assistant, powered by Google Gemma-2b via Hugging Face, can help you:
 
-## ⚡ Tech Stack
-- Python 3
-- Flask
-- Pandas, scikit-learn, XGBoost, Keras/TensorFlow
-- HTML5, CSS3, JavaScript (vanilla!)
+- Understand trade terminology and HS codes
+- Explore interesting trade patterns
+- Find specific trade data
+- Interpret visualization results
+- Provide insights about international trade
 
----
+## Using the Application
 
-## 🤝 Contributing
-PRs welcome! Open an issue, fork, or just say hi. 😊
+1. Explore different tabs to access various features
+2. Select countries, products, and years to visualize data
+3. Try the machine learning predictions for future trade values
+4. Chat with the AI Assistant for trade-related questions
 
----
+## Privacy & Data Usage
 
-## 📬 Contact
-- [GitHub Issues](https://github.com/jomasego/international_trade_predictions/issues)
+This application does not collect any personal data. API usage is subject to fair use policies of the data sources (UN COMTRADE, WTO, World Bank).
 
----
+## About
 
-Enjoy exploring the world of trade! 🌐🚀
+Created by José María Seguí Gómez-Olea. The AI component is powered by Google Gemma-2b through Hugging Face Inference API.