# Hugging Face Upload Instructions

This folder contains everything needed to publish the African CP Synthetic Dataset to Hugging Face.

## Folder Contents

### Dataset Files (7 CSVs - 3.1 MB total)
- ✅ `africa_cp_train_1000.csv` - Small training set
- ✅ `africa_cp_train_5000.csv` - Medium training set  
- ✅ `africa_cp_train_10000.csv` - Large training set
- ✅ `africa_cp_balanced_1000.csv` - Balanced 50/50 set
- ✅ `africa_cp_preterm_2000.csv` - High-risk preterm cohort
- ✅ `africa_cp_cases_only_500.csv` - CP cases only
- ✅ `africa_cp_test_2000.csv` - Hold-out test set

### Documentation
- ✅ `README.md` - Dataset card (auto-displays on HF)
- ✅ `CITATION.bib` - Citation information
- ✅ `LICENSE` - CC BY-NC 4.0 license
- ✅ `dataset_info.yaml` - Dataset metadata

### Code
- ✅ `cp_data_generator.py` - Generator script
- ✅ `load_dataset.py` - Example loading code

### Configuration
- ✅ `.gitattributes` - Git LFS configuration for large files

---

## Upload Steps

### Option 1: Using Hugging Face Web Interface (Easiest)

1. **Create account** at https://huggingface.co/join

2. **Create new dataset**:
   - Go to https://huggingface.co/new-dataset
   - Dataset name: `african-cp-synthetic` (or your choice)
   - License: `cc-by-nc-4.0`
   - Make public or private

3. **Upload files**:
   - Click "Files" tab
   - Drag and drop all files from this folder
   - Or use "Add file" → "Upload files"
   - Commit message: "Initial dataset upload"

4. **Verify**:
   - README.md should auto-display as dataset card
   - All 7 CSV files should be visible
   - Check file sizes are correct

### Option 2: Using Git (Advanced)

```bash
# 1. Install git-lfs (for large files)
git lfs install

# 2. Clone your dataset repo
git clone https://huggingface.co/datasets/[your-username]/african-cp-synthetic
cd african-cp-synthetic

# 3. Copy files from this folder
cp /path/to/huggingface_upload/* .

# 4. Add files
git add .

# 5. Commit
git commit -m "Initial dataset upload"

# 6. Push
git push
```

### Option 3: Using Hugging Face CLI

```bash
# 1. Install huggingface_hub
pip install huggingface_hub

# 2. Login
huggingface-cli login

# 3. Upload folder
huggingface-cli upload [your-username]/african-cp-synthetic /path/to/huggingface_upload
```

---

## After Upload

### 1. Update README.md

Replace placeholders in README.md:
- `[your-username]` → Your actual Hugging Face username
- Add contributors/authors if applicable

### 2. Add Dataset Card Metadata (Top of README.md)

```yaml
---
license: cc-by-nc-4.0
task_categories:
- tabular-classification
- medical
task_ids:
- binary-classification
pretty_name: African Cerebral Palsy Synthetic Dataset
size_categories:
- 10K<n<100K
tags:
- cerebral-palsy
- medical
- healthcare
- africa
- synthetic-data
- pediatrics
---
```

### 3. Test Loading

```python
from datasets import load_dataset

# Load specific split
dataset = load_dataset(
    "[your-username]/african-cp-synthetic",
    data_files="africa_cp_train_5000.csv"
)

print(dataset)
```

### 4. Share!

Your dataset will be available at:
`https://huggingface.co/datasets/[your-username]/african-cp-synthetic`

---

## Recommended Dataset Settings

**Visibility**: Public (for research community benefit)

**Categories**:
- Medical
- Tabular Classification
- Healthcare
- Synthetic Data

**Tags**:
- cerebral-palsy
- africa
- pediatrics
- neurology
- early-detection
- low-resource-settings
- synthetic-data
- machine-learning

**License**: CC BY-NC 4.0 (Attribution-NonCommercial)

---

## Dataset Page Features to Enable

- ✅ **Dataset Viewer**: Auto-enabled for CSV files
- ✅ **Community**: Enable discussions for questions
- ✅ **Paper**: Link to any publications using this data
- ✅ **Citation**: Automatically generated from CITATION.bib

---

## Expected File Sizes

Verify these after upload:

```
africa_cp_train_1000.csv        137 KB
africa_cp_train_5000.csv        685 KB
africa_cp_train_10000.csv       1.4 MB
africa_cp_balanced_1000.csv     146 KB
africa_cp_preterm_2000.csv      277 KB
africa_cp_cases_only_500.csv     79 KB
africa_cp_test_2000.csv         275 KB
```

Total: ~3.1 MB

---

## Common Issues & Solutions

**Issue**: CSV files too large for web upload  
**Solution**: Use Git LFS or CLI upload method

**Issue**: README not displaying properly  
**Solution**: Ensure file is named exactly `README.md`

**Issue**: Dataset viewer not working  
**Solution**: Check CSV format, ensure proper headers

**Issue**: Missing dataset card metadata  
**Solution**: Add YAML frontmatter to README.md

---

## Questions?

- Hugging Face Docs: https://huggingface.co/docs/datasets
- Dataset Upload Guide: https://huggingface.co/docs/datasets/upload_dataset
- Community: https://discuss.huggingface.co/

---

## Checklist Before Upload

- [ ] All 7 CSV files present
- [ ] README.md is the dataset card
- [ ] LICENSE file included
- [ ] CITATION.bib present
- [ ] .gitattributes configured
- [ ] Replace `[your-username]` placeholders
- [ ] Test dataset locally first
- [ ] Choose appropriate license (CC BY-NC 4.0)
- [ ] Add yourself as author/contributor

---

**Ready to upload!** 🚀

This dataset will help researchers worldwide develop CP detection tools for African populations.