Function bootstrap

sdk/python/feast/templates/pytorch_nlp/bootstrap.py:249–301

Bootstrap the pytorch_nlp template with sample data.

()

def bootstrap():
    """Bootstrap the pytorch_nlp template with sample data."""
    repo_path = pathlib.Path(__file__).parent.absolute() / "feature_repo"
    raw_project_name = pathlib.Path(__file__).parent.absolute().name

    # Sanitize project name for SQLite compatibility (no hyphens allowed)
    project_name = raw_project_name.replace("-", "_")
    if project_name != raw_project_name:
        print(f"   ℹ️  Project name sanitized: '{raw_project_name}' → '{project_name}'")
        print("   💡 SQLite table names cannot contain hyphens")

    data_path = repo_path / "data"
    data_path.mkdir(exist_ok=True)

    print("🎭 Setting up sentiment analysis data for PyTorch NLP demonstration...")

    parquet_file = data_path / "sentiment_data.parquet"

    # Generate sentiment data
    print("   📝 Generating synthetic sentiment analysis dataset...")
    df = create_sentiment_data(num_samples=1000)

    # Save to parquet
    table = pa.Table.from_pandas(df)
    pq.write_table(table, parquet_file)

    print(f"   ✅ Created sentiment dataset with {len(df)} samples")
    print("   📊 Sentiment distribution:")
    sentiment_counts = df["sentiment_label"].value_counts()
    for sentiment, count in sentiment_counts.items():
        print(f"      - {sentiment.capitalize()}: {count} samples")

    # Replace template placeholders
    example_py_file = repo_path / "example_repo.py"
    replace_str_in_file(example_py_file, "%PROJECT_NAME%", str(project_name))

    test_workflow_file = repo_path / "test_workflow.py"
    replace_str_in_file(test_workflow_file, "%PROJECT_NAME%", str(project_name))

    print("🚀 PyTorch NLP template initialized successfully!")

    print("\n🎯 To get started:")
    print(f"  1. cd {project_name}")
    print("  2. pip install -r requirements.txt")
    print("  3. cd feature_repo")
    print("  4. feast apply")
    print("  5. feast materialize")
    print("  6. python test_workflow.py")
    print("\n💡 This template demonstrates:")
    print("  - Text feature engineering with Feast")
    print("  - PyTorch + Hugging Face transformers integration")
    print("  - Sentiment analysis with pre-trained models")
    print("  - Online and offline feature serving")


if __name__ == "__main__":