README.md

metadata

title: ThoughtSpot Demo Builder
emoji: 🚀
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.44.1
app_file: app.py
pinned: false
license: mit
python_version: '3.11'

DemoPrep — AI-Powered ThoughtSpot Demo Builder

A Gradio application that generates and deploys complete ThoughtSpot demo environments end to end — company research, an LLM-authored data blueprint, deterministic data generation, a Snowflake schema + data load, and a ThoughtSpot semantic model and liveboard.

Deployed on Hugging Face Spaces. app.py is the entry point. See DEPLOYMENT.md for setup.

🖥️ The interface

The app is a tabbed Gradio UI, and the App tab is the primary way to build a demo: fill in a short form and press GO, and the full pipeline runs automatically with live progress.

Naming note: the whole application lives in chat_interface.py — that name is historical. It is not just a chat interface; that single file hosts every tab (App, Chat, Admin, Settings, Run History, …), and the App tab is the main path. A Chat tab offers an alternative conversational flow that runs the same pipeline. app.py simply configures and launches this app.

🚀 Features

• AI-powered research — researches the company and industry with an LLM
• Single blueprint pipeline — one path, no keyword routing or fallback: research → LLM-authored DemoBlueprint → deterministic engine → validation → derived DDL → load
• Realistic data — bounded/related measures (a part never exceeds its whole, rates stay ≤ 1, prices are stable per entity) plus planted, discoverable "Demo to Win" insights the validator proves are visible
• ThoughtSpot integration — deploys the connection, tables, semantic model, and an enhanced liveboard
• Interactive UI — tabbed Gradio app with real-time progress

🛠️ Tech Stack

• Frontend: Gradio (Python web UI)
• Backend: Python 3.11
• Database: Snowflake (keypair auth)
• Analytics: ThoughtSpot (per-environment trusted auth)
• AI: multiple LLM providers (OpenAI and Anthropic/Claude), routed via llm_config.py
• Settings: Supabase (admin + per-user settings)

📋 Prerequisites

• Python 3.11
• Snowflake account (keypair configured)
• ThoughtSpot Cloud account
• An LLM API key (OpenAI and/or Anthropic)
• Supabase project (settings storage)

🚀 Quick Start

1. Clone the repository

   git clone <repo-url>
   cd demoprep
   

2. Set up a virtual environment

   python -m venv .venv
   source .venv/bin/activate   # Windows: .venv\Scripts\activate
   

3. Install dependencies

   pip install -r requirements.txt
   

4. Configure bootstrap environment variables (see Configuration)

   cp .env.example .env
   # edit .env
   

5. Run the application

   python app.py
   

   app.py launches the full Gradio app defined in chat_interface.py.

6. Open your browser Navigate to http://localhost:7860 and use the App tab.

⚙️ Configuration

.env holds only the bootstrap secrets needed to start the app and reach Supabase and the LLM providers. All other credentials — the Snowflake keypair and the ThoughtSpot per-environment trusted-auth keys — live in Supabase admin settings and are loaded at runtime.

# Supabase (settings storage)
SUPABASE_URL=...
SUPABASE_ANON_KEY=...

# LLM providers
OPENAI_API_KEY=...
GOOGLE_API_KEY=...            # optional (Gemini)

# Slack deployment notifications (optional, outbound-only)
SLACK_BOT_TOKEN=xoxb-...
SLACK_DEPLOYMENT_CHANNEL_ID=C0123456789

Slack notifications use the Slack Web API to post deployment status into one approved channel. This path is outbound-only: no Socket Mode, event subscriptions, slash commands, or public request URL. The Slack app needs only the chat:write bot scope and must be invited to the target channel.

🎯 Usage

The App tab is the primary interface — fill the form, press GO, and the pipeline runs end to end:

1. Fill the form — vertical / line of business / function (or a custom use case), the company URL, and the target ThoughtSpot environment
2. Press GO — research → blueprint → data generation → validation → Snowflake load → ThoughtSpot model + liveboard
3. Review — model and liveboard links, a Demo Pack, and a Spotter Viz story appear on completion

📁 Project Structure

demoprep/
├── app.py                    # Entry point — launches the Gradio app on :7860
├── chat_interface.py         # The full Gradio app (App/Chat/Admin/Settings/…) — historical name
├── thoughtspot_deployer.py   # ThoughtSpot deploy: connection, tables, model, liveboard
├── liveboard_creator.py      # MCP liveboard creation + TML post-processing (enhance_mcp_liveboard)
├── snowflake_auth.py         # Snowflake keypair authentication
├── llm_config.py             # LLM provider/model routing (single source of truth)
├── demo_personas.py          # Vertical × Function use-case configs
├── demoprep_app/             # The demo-generation pipeline package
│   ├── pipeline/build_demo.py    #   the single build entry point
│   ├── scenario/                 #   blueprint contract + LLM authoring + directives
│   ├── dataset/                  #   deterministic engine + validator
│   ├── ddl/                      #   Snowflake DDL derived from the dataset
│   └── integrations/snowflake/   #   Snowflake row loader
├── requirements.txt          # Python dependencies
├── docs/                     # Documentation
├── tests/                    # Unit tests + the e2e quality harness (e2e_quality.py)
└── results/                  # Generated demo results

🧪 Testing

# Unit tests (pytest collects test_*.py)
python -m pytest tests/

# End-to-end quality harness — drives a running app via the browser and grades the output
python tests/e2e_quality.py --env-name test

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• ThoughtSpot for the analytics platform
• Snowflake for the data warehouse
• OpenAI and Anthropic for the AI capabilities
• Gradio for the web interface

📚 Development Notes

Sprint planning and working notes live in dev_notes/ (gitignored); architecture and handoff docs are in docs/.

📞 Support

For support, create an issue in this repository.

---

Built with ❤️ for the ThoughtSpot community