dataface-playground 0.0.2.dev1509

Dataface interactive playground — OSS dashboard editor

Dataface Playground

Interactive editor for Dataface dashboards — live YAML editing with a real-time preview, share links, and an AI copilot for dashboard authoring.

Install

pip install dataface-playground

The package installs both the dft-playground console script and the dft playground subcommand on the dft CLI.

Run

dft-playground            # auto-resolved port
dft playground            # equivalent via the dft CLI
dft playground --port 3000
dft playground --base-dir ./examples
dft playground --data ./my_data.db
dft playground --openai-key sk-...    # enables AI copilot

The playground opens at http://127.0.0.1:<port> and ships a curated gallery of example dashboards covering charts, layouts, tables, KPIs, and variable-driven interactions.

What's in the wheel

• FastAPI app: live YAML editor, WebSocket preview, share-link cache, AI copilot (/api/ai/chat — OpenAI-backed when OPENAI_API_KEY is set)
• A small curated example gallery under examples_data/examples/ (DuckDB-backed) so a fresh install renders something on first launch

Configuration

Variable	Default	Description
OPENAI_API_KEY	—	Enables the AI copilot sidebar
DFT_PLAYGROUND_MODE	local (set by CLI)	local allows inline data sources; deployed requires an allowlisted source
DFT_DOCS_URL	—	Base URL for per-face docs: links rendered in the gallery

Known limitations

Share cache is in-memory. Shared dashboard links are invalidated when the server restarts. The playground is meant for local editing and short-lived demos, not as a persistent share-link service.

License

Apache-2.0 — see LICENSE.