docs/liveboard_lab.md

# Liveboard Lab

The Liveboard Lab is an internal workflow for iterating on liveboard quality
from existing ThoughtSpot model IDs. It preserves the supported architecture:

1. MCP creates answers and the initial liveboard.
2. TML post-processing polishes the same liveboard.
3. The lab opts into the experimental `golden_two_tab` layout strategy.

The normal app path is not changed by the lab. Production enhancement still
defaults to the existing single-tab layout unless `layout_strategy` is passed.

## Plan Only

```bash
python3 tools/liveboard_lab.py plan \
  --company Nike \
  --use-case "Retail Sales" \
  --target-viz-count 16
```

This prints the dashboard plan without importing Snowflake or connecting to
ThoughtSpot.

## Create From Existing Model

```bash
python3 tools/liveboard_lab.py create \
  --model-id "<thoughtspot-model-guid>" \
  --model-name "<thoughtspot-model-name>" \
  --company Nike \
  --use-case "Retail Sales" \
  --target-viz-count 16
```

Required ThoughtSpot settings can be passed as CLI flags or environment
variables:

- `THOUGHTSPOT_URL`
- `THOUGHTSPOT_USERNAME`
- `THOUGHTSPOT_TRUSTED_AUTH_KEY`

For older DemoPrep `.env` files that use `TS_ENV_1_URL`, `TS_ENV_1_KEY_VAR`,
and `TEST_USER`, pass:

```bash
python3 tools/liveboard_lab.py --env-file /path/to/.env --ts-env-index 1 create ...
```

## Re-Enhance Existing Liveboard

```bash
python3 tools/liveboard_lab.py enhance-existing \
  --liveboard-guid "<thoughtspot-liveboard-guid>" \
  --company Nike \
  --use-case "Retail Sales"
```

Use this when MCP already created a liveboard and the iteration is only on TML
layout/style.

## Current Golden Target

The reference is `goldendemo/Vizio.com.liveboard.tml`.

The lab target is smaller than the reference:

- 14-20 visualizations instead of 40+
- 2 tabs: `Executive Overview`, `Operational Detail`
- Thematic groups inspired by the reference
- Curved borders, group colors, highlighted KPI tiles