First 15 minutes (no Spark)

Install tablespec, author one split UMF, validate it, compile artifacts, and open a guidebook — without Spark or a Databricks workspace.

1. Install

uv add tablespec --index-url https://documentdrivendx.github.io/tablespec/simple/
# or: pip install tablespec --index-url https://documentdrivendx.github.io/tablespec/simple/

2. Author a minimal split UMF

Use the Getting Started authoring example, or any split directory with table.yaml + columns/*.yaml (and provenance columns if you will run full pipeline validate).

For a quick try from the repo fixtures after cloning tablespec:

# from a tablespec checkout
ls tests/e2e/fixtures/member.umf.yaml

3. Validate

tablespec validate tables/medical_claims/
# or a fixture:
tablespec validate tests/e2e/fixtures/member.umf.yaml

4. Compile artifacts (Path B one-shot)

tablespec bootstrap tables/medical_claims/ -o build/artifacts --dialect duckdb
# Databricks-facing cast SQL spelling (normalizes to Spark-family SQL):
# tablespec bootstrap tables/medical_claims/ -o build/artifacts --dialect databricks

This is the public CLI for bootstrap_from_specs: load UMF → write the committed artifact tree (ingest SQL, schemas, suites, dbt/LDP projects as applicable).

Inspect build/artifacts/ for generated files.

5. Generate single-format outputs (optional)

tablespec generate tables/medical_claims/ -f sql
tablespec generate tables/medical_claims/ -f ingest
tablespec emit tables/medical_claims/ out/dbt --backend dbt --dialect duckdb

6. Guidebook

tablespec guidebook tables/ -o out/guidebook
# open out/guidebook/index.html in a browser

Empty directories fail with an actionable error (exit 1).

What you proved

Step	Result
Install	Package from the Pages index
Validate	UMF structure accepted
Bootstrap	Deterministic artifact tree under `-o`
Guidebook	Browsable HTML for the same UMF set

Next: In a workspace when you have Databricks; Deploy the app for the profiling UI.