Publish a second end-to-end example (non-jaffle)

[KKranthi6881]

Problem

The only zero-setup, runnable example is examples/jaffle_shop_demo/. That's a great on-ramp for dbt users, but it under-sells DataLex's broader surface — contracts, cross-repo packages, policy, glossary, multi-dialect emission.

Goal

Add a second end-to-end example that demonstrates something the jaffle demo doesn't. Candidate shapes (pick one):

• examples/ecommerce_core/ — a medium-size DataLex project (50 entities) with: contracts enabled, a .datalex/snippets/audit_columns.yaml applied across entities, a policies/require_owner.yaml, a glossary of 10 terms, and two dialect outputs (Postgres + Snowflake). No dbt.
• examples/cross_repo_shared_dims/ — one DataLex project that imports a second one via imports:, showing datalex packages resolve end-to-end with a lockfile.
• examples/governance_starter/ — a project with sensitivity-tagged columns, policies, and a pre-built datalex emit erd Markdown output so a visitor can open it on GitHub.

Each example must include:

• A README.md with 5-minute instructions.
• A setup.py (or equivalent) that runs end-to-end with zero external credentials.
• A link from the main README's example index.

Acceptance criteria

• python examples/<new>/setup.py succeeds on a clean clone after pip install -e '.[duckdb]'.
• The resulting DataLex project validates under datalex validate.
• Pattern matches the existing examples/jaffle_shop_demo/ layout and tone.

Why this is a good first issue

You'll read (not write) the core engine, which is a great way to learn the surface. The scope is self-contained and the acceptance criteria are concrete.

[snakefood3232]

The jaffle demo alone under-sells DataLex's broader surface like contracts and multi-dialect emission. I'd handle this by building the ecommerce_core example with Postgres/Snowflake outputs, contracts, and a glossary to demonstrate these features. Can have a PR up within a day—I've built similar production data catalogs with automated governance tooling.