Quickstart

Run a graph query against a small Parquet dataset, then explore the results in the browser workbench.

On this page

1. Generate the demo data

Install Git, Rust 1.95 or newer, and a C/C++ toolchain. The workbench step also requires Node.js 24. The built-in demo writes three accounts, four transfers, and a graph mapping to a temporary catalog.

Shell
git clone --branch v1.1.1 --depth 1 https://github.com/SamuelSupe/graphlake.git
cd graphlake
demo_root=$(mktemp -d)
demo_dir="$demo_root/catalog"
cargo run --locked --example demo -- "$demo_dir"

2. Run a Cypher query

In the same shell, find destinations two transfers away from account 10 in tenant 1. The amount sums the last transfer on each matched path.

Shell
cargo run --locked -- --catalog "$demo_dir/graph.yaml" query \
  'MATCH (a:Account)-[:Transfer]->(b:Account)-[t:Transfer]->(c:Account)
WHERE a.tenant_id = 1 AND a.account_id = 10
RETURN c.account_id AS account, count(*) AS paths, sum(t.amount) AS amount'

Expected result Account 30, 2 paths, total amount 600.

3. Open the workbench

Build the browser UI, set the local admin password, and start the service. The catalog's valid model is adopted on startup.

Shell
npm --prefix web ci
npm --prefix web run build
cargo run --locked --features admin --bin graphlake-admin -- \
  --catalog-dir "$demo_dir" auth init
cargo run --locked --features admin --bin graphlake-admin -- \
  --catalog-dir "$demo_dir" --assets-dir web/dist

Open http://127.0.0.1:7337 and sign in as admin. In Graph Models (图模型), open graph and go to Query (查询).

4. Explore the transfers

Choose Cypher, run the query below, then open the Graph (图) result tab. Select a node or relationship to inspect its properties.

Cypher
MATCH p = (:Account)-[:Transfer]->(:Account)
RETURN p
GraphLake workbench showing three accounts linked by transfers from the Parquet demo
The transfer graph in the GraphLake workbench.

Ready for your own data? Connect a source and publish a graph model.