Inventory Lookup

Queries a SQLite database to check product stock levels and reports whether inventory is sufficient or running low.

Full source

config:
    name: "Inventory Lookup"
    version: 1

services:
    DB is a database at "./inventory.sqlite"

workflow:
    trigger: when a stock check is requested

    step CheckProduct:
        get product using DB at "products" with id request.product-id
            save the result as product

        if product is empty:
            reject with "Product not found"

        log "Found product: {product.name}"

    step EvaluateStock:
        if product.stock is below 10:
            log "LOW STOCK WARNING: {product.name} has only {product.stock} units"
            complete with status "low-stock" and product-name product.name and stock product.stock

        otherwise:
            log "Stock level OK: {product.name} has {product.stock} units"
            complete with status "in-stock" and product-name product.name and stock product.stock

What this does

1. Declares a database service — DB is a database at "./inventory.sqlite" connects to a local SQLite file
2. Queries a single row — get product using DB at "products" auto-generates a SELECT * FROM products WHERE id = :id LIMIT 1
3. Handles missing data — if the query returns no rows, product is empty and the workflow rejects
4. Reads fields from the result — product.name and product.stock access columns from the returned row
5. Evaluates stock level — compares product.stock against a threshold using is below
6. Completes with status — returns either "low-stock" or "in-stock" along with the product details

Concepts demonstrated

• Database service type — is a database at "path" backed by SQLite via better-sqlite3
• Table mode queries — at "tablename" with with params auto-generates SQL (no raw SQL needed)
• Verb-to-operation mapping — get maps to SELECT ... LIMIT 1, returning a single record or empty
• Empty checks — if product is empty: handles the case where no database row matched
• Dot access on query results — product.name, product.stock read columns from the returned record
• Multiple outputs — complete with status ... and product-name ... and stock ...

Running it

# Test with mock services (no database needed)
flow run examples/inventory-lookup.flow --mock \
  --input '{"request": {"product-id": 1}}'

# Run with a real SQLite database
flow run examples/inventory-lookup.flow \
  --input '{"request": {"product-id": 42}}'

Database setup

For real execution, create the SQLite database first:

sqlite3 inventory.sqlite <<'SQL'
CREATE TABLE products (
    id INTEGER PRIMARY KEY,
    name TEXT NOT NULL,
    stock INTEGER DEFAULT 0
);
INSERT INTO products (name, stock) VALUES ('Widget', 100);
INSERT INTO products (name, stock) VALUES ('Gadget', 3);
INSERT INTO products (name, stock) VALUES ('Doohickey', 50);
SQL

As a webhook

flow serve examples/inventory-lookup.flow --mock

curl -X POST http://localhost:3000 \
  -H "Content-Type: application/json" \
  -d '{"request": {"product-id": 1}}'