Sales history

Volume-weighted average sale price per item over the last 30 days, in cents, in one response with no pagination. This is the number to price a lister against: `average_price` is `sum(price * amount) / sum(amount)` across every sale of that item in the window, `sales` is how many separate sales made it up and `volume` how many copies changed hands. Thin items move on very few sales, so check `sales` before trusting an average. Refunded and failed orders are excluded, which is the one way this differs from `GET /public/v1/sales`. Optional `app_id` filter, worth passing since it cuts the payload to the game you trade. The whole set is built once and cached for 10 minutes; responses carry an `ETag`, so send it back as `If-None-Match` and you get a `304` when nothing has changed. `generated_at` is when the cached copy was built.