Preventing duplicate server writes during PyScript network retries

Idempotency in PyScript Network Requests

When implementing retry logic for HTTP POST requests in PyScript (using pyodide.http.pyfetch or the js module FFI), there is a risk of duplicating server-side writes. Because PyScript operates within the browser's asynchronous environment, a request may time out or return a network error after the server has already processed the data but before the client receives the confirmation.

Since PyScript does not provide a built-in idempotency layer or automatic request deduplication, the responsibility for ensuring "at-most-once" delivery falls to the developer. This requires a coordination strategy between the Python code running in WebAssembly and the backend API.

What is the most reliable pattern for implementing client-side idempotency keys within a PyScript async loop to ensure that retries triggered by transient 5xx errors or rejected promises do not result in duplicate records?