You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Even though the technology is well known and supported by browsers it seems to be rarely used in backend-2-backend communication.
With the boom of chatGPT this become a case as it is supporting streaming using Server-sent events: https://platform.openai.com/docs/api-reference/chat/create
Suggested implementation
Currently payload is available as a text when the request is completed.
My implementation suggestion would be to pass eventEmitter as a parameter to resource method and emit events from inside mappersmith on every new chunk of data.
I think it looks very interesting and sounds like a fun project 👍
Since gateways are configurable I guess ideally it should be implemented in both HTTP and Fetch gateways, even though HTTP is the default it is not uncommon to override it to Fetch for node:
import forge, { configs } from 'mappersmith';
import fetchGateway from 'mappersmith/gateway/fetch';
import fetch from 'node-fetch';
configs.gateway = fetchGateway;
configs.fetch = fetch;
const client = forge({ ... })
Server-sent events client implementation
Purpose
I want to bump with the community and understand if this feature would be interesting and useful in scope of mappersmith.
Docs
https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events
Use case
Even though the technology is well known and supported by browsers it seems to be rarely used in backend-2-backend communication.
With the boom of chatGPT this become a case as it is supporting streaming using Server-sent events: https://platform.openai.com/docs/api-reference/chat/create
Suggested implementation
Currently payload is available as a text when the request is completed.
My implementation suggestion would be to pass eventEmitter as a parameter to resource method and emit events from inside mappersmith on every new chunk of data.
The text was updated successfully, but these errors were encountered: