How to communicate between a NodeJS graphql-apollo server and a Python project?

Question

Since this stack is new to me I'd figure it wouldnt hurt to ask the community for input.

I'm exploring a new type of project and have been brainstorming how/ if it's possible to communicate between an apollo-graphql server (NodeJS) and a client Python program. I have a web development background and I intend to keep my app and server in Typescript/Node.

However...

As a project and exercise, I would like to play around with connecting a watering system via the GPIO package in Python + RaspberryPi and then send the updates to my apollo-graphql server. I think conceptually, in a traditional RESTful api it would be straightforward... However, with GraphQL things are a bit more nebulous to me. It would be nice to build out hardware connected services in various languages as appropriate and not be bottled into having everything in typescript/node and to that end Python is the most widely used language within the RaspberryPi community.

Has anyone done this before, or something similar, and have insights and experience to share?

---

edit : since posting I became more familiar with graphql and played around with it in Postman. It simply has an endpoint that I can post to it with queries (and I assume mutations and everything else) and it's captured by my NodeJS apollo-graphQL server. I ended up using this python package to query against my node API and it seems to work well.

Answer 1

Your server is node and your client is python. The server doesn't care at all whether your client is python or java or C#, it only wants to see GraphQL requests.

I suggest using a python-based GraphQL client package such as python-graphql-client

You're right that in the end everything looks like a POST. Using an actual client might make life a little bit easier for you though.

Answer 2

You can try calling your GraphQL service using a RESTful wrapper: https://graphql.org/blog/rest-api-graphql-wrapper/ . There's a similar question asked before: link .

Secondly, instead of calling the server directly, you could also try using a message broker in the middle. In short: the raspi sends messages to a queue that is polled by your server. There are several solutions available like: Rabbitmq .