Checking for existence of key in Collections object causes an infinite stall that consumes all memory

[martinm07]

Description

When trying to check for the existence of a Collection in client.collections using "collection_name" in client.collections, the program stalls seemingly forever, consuming all the available RAM relatively quickly, regardless of if the collection does indeed exist or not.

Steps to reproduce

import typesense

client = typesense.Client(
    {
        "nodes": [{"host": "localhost", "port": "8108", "protocol": "http"}],
        "api_key": "yourapikey",
        "connection_timeout_seconds": 2,
    }
)

if "documentname" in client.collections:
    print("'documentname' collection exists")

My particular environment is on Windows 10, using Docker Desktop with Docker Engine version 26.1.4 and the typesense container created using

docker run -p "8108:8108" -v "${pwd}/typesense-data:/data" typesense/typesense:26.0 --data-dir "/data" --api-key "$env:TYPESENSE_API_KEY" --enable-cors

on Powershell.

Expected Behavior

Either for it return if the collection exists or not, or to error gracefully.

Actual Behavior

The program stalls forever, consuming more and more RAM until the script is interrupted and the Docker container shut down.

Metadata

Typesense Version: 26.0
typesense-python Version: 0.21.0

OS: Windows 10 x64

[tharropoulos]

The client.collections object is an instance of the Collections class. If you're trying to access the dictionary of collections inside that, it would actually be available through client.collections.collections. Still, that's in-memory, not the actual collections found on the Typesense server, so even then, this isn't particularly useful. Might I know what the use-case for this is?

The reason for the infinite loop is that, if you're trying to access the "documentname" with the in keyword, Python will try to use the __contains__ method, but since that doesn't exist, it will fall back to the to __getitem__ method. Since that will return a new Collection object if it doesn't exist (which it won't, since you haven't initialized it yet). Then, it recursively and infinitely creates new Objects. This use-case isn't something that should normally be done while using the Typesense API

[martinm07]

@tharropoulos Quite simply, I wanted to check if a collection existed or not, so that I may delete it and recreate it. This was for a CLI command I was making (for dev purposes).
"documentname" in client.collections seemed a very natural way of going about that, so I was rather surprised to see my entire computer grind to a halt because of it. I ended up creating the following helper function to circumvent this:

def collection_exists(client, collection_name: str):
    return collection_name in [
        schema["name"] for schema in client.collections.retrieve()
    ]

If I missed a more obvious method, I'd love to hear about it.
Otherwise, I think it'd be good to at least replace this with a less... destructive, behavior— for when others try do the same thing I did.
Anyway, thanks for explaining what's going on!

[tharropoulos]

@jasonbosco should we override the __contains__ dunder here to avoid this behavior?