Few suggestions to improve CX

Hey Guys, Thanks for creating this wonderful extension. I'm trying to integrate this extension to one of our existing databases to be able to perform similarity search and clustering. While I was setting up this extension in my test database, I faced few issues and thought of sharing them and a few suggestions to improve the Dev Experience.

API Key

I'm not entirely sure why (maybe because I'm using macOS) but ALTER DATABASE and ALTER SYSTEM didn't set the Pinecone API Key.

ALTER DATABASE succeeded but didn't actually set the value so running CREATE INDEX failed.
ALTER SYSTEM failed with unrecognized configuration parameter "pinecone.api_key" error
I was finally able to get it working using SET pinecone.api_key To 'xxxx' and verifying with SHOW pinecone.api_key

It would be great if all these options were available in README and also how to verify that api_key was set.

Handle `https://` in host for CREATE INDEX

I initially ran the following command to create an index, it failed with error curl_easy_perform(): Timeout was reached. The error was not intuitive. Removing https:// worked, having an appropriate error or automatically handing this would be great.

CREATE INDEX my_remote_index 
ON products 
USING pinecone (embedding) 
WITH (host = 'https://xxx.svc.aped-4627-b74a.pinecone.io');

Option to migrate from `pgvector` to `pgvector-remote`

Having this option would be great in the long run. I already have a database that uses pgvector. I tried to run ALTER EXTENSION UPDATE to use pgvector-remote in the existing database, postgres threw and exception

extension "vector" has no update path from version "0.6.2" to version "remote0.1.0"

It would be great allowing existing users for pgvector to seamlessly transition to pgvector-remote without having to create a new database and spend hours to migrate from older to newer database.

nit: Installation Instructions

This is a nit pick but the installation instructions were confusing at first glance. I wasn't sure if I had to also running the commands for Milvus or not. I elected to skip them and was able to use pinecone without any issues. Having simple steps to follow would make it easier for dev trying to test. And it would also be great to have instructions for Mac.

Hello Hemanth,

Thank you for your kind words about our extension and for choosing to integrate it into your project. We truly appreciate the time you took to share the challenges you encountered during setup, as well as your suggestions for improvement.

We are actively developing and enhancing our extension, and your detailed feedback is invaluable in this process. We aim to clarify our documentation to ensure a smoother setup experience and are particularly focused on making it easier for developers to test our tools. Additionally, we acknowledge the importance of facilitating a seamless transition from pgvector to pgvector-remote and are committed to implementing this feature in our upcoming release which supports Pinecone and Milvus.

Thank you once again for your feedback. We will keep you updated on our progress.

georgia-tech-db / pgvector-remote