We will be uploading parquet and pickle files to S3, as a start. Later we will add actual vector databases to enable visualizations and so on.
When we upload a file, we should be able to find it globally, so we can list it for download. We should also associate it with a user, so they can see their own file uploads.
The users table should have a foreign key to a collections table. Or rather, a one to many relationship. That way you can create several collections.
We will be uploading parquet and pickle files to S3, as a start. Later we will add actual vector databases to enable visualizations and so on.
When we upload a file, we should be able to find it globally, so we can list it for download. We should also associate it with a user, so they can see their own file uploads.
The users table should have a foreign key to a collections table. Or rather, a one to many relationship. That way you can create several collections.
The collection can have all the object metadata.