Thanks for putting together this adapter! Noticed that you recommended setting the dataset to private in order to make the user documents not public. This PR introduces another approach where dataset visibility doesn't matter because documents stored on a path won't be accessible through the public API (for example user.214d0e86-417d-48fb-9d89-fbe3d7da87d9).
This means that developers can leave their datasets public and get access to query CDNs and all that good stuff, without leaking the user documents.
Hi!
Thanks for putting together this adapter! Noticed that you recommended setting the dataset to private in order to make the user documents not public. This PR introduces another approach where dataset visibility doesn't matter because documents stored on a path won't be accessible through the public API (for example
user.214d0e86-417d-48fb-9d89-fbe3d7da87d9
).This means that developers can leave their datasets public and get access to query CDNs and all that good stuff, without leaking the user documents.