feat: add Confluence source permission metadata

[shreyanid]

Retrieves permission data from each Confluence document and its parent space, normalizes the format (see below), and adds it to the permissions_data field in the metadata.

The output format is a dictionary of roles (read, update, delete) with a lists of "groups" and users" having that permission under each.

"permissions_data": [{
     "read": {
       "groups": [...,],
       "users": [...,], 
     },
     "update": {
       "groups": [...,],
       "users": [...,], 
     },
     "delete": {
       "groups": [...,],
       "users": [...,], 
     },
 }]

The output type is currently wrapped in a list to match the existing FileDataSourceMetadata type definition shared by all connectors. TODO to update that as the other connectors get normalized

The current implementation limits the results to approximately 250 permission entities.

Further improvement items:

• expand users and groups (apply group permissions as enabled for user of the group)
• retrieve email based on ID to populate the email_address field
• enforce the normalized format via the type of permissions_data in FileDataSourceMetadata (instead of a generic list[dict])

This PR also clarifies "space_id" (int) vs "space_key" (str). The previous space_id used throughout the code was actually the space key. Does not change the user facing experience (the space_id field returned in metadata is still the space key as before)

[vangheem]

You refactored this to no longer use requests. I guess it's fine to still use requests and put the import inline.

[shreyanid]

either works right? I went ahead with catching a generic exception since we really don't want this to causing jobs to err, so ill remove this dep