Enhancement: Add CalcKey Instance Support

[MrHamel]

CalcKey's API documentation can be found at https://calckey.social/api-doc

For both pulling in posts, but also updating timelines, etc.

[Teqed]

Additional information:

Calckey is downstream of Misskey, differences are noted.

• .well-known/nodeinfo response: Yes
• Nodeinfo path: {server}/nodeinfo/2.0 and {server}/nodeinfo/2.1
• Nodeinfo response: misskey or calckey
• Users URL pattern: {server}/@{user}
• "Posts" URL pattern: {server}/notes/{id}

Calckey's API documentation mentions notes/children and notes/conversation endpoints which can be POSTed to using a noteID as parsed from the URL. This appears to function similar to Mastodon's Context API and respectively fetch ancestors and descendants of a "note".

A CLI example: curl -X POST -H "Content-type: application/json" -H "Accept: application/json" -d '{ "noteId": "{id}" }' {server}/api/notes/children will return a JSON with the replies, including a URL that can be handed back to the local Mastodon server's search API for lookup.

For user profile backfilling, there is a user/notes endpoint but it requires a userID which does not parse from the user's name in the URL. It should be possible to make a request against the users/search endpoint and provide the parsed username as the query parameter, which will return the required userID.

Because the user URLs match the same pattern as Mastodon profiles, the choices here would be to either:

1. After failing against any Mastodon API, attempt Misskey/Calckey API.
2. Implement nodeinfo as discussed here to determine the appropriate API.

[nanos]

Thanks for the suggestion @MrHamel and thanks for the detailed report @Teqed !

For user profile backfilling, there is a user/notes endpoint but it requires a userID which does not parse from the user's name in the URL. It should be possible to make a request against the users/search endpoint and provide the parsed username as the query parameter, which will return the required userID.

That's the same with Mastodon, see

FediFetcher/find_posts.py

Line 216 in 213ef57

def get_user_id(server, user = None, access_token = None):

annoyingly a different path, though.

Because the user URLs match the same pattern as Mastodon profiles, the choices here would be to either:

1. After failing against any Mastodon API, attempt Misskey/Calckey API.
2. Implement nodeinfo as discussed here to determine the appropriate API.

I'm personally strongly in favour of the 2nd option here. Doing 2 requests per user seems wasteful.

I'm not currently having capacity to actually implement this, but hopefully later in the summer. If you have an itch, and the time, I'd be very happy, too, though 😉

[MrHamel]

Let's reopen this issue since the script still does not import posts into CalcKey/FireFish instances, thanks.

[nanos]

Apologies @MrHamel. Like @ToadKing i understood this issue to mean to offer support for pulling in content from Calckey/Firefish instances, not to actually support running FediFetcher against them.

I need to be honest though: I do not have a Firefish instance, so I’m unable to implement Firefish support myself. Unless someone else came along and implemented this, and was willing to commit to ongoing testing and support Firefish, this is unfortunately very unlikely to be implemented in FediFetcher.