Closed LudovicMalot closed 1 week ago
What kind of data would we like for this?
As I see it, there are at least 3 different source models.
Unrelated:
Regarding the Hoyo API, as it's accessible with no authentication (public), it "should" be fine. I noticed that for all characters as well there's an API.
GET
https://sg-wiki-api-static.hoyolab.com/hoyowiki/genshin/wapi/entry_page?entry_page_id=4577entry_page_id
can be found there: POST
https://sg-wiki-api.hoyolab.com/hoyowiki/genshin/wapi/get_entry_page_list
curl 'https://sg-wiki-api.hoyolab.com/hoyowiki/genshin/wapi/get_entry_page_list' \
-H 'Referer: https://wiki.hoyolab.com/' \
--data-raw '{"filters":[],"menu_id":"2","page_num":1,"page_size":30,"use_es":true}' \
--compressed
There's also the type of events that we want to collect. Only the ones that actually appear in-game?
All the community events? How about the web events?
Is the goal to create a timeline ( https://paimon.moe/timeline )?
In this case, we can get a very limited subset of events (they usually reuse keywords).
All the community events? How about the web events?
Is the goal to create a timeline ( https://paimon.moe/timeline )?
In this case, we can get a minimal subset of events (they usually reuse keywords).
As we are the data team I would suggest we save the most data possible but with proper classification for web events, in-game events etc... I think the Hoyolab API should be fine for all types of events. No need to scrap Honey or the Wiki. Especially with the language issue.
To have the history of the event, we probably will need to parse (or manually) the wiki. It's possible that some GitHub repos already store those in a friendlier format. But this is not important and seems like a lot of work for not necessarily a significant impact. If it's too time-consuming, better to raise an issue with ultra-low priority for this task.
To be noted that I cannot assign myself, I don't have the permissions.
To be noted that I cannot assign myself, I don't have the permissions
should be fixed
Still nope :)
Create and fill a JSON with all data following the event types