I’d like to make the large-scale analysis of the games. The answers I’d like to find are:
what is the difference between given level and another in terms of points loss
how does the user compare to other players e.g. should one practice midgame/joseki/fuseki, etc.
This knowledge would be helpful for me to understand the nature of a game, set the goals etc.
Basically the public data would be already a lot of information, especially the standard performance trend plotted for every game. Obviously more AI metadata (average points lost) would be also helpful.
I tried to access OGS API and even with token it looks like I can really download the game, and further reviews.
Can anyone confirm/deny if I should be able to get the data I’m looking for via API?
I think getting AI information should already be possible on a per-game basis, because the web client can get that information too. So checking out the (open) source code for the web client could tell you how to do that.
It would be all much easier to understand for me, if I could successfully parse endpoints. Maybe I can get help here?
This is the Python code I have. After obtaining the token correctly, this code is returning me a nice parsed dictionary with all my games:
url = "https://online-go.com/api/v1/me/games"
games = requests.get(url, headers={"Authorization": "Bearer " + token})
games = games.json()
However, when I’m hitting other endpoints, for example http://online-go.com/v1/games/, I am getting some html page in the games.content that starts with:
'<!doctype html><html lang="en"><head><meta charset="utf-8"><title>Play Go at online-go.com! | OGS</title>
So it looks like I can just access the /me information. Anyone knows the reason?
Hello, thanks for the response! When I download the sgf from the link you included, I can see there is no other metadata than just a game. The only mention of AI is:
“C[AI Review by KataGo strength 12000 using the 40x256 network b40c256-s11101799168-d2715431527.”
