Requests/sec

[xeacott]

I wanted to open an issue to quickly ask a question to the developers as I am coming from nba_py. I wish to use this package in a GUI that is actually turning into a class project ... however I built the GUI previously with nba_py adhering to an MVC architecture. The model I used would pull information on the fly and store whatever data the user wished. For example, the session would start clean (or could import previous player data). Upon the latter, if a user has 12 players loaded in a session looking at say 20 different categories, I would load all the information at startup time. (Through sending requests for each player.)

I know this may present an issue as I would be asking for lots of requests from stats.nba.com and I know they have been limiting requests/sec. Do the developers have preferences or suggestions on allowing/disallowing this behavior?

edit: typos, grammer, etc.

[swar]

I'm sure people will have varying answers to this question. Personally, I wouldn't want to flood any website with many requests. I don't want to abuse something that's publicly available, but I do think there are ways around it while maintaining respect for the website. You could make sure you're pausing between requests and caching the information so that you don't have to make multiple requests for the same information.

Depending on how many requests you plan on submitting, just be mindful that they can ban your IP or even a group of IPs if they are all coming from the same location, such as a university.

[xeacott]

Understood - Beyond the scope of the 'requirements' i have to fill, I will probably build around this so it doesn't become an issue. Thank you for the response.

[jaypeedevlin]

Jumping into this topic, one of the things I loved about nba_py when using it was that it used requests-cache which meant that it avoiding hitting the endpoints repeatedly when you were optimizing code etc.

I wonder if that might be an easy drop-in which would mitigate some of the cases (and be generally more respectful to the underlying API)

[xeacott]

I do agree with Jay, respecting public APIs is important and hopefully that is something that can be implemented. There ought to be a package that is the gold standard...

[swar]

I am going to add an issue per @jaypeedevlin so that we can look into adding in requests-cache functionality as that'll be a good idea and a replacement to DEBUG_STORAGE in debug.py

[xeacott]

Also, this thread may be closed at your discretion @swar

[swar]

@xeacott I'll leave it open until the end of the month in case anyone else has anything to add. Thank you!