>>> import srcomapi
>>> api=srcomapi.SpeedrunCom()
>>> api.get_games()

results in an error, same with api.get_game

Hi!
I'm kind of new to REST APIs but I noticed that sometimes not all data is returned in one http request, because max value is set to 20. So there are problems if for example a category has over 20 levels.
An easy fix is to set a max parameter above 20 like this:
https://www.speedrun.com/api/v1/categories/7dg5vw72/records?max=100

Hello,

I am trying to gather data for a statistical modeling project. However, I can seem to figure out how to extract all the run times, dates, users and category (any% , etc) from one single game. Could you provide an examples how to do it?

Might be a very basic question but I am not that familiar with python syntax.

Thank you,

Tiago

How are obsolete runs stored?

Originally posted by @SDuranNebreda in #11 (comment)

Title ^

If you try to install this package, setup doesn't copy .version

To fix it just add package_data={'': ['.version']} to the setup method

Currently there is only a way to get all games or a specific game by its id or abbreviation, but games should be searchable by name and a lot of other parameters listed in the Speedrun.com API-documentation. Same with users.

Executing the following code

api = srcomapi.SpeedrunCom()
g = api.get_game('sms')
dir(g)

shows that g has the attribute 'release-date'.

g.release-date

fails, because Python interpretes the '-' as an operator for g.release and date, which causes a AttributeError.
It's still possible to get it with g.__getattr__('release-date') but I don't think this is the way it should work.

Hi! Thank you so much for writing this tool, I've found it very helpful.

I wanted to ask if there is a direct way of getting the data for WR progressions, as in the gamestat page of each game (example).

Concretely, I would like to end up with a list of dictionary entries, where each dictionary is of the form

{
'game' : 'Super Mario Sunshine', 
'category' : 'any%', 
'time' : <primary_t>,
'date'; <date>,
'runner' : '<player.id(s)>',
}

I tried downloading all runs with the intention of filtering them myself later to only include world record improvements, as this:

import srcomapi, srcomapi.datatypes as dt

api = srcomapi.SpeedrunCom(); api.debug = 1
search_results = api.search(srcomapi.datatypes.Game, {"name": "super mario sunshine"})
game = search_results[0]
category = game.categories[0]
leaderboard = dt.Leaderboard(api, data=api.get(f"leaderboards/{game.id}/category/{category.id}?embed=variables"))

data = []
for run in leaderboard.runs:
  run = run['run']
  row = {'date' : run.date,
         'time' : run.times['primary_t'],
         'runner' : ", ".join([player.name for player in run.players])}
  data.append(row)

But this method is very slow.

So, given this code (for reproducibility) :

import srcomapi, srcomapi.datatypes as dt
import pandas as pd

api = srcomapi.SpeedrunCom(); api.debug = 1
test_df = pd.DataFrame(['Castlevania: Aria of Sorrow','Pokémon: Yellow'],columns=['Game'])
nb_games = len(test_df)

lb_per_game = []

#getting the game for each name
for i in range(nb_games):
    try: 
      test_df.loc[i,'SRC_game'] = api.search(srcomapi.datatypes.Game, {"name": test_df.loc[i]['Game']})[0]
    except IndexError:
      pass

#collecting leaderboards for each game
for i in range(nb_games):
    lb = {}
    for category in test_df.loc[i,'SRC_game'].categories:
      if not category.name in lb:
        lb[category.name] = {}
      if category.type == 'per-level':
        pass
      # if category.miscellaneous:
      #   pass
      else:
        lb[category.name] = dt.Leaderboard(api, data=api.get("leaderboards/{}/category/{}?embed=variables".format(test_df.loc[i,'SRC_game'].id, category.id)))
    lb_per_game.append(lb)

This piece of code works :

platform = lb_per_game[0]['Soma Any%'].game.__getattr__('platforms')[0]
 print(platform)

But this one does not :

for i in range(2):
  game = lb_per_game[0]['Soma Any%'].game
  platform = game.__getattr__('platforms')[0]
  print(platform)

It raises a Type Error at the second time you enter the loop on second line :

object of type 'Platform' has no len()
  File "C:\Users\charl\OneDrive\Documents\Projects\OWREstimator\srcdata.py", line 68, in <module>
    platform = game.__getattr__('platforms')[0]

By inspecting the debugger, it seems the game.platforms attribute equals to this at this moment :
'Traceback (most recent call last):\n File "c:\\Users\\charl\\.vscode\\extensions\\ms-python.python-2021.6.944021595\\pythonFiles\\lib\\python\\debugpy\\_vendored\\pydevd\\_pydevd_bundle\\pydevd_resolver.py", line 193, in _get_py_dictionary\n attr = getattr(var, name)\n File "C:\\Users\\charl\\miniconda3\\lib\\site-packages\\srcomapi\\datatypes.py", line 46, in __getattr__\n if type(self.data[attr]) is list and len(self.data[attr]) > 0 and len(self.data[attr][0]) == 8:\nTypeError: object of type \'Platform\' has no len()\n'
(same for all not null attributes)

I have an application which gets unverified runs every minute (SRC api appears to update every 5mins), but it appears that the current raw requests.get() hits a cache and continues to fetch stale data that eg. firefox does not retrieve for around 10 minutes.

Isolated to the PHPSESSID cookie - there may be some optimisation on SRC's end that forces new sessions to hit a cache but allows existing sessions to fetch fresh data. Manually inserting the cookie from my Firefox into the headers correctly fetches fresh data.

.game and .category, before the object is printed, refer to ID strings. This behavior is inconsistent, especially when working between the interpreter and scripting.

I'm writing a tool that will use srcomapi and it would be really nice if it was part of PyPI so it could easily be installed as a dependency. Also, for future developers it would be nice if you could just pip install srcomapi to use this lib.

Speedrun.com uses status code 420 to indicate that the client is being rate-limited. Details here. (I say "possible" because I have no idea how to test for this, but I'm pretty sure it's the reason my code failed.)

If you call __repr__ the first time, everything works fine, but after this, the method always returns an error