General Discussion

General DiscussionDotabuff API

Dotabuff API in General Discussion
Veni Vidi Vodka

    This may have been asked many times, but why not have a nice Dotabuff API?

    There are some stats I don't want to pull with a small IRCBot, which could be supplied by a service like Dotabuff, which keeps track of all games. The scope of my IRCbot is simply too small to support +2k games per player and maintain my own stats.
    Is this possible somehow?

    game is bad

      I guess because then why would you buy dotabuff plus.

      swoleytrinity

        @Veni what are you trying to pull and what are you trying to ignore.

        Veni Vidi Vodka

          Stats like "Top 3 Matches" or "Top 3 Heroes", some records and stuff like that, which needs to be processes manually with the entire data-segment.

          swoleytrinity

            Specifics! Those examples are broad and vague, give me a real example of a dataset you're trying to pull.

            Top 3 matches? Based on what? Time? Kills? Deaths?

            Top 3 Heroes? Based on what? KDA? Play time?

            Specifics are important when dealing with information.

            Veni Vidi Vodka

              Most Played Heroes (top 3), Winrate in general, Recordstats like GPM, XPM, Kills / Killparticipation, Minutes. Stuff people brag about.
              The bot is rather minor and I just threw a few ideas together in a basic wrapper for the Steam API.

              CaptainDreads

                Theoretically this should get 3 most played heroes for a player

                curl http://dotabuff.com/players/<player_id>/heroes?metric=played | grep -om3 title=\"[a-zA-Z]*\" | sed -n 3,5p | grep -o \".*\" | sed 's/"//' | sed 's/"//'

                PLEASE DON'T ACTUALLY USE THIS - SCREEN SCRAPING IS BAD

                I was bored and wanted to bodge something... I really need to learn how to awk

                Tento komentář byl upraven
                Veni Vidi Vodka

                  Nobody is that stupid. I have already an idea how I could design a very minimal database which could run through the SteamAPI very effective - but still, it isn't the most optimal solution (and needs way more computation-power).