pulsetriada.blogg.se - Dota replay player

DOTA REPLAY PLAYER HOW TO
DOTA REPLAY PLAYER FULL
DOTA REPLAY PLAYER PRO
DOTA REPLAY PLAYER CODE

Player/match caching: We cache matches in Redis in order to reduce DB lookups on repeated loads.When an error occurs or the job finishes, it either displays the error or redirects to the match page. The client uses AJAX to poll the server.This reads the match data from the steam API, then uses operations.insertMatchProgress in order to force waiting for the parse to finish. Request: Requests are processed from the Request page.Sequential: We read a match from the Steam API that either has leagueid>0 or contains a player in the trackedPlayer set.Pipeline: Generally parses come in one of two ways:.

DOTA REPLAY PLAYER FULL

fullhistory: Processes full history requests.

By permuting all three skill levels with the list of heroes, we can get up to 500 matches for each combination.

Applying the following filters increases the number of matches we can get skill data for.

skill: Reads the GetMatchHistory API in order to continuously find matches of a particular skill level.

proxy: A standalone HTTP server that simply proxies all requests to the Steam API.

If match.parse_status is explicitly set to 0, the match is queued for parse. If a match is found passing the criteria for parse.

scanner: Reads the Steam sequential API to find the latest matches.

Read the response from parser worker and save as match.parsed_data.

Get the replay URL: getReplayUrl takes care of this.

This uses the Node cluster module to fork as many workers as there are available parsing cores.

A single endpoint may appear multiple times (as many cores as it has).

parseManager: This reads Redis to find the currently available list of parse workers.The schema for the current parsed_data structure can be found in utility.getParseSchema.The parser emits a newline-delimited JSON stream of events, which is picked up and combined into a monolithic JSON object sent back to the client as the response.Each parser maintains a copy of the current heroes, which is used to map combat log names ("npc_dota_hero.") to hero_id, which can be used to match the combat log units to a player.dem.bz2 at this location, which it downloads, streams through bunzip2, and then through the compiled parser. parser: This is a standalone HTTP server that accepts a URL param url.Currently, this involves re-queueing currently active tasks on restart, and rebuilding the sets of tracked players, donated players, rating players, etc. worker: Takes care of background tasks.This is used in order to determine the list of users that have added a tracker as a friend.Accessing it without any params returns a list of the registered Steam accounts, and a hash mapping friends of those accounts to the Steam account.retriever: This is a standalone HTTP server that accepts URL params match_id and account_id, and interfaces with the Steam GC in order to return match details/account profile.

web: An HTTP server which serves the web traffic.

The project uses a microservice architecture, in order to promote modularity and allow different pieces to scale on different machines.

wget & unzip dota & mongorestore -dir dota to import a database dump if you want a medium-sized data set to work with.

Run all services in dev mode (this will run under nodemon so file changes automatically restart the server): npm run dev.

env file with required config values in KEY=VALUE format (see config.js for a full listing of options) cp.

DOTA REPLAY PLAYER HOW TO

If on Debian/Ubuntu: sudo bash init.sh Otherwise, you're responsible for figuring out how to install dependencies yourself.

DOTA REPLAY PLAYER CODE

Open Source: All code is publicly available for feedback and contributions from the Dota 2 developer community.Scalable: Designed to scale to thousands of users.Modular: Microservice architecture, with pieces that can be used independently.Rating Tracker: Keep track of MMR by adding a Steam account as a friend.Comparison Tool: Computes a percentile for a player against all users.

DOTA REPLAY PLAYER PRO

Pro Games: Professional matches are automatically parsed.

Word clouds (text said and read in all chat).

Teammates/Opponents (win rate playing with/against particular players).

Hero Matchups (win rate when playing as, with, against a hero).

Histogram (number of matches across Duration, LH, HD, TD, K, D, A, etc.).

Standard filters: patch, game mode, hero, etc.

Advanced Querying: Supports flexible querying and aggregation with the following criteria:.

Radiant advantage/Gold/XP/LH graphs per min.

Replay Parsing: Parses replays of Dota 2 matches to provide additional statistics per match.