ddstats-server

Summary

Date : 2020-01-12 21:02:43

Total : 42 files, 3887 codes, 176 comments, 528 blanks, all 4591 lines

Languages

language	files	code	comment	blank	total
Go	37	3,597	151	464	4,212
Markdown	1	139	0	45	184
SQL	1	109	3	14	126
HTTP	1	25	21	2	48
XML	1	16	0	3	19
Properties	1	1	1	0	2

Directories

path	files	code	comment	blank	total
.	42	3,887	176	528	4,591
cmd	1	101	1	26	128
cmd/server	1	101	1	26	128
pkg	36	3,496	150	438	4,084
pkg/api	7	817	16	155	988
pkg/ddapi	2	303	34	54	391
pkg/discord	14	888	15	88	991
pkg/models	8	915	57	84	1,056
pkg/models/postgres	7	778	41	66	885
pkg/socketio	1	266	9	25	300
pkg/websocket	4	307	19	32	358

Building protos

protoc --go_out=. --go_opt=paths=source_relative
--go-grpc_out=. --go-grpc_opt=paths=source_relative
gamesubmission/gamesubmission.proto

Building for linux

GOOS=linux GOARCH=amd64 go build -o dist/server -v ./cmd/server

GOOS=linux GOARCH=amd64 go build -o dist/collector -v ./cmd/collector

Automatically restarting server during dev

Go 1.11+ installation required

go get github.com/cespare/reflex

Run reflex using the following command:

reflex -d none -c reflex.conf

Things Left to Do...

authorization

ui

collector

add game_time_improvement to collector_active_player
add the following to runs: bronze_daggers, silver_daggers, gold_daggers, devil_daggers, total_bronze, total_silver, total_gold, total_devil
add last_active to player table and collect using collector
possibly add deaths and total_game_time to active players column

discord-bot commands

MAKE SURE TO CHECK FOR SURVIVAL HASH WHEN MAKING THE NOTIFICATION FUNCTION!

submission stuff

finish the logic behind game submissions and filtering (i think this is done)

live notification functionality

verify & fix bugs in client that relate to game submissions https://github.com/alexwilkerson/ddstats-go/blob/master/net.go#L80-L86
create socket.io listener for game_submitted function which has arguments (gameID int, notifyPlayerBest bool, notifyAbove1000 bool)
create discord bot listener which runs in its own goroutine and has channels for notifying player best and notifying when player is above 1000
create connection between socket.io -> websocket -> discord for player best and above 1000 notifications

TODO: endpoints

collector

api/v2/daily returns most recent
api/v2/daily?date={?}

index

set up /static endpoint to host static files
GET player/live (gets live players from websocket)
POST api/v2/client_connect (client sends client version and receives info)
GET api/v2/motd
create table in database for motd
GET api/v2/client/releases?page_size={int}&page_num={int}
create table for client releases
GET api/v2/news?page_size={int}&page_num={int}
create table in database for news

game

player

GET api/v2/player/update?id={int} (updates or creates user in database from dd backend)
GET api/v2/player/all?page_size={int}&page_num={int} (list all users)
GET api/v2/player?id={int}
~~GET api/v2/player/games?id={int}page_size={int}&page_num={int} (paginated list of games by player id)~~ (see api/v2/game/recent)

dd backend

GET api/v2/ddapi/get_user_by_rank?rank={int}
GET api/v2/ddapi/get_user_by_id?id={int}
GET api/v2/ddapi/user_search?user={string}
GET api/v2/ddapi/get_scores?offset={int}&limit={int}

Experimental stuff

store daily data about ddstats users
- each day, download their information to the database
- end goal would be to eventually show stats over a period of time so users could continuously track their improvement day-to-day
- possible daily data:
  - all data for that day
    - play time
    - number of deaths
    - gems
    - daggers fired
    - daggers hit
    - enemies killed
    - accuracy
    - etc, etc, etc... basically all of the player data
a function which runs every day at midnight which goes through the entire devil daggers backend, retrieves user information, stores it to the database. during the process of data collection, the function will analyze the informatiod and store the data to the database. it's possible that this function will be too expensive to process or take too much time, but hopefully it would work.
- the function would run daily
- possible daily data:
  - users active for that day
  - how many players got a new score
  - how many new players
  - average improvement among players who got new scores

Exporting data from SQLite3 database to PostgreSQL database

importing csvs from app.db

In order to port the database from SQLite3 to Postgres, the following shenanigans must occur:

scp app.db from server, then: sqlite3 app.db
Afterwards, import the csvs via Postico.
See schema.sql for creating the Postgres database.

game

.headers on
.mode csv
.output game.csv
-- printf required because of sqlite's weird float formatting
select id, player_id, granularity, printf("%.6f", game_time) as game_time, death_type, gems, homing_daggers, daggers_fired, daggers_hit, enemies_alive, enemies_killed, time_stamp, replay_player_id, survival_hash, version, printf("%.6f", level_two_time) as level_two_time, printf("%.6f", level_three_time) as level_three_time, printf("%.6f", level_four_time) as level_four_time, printf("%.6f", homing_daggers_max_time) as homing_daggers_max_time, printf("%.6f", enemies_alive_max_time) as enemies_alive_max_time, homing_daggers_max, enemies_alive_max from game;

state

.headers on
.mode csv
.output state.csv
-- remove any inconsistencies between state and game tables
delete from state where game_id not in (select id from game where id is not null);
-- printf required because of sqlite's weird float formatting
select id, game_id, printf("%.6f", game_time) as game_time, gems, homing_daggers, daggers_hit, daggers_fired, enemies_alive, enemies_killed from state;

might have to split up state to get it into postgres split -l 2500000 state.csv works

spawnset

.headers on
.mode csv
.output spawnset.csv
select * from spawnset;

player (renamed from user)

.headers on
.mode csv
.output player.csv
select id, username as player_name, rank, game_time, death_type, gems, daggers_fired, daggers_hit, enemies_killed, accuracy, time_total as overall_game_time, deaths_total as overall_deaths, gems_total as overall_gems, enemies_killed_total as overall_enemies_killed, daggers_fired_total as overall_daggers_fired, daggers_hit_total as overall_daggers_hit, accuracy_total as overall_accuracy from user;

add overall_game_time: UPDATE player SET overall_average_game_time=TRUNC(DIVZERO(overall_game_time, overall_deaths)::numeric, 4);

live

.headers on
.mode csv
.output live.csv
select * from live;

discord bot database

discord_users

.headers on
.mode csv
.output discord_users.csv
select * from users;

little helper to delete non-ddstats-users from database

DELETE FROM player
WHERE id<>-1 AND id IN (
	SELECT player.id
	FROM player LEFT JOIN game ON player.id=game.player_id
	WHERE game.id IS NULL
	GROUP BY player.id
);```

![DDSTATS Logo](/ui/static/ddstats_logo_v2_black_100px.png)

deployment commands

yarn build --mode production (from ui dir) scp -r dist casd:~/ddstats/ui (from ui dir)

lsaa/ddstats-server

ddstats-server

Summary

Languages

Directories

Building protos

Building for linux

Automatically restarting server during dev

Things Left to Do...

ui

pages

collector

discord-bot commands

submission stuff

live notification functionality

TODO: endpoints

collector

index

game

player

dd backend

Experimental stuff

Exporting data from SQLite3 database to PostgreSQL database

importing csvs from app.db

game

state

spawnset

player (renamed from user)

live

discord bot database

discord_users

little helper to delete non-ddstats-users from database

deployment commands