feat: update language leaderboard calculation

This commit is contained in:
2024-08-09 21:45:14 +07:00
parent a5cd6383f3
commit 0142224f60
19 changed files with 330 additions and 59 deletions
@@ -1,3 +1,11 @@
CREATE TABLE `repository_languages` (
`id` integer PRIMARY KEY AUTOINCREMENT NOT NULL,
`repo_id` integer NOT NULL,
`name` text NOT NULL,
`percentage` real NOT NULL,
FOREIGN KEY (`repo_id`) REFERENCES `repositories`(`id`) ON UPDATE no action ON DELETE no action
);
--> statement-breakpoint
CREATE TABLE `repositories` (
`id` integer PRIMARY KEY AUTOINCREMENT NOT NULL,
`user_id` integer NOT NULL,
@@ -7,7 +15,6 @@ CREATE TABLE `repositories` (
`stars` integer NOT NULL,
`forks` integer NOT NULL,
`last_update` text NOT NULL,
`languages` text,
`contributors` text,
`is_pending` integer DEFAULT false NOT NULL,
`is_error` integer DEFAULT false NOT NULL,
@@ -34,6 +41,7 @@ CREATE TABLE `users` (
`updated_at` text NOT NULL
);
--> statement-breakpoint
CREATE INDEX `repository_languages_name_idx` ON `repository_languages` (`name`);--> statement-breakpoint
CREATE INDEX `repositories_name_idx` ON `repositories` (`name`);--> statement-breakpoint
CREATE INDEX `repositories_uri_idx` ON `repositories` (`uri`);--> statement-breakpoint
CREATE INDEX `repositories_language_idx` ON `repositories` (`language`);--> statement-breakpoint
+60 -8
View File
@@ -1,9 +1,68 @@
{
"version": "6",
"dialect": "sqlite",
"id": "13d859a6-4c32-46a6-90e9-19740b2bb13a",
"id": "a268ec23-239a-4d86-8830-2f3c1c2110df",
"prevId": "00000000-0000-0000-0000-000000000000",
"tables": {
"repository_languages": {
"name": "repository_languages",
"columns": {
"id": {
"name": "id",
"type": "integer",
"primaryKey": true,
"notNull": true,
"autoincrement": true
},
"repo_id": {
"name": "repo_id",
"type": "integer",
"primaryKey": false,
"notNull": true,
"autoincrement": false
},
"name": {
"name": "name",
"type": "text",
"primaryKey": false,
"notNull": true,
"autoincrement": false
},
"percentage": {
"name": "percentage",
"type": "real",
"primaryKey": false,
"notNull": true,
"autoincrement": false
}
},
"indexes": {
"repository_languages_name_idx": {
"name": "repository_languages_name_idx",
"columns": [
"name"
],
"isUnique": false
}
},
"foreignKeys": {
"repository_languages_repo_id_repositories_id_fk": {
"name": "repository_languages_repo_id_repositories_id_fk",
"tableFrom": "repository_languages",
"tableTo": "repositories",
"columnsFrom": [
"repo_id"
],
"columnsTo": [
"id"
],
"onDelete": "no action",
"onUpdate": "no action"
}
},
"compositePrimaryKeys": {},
"uniqueConstraints": {}
},
"repositories": {
"name": "repositories",
"columns": {
@@ -63,13 +122,6 @@
"notNull": true,
"autoincrement": false
},
"languages": {
"name": "languages",
"type": "text",
"primaryKey": false,
"notNull": false,
"autoincrement": false
},
"contributors": {
"name": "contributors",
"type": "text",
+2 -2
View File
@@ -5,8 +5,8 @@
{
"idx": 0,
"version": "6",
"when": 1723210265266,
"tag": "0000_young_emma_frost",
"when": 1723211354217,
"tag": "0000_tough_jubilee",
"breakpoints": true
}
]
+34
View File
@@ -0,0 +1,34 @@
import db from "./db";
import queue from "./lib/queue";
import { users } from "./models";
const main = async () => {
const username = process.argv[2];
if (!username) {
throw new Error("Missing username");
}
const [user] = await db
.insert(users)
.values({ username, name: username })
.onConflictDoUpdate({
target: users.username,
set: { username },
})
.returning();
await queue.add(
"fetchUserProfile",
{ userId: user.id },
{ jobId: `fetchUserProfile:${user.id}` }
);
await queue.add(
"fetchUserRepos",
{ userId: user.id },
{ jobId: `fetchUserRepos:${user.id}` }
);
process.exit();
};
main();
+5 -5
View File
@@ -37,15 +37,15 @@ export const calculateUserPoints = async (data: CalculateUserPointsType) => {
: 0;
// User repositories
const repos = await db
.select()
.from(repositories)
.where(eq(repositories.userId, user.id));
const repos = await db.query.repositories.findMany({
where: eq(repositories.userId, user.id),
with: { languages: true },
});
points += repos.length * weights.repositories;
// Languages known
const languages = new Set(
repos.flatMap((i) => i.languages?.map((j) => j.lang))
repos.flatMap((i) => i.languages?.map((j) => j.name))
);
points += languages.size * weights.languagesKnown;
+49 -10
View File
@@ -2,7 +2,8 @@ import db from "@server/db";
import github from "@server/lib/github";
import queue from "@server/lib/queue";
import { repositories } from "@server/models";
import { eq } from "drizzle-orm";
import { repoLanguages } from "@server/models/repo-languages";
import { and, eq, notInArray } from "drizzle-orm";
export type FetchRepoDataJobType = {
id: number;
@@ -10,17 +11,55 @@ export type FetchRepoDataJobType = {
};
export const fetchRepoData = async (data: FetchRepoDataJobType) => {
const details = await github.getRepoDetails(data.uri);
const [repository] = await db
.select()
.from(repositories)
.where(eq(repositories.id, data.id));
const [result] = await db
.update(repositories)
.set({ languages: details.languages })
.where(eq(repositories.id, data.id))
.returning();
if (!result) {
if (!repository) {
throw new Error("Repository not found!");
}
await queue.add("calculateUserPoints", { userId: result.userId });
const details = await github.getRepoDetails(data.uri);
const { languages } = details;
await db.transaction(async (tx) => {
// Remove languages that don't exist anymore
const purgeLangFilter = and(
eq(repoLanguages.repoId, repository.id),
notInArray(
repoLanguages.name,
languages.map((i) => i.lang)
)
);
await tx.delete(repoLanguages).where(purgeLangFilter);
// Add or update languages
for (const lang of languages) {
const [existing] = await tx
.select({ id: repoLanguages.id })
.from(repoLanguages)
.where(
and(
eq(repoLanguages.repoId, repository.id),
eq(repoLanguages.name, lang.lang)
)
);
if (existing) {
await tx
.update(repoLanguages)
.set({ percentage: lang.amount })
.where(eq(repoLanguages.id, existing.id));
} else {
await tx.insert(repoLanguages).values({
repoId: repository.id,
name: lang.lang,
percentage: lang.amount,
});
}
}
});
await queue.add("calculateUserPoints", { userId: repository.userId });
};
+1
View File
@@ -19,6 +19,7 @@ export const fetchUserProfile = async (data: FetchUserProfileType) => {
.update(users)
.set({
name: details.name,
avatar: details.avatar,
followers: details.followers,
following: details.following,
location: details.location,
+4
View File
@@ -8,6 +8,7 @@ const GITHUB_API_URL = "https://api.github.com";
const selectors = {
user: {
name: "h1.vcard-names > span.vcard-fullname",
avatar: ".js-profile-editable-replace img.avatar-user",
location: "li[itemprop='homeLocation'] span",
followers: ".js-profile-editable-area a[href$='?tab=followers'] > span",
following: ".js-profile-editable-area a[href$='?tab=following'] > span",
@@ -30,6 +31,8 @@ const github = {
const $ = cheerio.load(response);
const name = $(selectors.user.name).text().trim();
const avatar = $(selectors.user.avatar).attr("src");
console.log({ avatar });
const location = $(selectors.user.location).text().trim();
const followers = intval($(selectors.user.followers).text().trim());
const following = intval($(selectors.user.following).text().trim());
@@ -43,6 +46,7 @@ const github = {
return {
name: name || username,
avatar,
username,
location,
followers,
+3
View File
@@ -12,6 +12,9 @@ export const intval = (value: any) => {
export const getLanguageLogo = (language: string) => {
const alias: Record<string, string> = {
gdscript: "godot",
gap: "godot",
html: "html5",
css: "css3",
};
let lang = language.toLowerCase().replace(/[^a-zA-Z]/g, "");
+2 -1
View File
@@ -1,2 +1,3 @@
export { users } from "./users";
export { repositories } from "./repositories";
export { repositories, repositoriesRelations } from "./repositories";
export { repoLanguages, repoLanguagesRelations } from "./repo-languages";
+35
View File
@@ -0,0 +1,35 @@
import { InferInsertModel, InferSelectModel, relations } from "drizzle-orm";
import {
text,
sqliteTable,
integer,
index,
real,
} from "drizzle-orm/sqlite-core";
import { repositories } from "./repositories";
export const repoLanguages = sqliteTable(
"repository_languages",
{
id: integer("id").primaryKey({ autoIncrement: true }),
repoId: integer("repo_id")
.notNull()
.references(() => repositories.id),
name: text("name").notNull(),
percentage: real("percentage").notNull(),
},
(t) => ({
nameIdx: index("repository_languages_name_idx").on(t.name),
})
);
export const repoLanguagesRelations = relations(repoLanguages, ({ one }) => ({
repository: one(repositories, {
fields: [repoLanguages.repoId],
references: [repositories.id],
}),
}));
export type RepositoryLanguage = InferSelectModel<typeof repoLanguages>;
export type CreateRepositoryLanguage = InferInsertModel<typeof repoLanguages>;
+12 -3
View File
@@ -1,7 +1,13 @@
import { Contributor, Language } from "@server/lib/github";
import { InferInsertModel, InferSelectModel, sql } from "drizzle-orm";
import { Contributor } from "@server/lib/github";
import {
InferInsertModel,
InferSelectModel,
relations,
sql,
} from "drizzle-orm";
import { text, sqliteTable, integer, index } from "drizzle-orm/sqlite-core";
import { users } from "./users";
import { repoLanguages } from "./repo-languages";
export const repositories = sqliteTable(
"repositories",
@@ -17,7 +23,6 @@ export const repositories = sqliteTable(
stars: integer("stars").notNull(),
forks: integer("forks").notNull(),
lastUpdate: text("last_update").notNull(),
languages: text("languages", { mode: "json" }).$type<Language[]>(),
contributors: text("contributors", { mode: "json" }).$type<Contributor[]>(),
isPending: integer("is_pending", { mode: "boolean" })
@@ -39,5 +44,9 @@ export const repositories = sqliteTable(
})
);
export const repositoriesRelations = relations(repositories, ({ many }) => ({
languages: many(repoLanguages),
}));
export type Repository = InferSelectModel<typeof repositories>;
export type CreateRepository = InferInsertModel<typeof repositories>;
+1 -1
View File
@@ -22,7 +22,7 @@ const onJobRetriesExhausted = async (job: Job) => {
const worker = new Worker(BULLMQ_JOB_NAME, handler, {
connection: BULLMQ_CONNECTION,
concurrency: Number(import.meta.env.QUEUE_CONCURRENCY) || 1,
concurrency: Number(import.meta.env.QUEUE_CONCURRENCY) || 10,
removeOnComplete: { count: 0 },
removeOnFail: { count: 0 },
});
+37 -14
View File
@@ -1,11 +1,11 @@
import db from "@server/db";
import { getLanguageLogo } from "@server/lib/utils";
import { repositories, users } from "@server/models";
import { repoLanguages, repositories, users } from "@server/models";
import { count, desc, eq, getTableColumns, inArray, sql } from "drizzle-orm";
import { HTTPException } from "hono/http-exception";
export type GetLeaderboardRes = {
columns: { title: string; selector: string }[];
columns: { title: string; selector: string; hideOnSmall?: boolean }[];
rows: {
id?: string | null;
rank: number;
@@ -48,12 +48,18 @@ export class LeaderboardRepository {
async getTopLanguages() {
const data = await db
.select({
language: repositories.language,
language: repoLanguages.name,
count: count(),
total: sql<number>`sum(${repoLanguages.percentage})`.as("total"),
avg: sql<number>`avg(${repoLanguages.percentage})`.as("avg"),
coverage: sql<number>`count(*) * avg(${repoLanguages.percentage})`.as(
"coverage"
),
})
.from(repositories)
.groupBy(repositories.language)
.orderBy(desc(count()));
.innerJoin(repoLanguages, eq(repoLanguages.repoId, repositories.id))
.groupBy(repoLanguages.name)
.orderBy(desc(sql`count(*) * avg(${repoLanguages.percentage})`));
const columns = [
{
@@ -62,8 +68,21 @@ export class LeaderboardRepository {
},
{
title: "Total Repo",
selector: "repo",
},
{
title: "Rerata",
selector: "avg",
hideOnSmall: true,
},
{
title: "Cakupan",
selector: "sub",
},
// {
// title: "Total Persentase",
// selector: "total",
// },
];
const rows = data
@@ -71,8 +90,11 @@ export class LeaderboardRepository {
.map((data, idx) => ({
rank: idx + 1,
name: data.language,
sub: `${data.count} repo`,
sub: data.coverage.toFixed(0),
image: getLanguageLogo(data.language),
repo: `${data.count} repo`,
total: data.total,
avg: data.avg.toFixed(1),
}));
return { columns, rows } satisfies GetLeaderboardRes;
@@ -91,7 +113,8 @@ export class LeaderboardRepository {
})
.from(repositories)
.innerJoin(users, eq(users.id, repositories.userId))
.where(inArray(sql`lower(${repositories.language})`, languages))
.innerJoin(repoLanguages, eq(repoLanguages.repoId, repositories.id))
.where(inArray(sql`lower(${repoLanguages.name})`, languages))
.groupBy(users.id)
.orderBy(desc(count()));
@@ -141,18 +164,18 @@ export class LeaderboardRepository {
throw new HTTPException(404, { message: "User not found!" });
}
const repos = await db
.select()
.from(repositories)
.where(eq(repositories.userId, user.id))
.orderBy(desc(repositories.stars), desc(repositories.forks));
const repos = await db.query.repositories.findMany({
where: eq(repositories.userId, user.id),
orderBy: [desc(repositories.stars), desc(repositories.forks)],
with: { languages: true },
});
const languageMap: Record<string, number> = {};
repos
.flatMap((i) => i.languages || [])
.forEach((i) => {
if (!languageMap[i.lang]) languageMap[i.lang] = 0;
languageMap[i.lang] += i.amount;
if (!languageMap[i.name]) languageMap[i.name] = 0;
languageMap[i.name] += i.percentage;
});
const totalLangWeight = Object.values(languageMap).reduce(