mirror of
https://github.com/yusufipk/RepoHub.git
synced 2026-09-11 18:46:07 +00:00
feat: Add smart package recommendations feature
- Implement onboarding modal with 3-step wizard - Category selection (up to 3 categories) - OS detection with manual override - Experience level selection - Add intelligent recommendation engine - Hybrid scoring algorithm (category 40%, popularity 30%, OS 20%, preset 10%) - 7 curated categories with preset packages - Support for all platforms (Windows, macOS, Ubuntu, Debian, Arch, Fedora) - Create recommendation UI components - RecommendationsSection with grid layout - Package cards with recommendation scores and reasons - User profile display with customization options - Add localStorage-based profile management - Persistent user preferences - Automatic OS detection - Profile CRUD operations via useRecommendationProfile hook - Implement API endpoint - POST /api/recommendations - GET /api/recommendations (query params) - Request validation and error handling - Add full i18n support - English and Turkish translations - Onboarding flow, categories, and UI labels - Update TypeScript config (lib: es2017 for array.includes) Closes #1
This commit is contained in:
ersaayan
@@ -0,0 +1,294 @@
|
||||
import { PackageService } from './packageService'
|
||||
import { Package } from '@/models/Package'
|
||||
import {
|
||||
RecommendationRequest,
|
||||
RecommendedPackage,
|
||||
UserCategory,
|
||||
ExperienceLevel
|
||||
} from '@/types/recommendations'
|
||||
import {
|
||||
getPresetPackageNames,
|
||||
getPresetPriority,
|
||||
getRecommendationReason
|
||||
} from '@/data/recommendationPresets'
|
||||
|
||||
/**
|
||||
* Recommendation scoring weights
|
||||
*/
|
||||
const SCORING_WEIGHTS = {
|
||||
CATEGORY_MATCH: 0.4,
|
||||
POPULARITY: 0.3,
|
||||
OS_COMPATIBILITY: 0.2,
|
||||
PRESET_BOOST: 0.1
|
||||
}
|
||||
|
||||
export class RecommendationService {
|
||||
/**
|
||||
* Generate package recommendations based on user profile
|
||||
*/
|
||||
static async generateRecommendations(
|
||||
request: RecommendationRequest
|
||||
): Promise<RecommendedPackage[]> {
|
||||
const { platform_id, categories, experienceLevel, limit = 20 } = request
|
||||
|
||||
// Step 1: Get preset package names for the user's categories and platform
|
||||
const presetPackageNames = getPresetPackageNames(categories, platform_id)
|
||||
|
||||
// Step 2: Fetch packages from database
|
||||
// First, get preset packages
|
||||
const presetPackages = await this.fetchPresetPackages(
|
||||
presetPackageNames,
|
||||
platform_id
|
||||
)
|
||||
|
||||
// Then, get additional packages from categories
|
||||
const categoryPackages = await this.fetchCategoryPackages(
|
||||
categories,
|
||||
platform_id,
|
||||
limit * 2 // Fetch more to ensure we have enough after filtering
|
||||
)
|
||||
|
||||
// Step 3: Combine and deduplicate
|
||||
const allPackages = this.deduplicatePackages([
|
||||
...presetPackages,
|
||||
...categoryPackages
|
||||
])
|
||||
|
||||
// Step 4: Score and rank packages
|
||||
const scoredPackages = allPackages.map(pkg =>
|
||||
this.scorePackage(pkg, categories, platform_id, presetPackageNames, experienceLevel)
|
||||
)
|
||||
|
||||
// Step 5: Sort by score and limit results
|
||||
scoredPackages.sort((a, b) => b.recommendationScore - a.recommendationScore)
|
||||
|
||||
return scoredPackages.slice(0, limit)
|
||||
}
|
||||
|
||||
/**
|
||||
* Fetch packages that match preset names
|
||||
*/
|
||||
private static async fetchPresetPackages(
|
||||
packageNames: string[],
|
||||
platformId: string
|
||||
): Promise<Package[]> {
|
||||
if (packageNames.length === 0) {
|
||||
return []
|
||||
}
|
||||
|
||||
try {
|
||||
// Fetch packages by exact name match
|
||||
const packages: Package[] = []
|
||||
|
||||
for (const name of packageNames) {
|
||||
const result = await PackageService.getMany({
|
||||
platform_id: platformId,
|
||||
search: name,
|
||||
limit: 1,
|
||||
sort_by: 'popularity_score',
|
||||
sort_order: 'desc'
|
||||
})
|
||||
|
||||
// Only add if exact match
|
||||
if (result.packages.length > 0 && result.packages[0].name === name) {
|
||||
packages.push(result.packages[0])
|
||||
}
|
||||
}
|
||||
|
||||
return packages
|
||||
} catch (error) {
|
||||
console.error('Error fetching preset packages:', error)
|
||||
return []
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Fetch packages based on categories
|
||||
*/
|
||||
private static async fetchCategoryPackages(
|
||||
categories: UserCategory[],
|
||||
platformId: string,
|
||||
limit: number
|
||||
): Promise<Package[]> {
|
||||
try {
|
||||
// Map user categories to database categories
|
||||
const categoryMap: Record<UserCategory, string[]> = {
|
||||
'development': ['Development', 'Internet'],
|
||||
'design': ['Graphics', 'Multimedia'],
|
||||
'multimedia': ['Multimedia', 'Graphics'],
|
||||
'system-tools': ['System', 'Utilities'],
|
||||
'gaming': ['Games'],
|
||||
'productivity': ['Office', 'Utilities'],
|
||||
'education': ['Science', 'Education']
|
||||
}
|
||||
|
||||
// Get all matching packages
|
||||
const packages: Package[] = []
|
||||
|
||||
for (const category of categories) {
|
||||
const dbCategories = categoryMap[category] || []
|
||||
|
||||
// Note: Since we don't have category filtering in current API,
|
||||
// we'll fetch by popularity and filter client-side
|
||||
// This is a limitation of current schema - categories are not well-utilized
|
||||
const result = await PackageService.getMany({
|
||||
platform_id: platformId,
|
||||
limit: Math.ceil(limit / categories.length),
|
||||
sort_by: 'popularity_score',
|
||||
sort_order: 'desc'
|
||||
})
|
||||
|
||||
packages.push(...result.packages)
|
||||
}
|
||||
|
||||
return packages
|
||||
} catch (error) {
|
||||
console.error('Error fetching category packages:', error)
|
||||
return []
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Remove duplicate packages (by ID)
|
||||
*/
|
||||
private static deduplicatePackages(packages: Package[]): Package[] {
|
||||
const seen = new Set<string>()
|
||||
return packages.filter(pkg => {
|
||||
if (seen.has(pkg.id)) {
|
||||
return false
|
||||
}
|
||||
seen.add(pkg.id)
|
||||
return true
|
||||
})
|
||||
}
|
||||
|
||||
/**
|
||||
* Score a package based on multiple factors
|
||||
*/
|
||||
private static scorePackage(
|
||||
pkg: Package,
|
||||
categories: UserCategory[],
|
||||
platformId: string,
|
||||
presetPackageNames: string[],
|
||||
experienceLevel?: ExperienceLevel
|
||||
): RecommendedPackage {
|
||||
let score = 0
|
||||
let reason = ''
|
||||
const isPresetMatch = presetPackageNames.includes(pkg.name)
|
||||
|
||||
// 1. Category Match Score (40%)
|
||||
// For preset packages, this is always high
|
||||
const categoryScore = isPresetMatch ? 1.0 : 0.5
|
||||
score += categoryScore * SCORING_WEIGHTS.CATEGORY_MATCH
|
||||
|
||||
// 2. Popularity Score (30%)
|
||||
// Normalize popularity_score (0-100) to 0-1
|
||||
const popularityScore = (pkg.popularity_score || 0) / 100
|
||||
score += popularityScore * SCORING_WEIGHTS.POPULARITY
|
||||
|
||||
// 3. OS Compatibility Score (20%)
|
||||
// All packages from DB should be compatible, so this is always 1.0
|
||||
const osScore = 1.0
|
||||
score += osScore * SCORING_WEIGHTS.OS_COMPATIBILITY
|
||||
|
||||
// 4. Preset Boost (10%)
|
||||
// Extra boost for preset packages based on priority
|
||||
let presetBoost = 0
|
||||
if (isPresetMatch) {
|
||||
const priority = getPresetPriority(pkg.name, categories, platformId)
|
||||
if (priority !== null) {
|
||||
presetBoost = priority / 10 // Normalize 1-10 to 0.1-1.0
|
||||
|
||||
// Get recommendation reason from preset
|
||||
const presetReason = getRecommendationReason(pkg.name, categories)
|
||||
if (presetReason) {
|
||||
reason = presetReason
|
||||
}
|
||||
}
|
||||
}
|
||||
score += presetBoost * SCORING_WEIGHTS.PRESET_BOOST
|
||||
|
||||
// Default reason if not from preset
|
||||
if (!reason) {
|
||||
if (pkg.popularity_score && pkg.popularity_score > 70) {
|
||||
reason = 'Popular choice in the community'
|
||||
} else {
|
||||
reason = 'Recommended for your selected categories'
|
||||
}
|
||||
}
|
||||
|
||||
// Normalize final score to 0-100
|
||||
const finalScore = Math.round(score * 100)
|
||||
|
||||
return {
|
||||
id: pkg.id,
|
||||
name: pkg.name,
|
||||
description: pkg.description || 'No description available',
|
||||
version: pkg.version || 'latest',
|
||||
category: typeof pkg.category === 'string' ? pkg.category : pkg.category?.name,
|
||||
license: typeof pkg.license === 'string' ? pkg.license : pkg.license?.name,
|
||||
type: pkg.type || 'cli',
|
||||
platform: pkg.platform,
|
||||
platform_id: pkg.platform_id,
|
||||
repository: pkg.repository || 'official',
|
||||
download_url: pkg.download_url,
|
||||
lastUpdated: pkg.last_updated ? pkg.last_updated.toString() : undefined,
|
||||
downloads: pkg.downloads_count,
|
||||
popularity: pkg.popularity_score,
|
||||
popularity_score: pkg.popularity_score,
|
||||
tags: pkg.tags,
|
||||
recommendationScore: finalScore,
|
||||
recommendationReason: reason,
|
||||
presetMatch: isPresetMatch
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Get quick start recommendations (top 5 most essential)
|
||||
*/
|
||||
static async getQuickStartRecommendations(
|
||||
platformId: string,
|
||||
primaryCategory: UserCategory
|
||||
): Promise<RecommendedPackage[]> {
|
||||
return this.generateRecommendations({
|
||||
platform_id: platformId,
|
||||
categories: [primaryCategory],
|
||||
limit: 5
|
||||
})
|
||||
}
|
||||
|
||||
/**
|
||||
* Get recommendations for multiple categories with balanced distribution
|
||||
*/
|
||||
static async getBalancedRecommendations(
|
||||
platformId: string,
|
||||
categories: UserCategory[],
|
||||
totalLimit: number = 20
|
||||
): Promise<RecommendedPackage[]> {
|
||||
const perCategory = Math.ceil(totalLimit / categories.length)
|
||||
const allRecommendations: RecommendedPackage[] = []
|
||||
|
||||
for (const category of categories) {
|
||||
const recommendations = await this.generateRecommendations({
|
||||
platform_id: platformId,
|
||||
categories: [category],
|
||||
limit: perCategory
|
||||
})
|
||||
allRecommendations.push(...recommendations)
|
||||
}
|
||||
|
||||
// Deduplicate by ID and re-sort
|
||||
const seen = new Set<string>()
|
||||
const deduplicated = allRecommendations.filter(pkg => {
|
||||
if (seen.has(pkg.id)) {
|
||||
return false
|
||||
}
|
||||
seen.add(pkg.id)
|
||||
return true
|
||||
})
|
||||
|
||||
deduplicated.sort((a, b) => b.recommendationScore - a.recommendationScore)
|
||||
|
||||
return deduplicated.slice(0, totalLimit)
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user