package main

import (
	"fmt"
	"math/rand"
	"time"

	"hr_receiver/config"
	"hr_receiver/models"
)

func main() {
	config.InitConfig()
	config.ConnectDB()

	// 生成100条测试数据
	count := 100
	records := make([]models.AIAnalysisRecord, 0, count)

	for i := 0; i < count; i++ {
		records = append(records, generateRecord())
	}

	if err := config.DB.CreateInBatches(records, 50).Error; err != nil {
		panic("failed to insert mock data: " + err.Error())
	}

	fmt.Printf("成功插入 %d 条 AI 分析记录\n", count)
}

func generateRecord() models.AIAnalysisRecord {
	// regionID 为 1 或 3
	regionID := uint32(1)
	if rand.Intn(2) == 1 {
		regionID = 3
	}

	// sourceType: upload 或 cloud
	sourceType := "upload"
	if rand.Intn(2) == 1 {
		sourceType = "cloud"
	}

	// docx 教案原始文件大小: 50KB ~ 500KB
	docxSize := int64(rand.Intn(451*1024) + 50*1024)

	// 心率 csv 原始文件大小: 约 80KB (70KB ~ 90KB)
	csvSize := int64(rand.Intn(20*1024) + 70*1024)

	// 步数 csv 原始文件大小: 约 20KB ~ 40KB (heart_rate_with_steps 时才有)
	var stepCsvSize int64
	analysisType := analysisType()
	if analysisType == "heart_rate_with_steps" {
		stepCsvSize = int64(rand.Intn(20*1024) + 20*1024)
	}

	originalFileSize := docxSize + csvSize + stepCsvSize

	// 压缩后内容大小: csv 每4行保留1行，大约压缩为 25% + 表头；docx 提取文本后大约 30%~60%
	compressedDocx := int64(float64(docxSize) * (0.3 + rand.Float64()*0.3))
	compressedCsv := int64(float64(csvSize) * (0.22 + rand.Float64()*0.08)) // ~22%-30%
	var compressedStepCsv int64
	if stepCsvSize > 0 {
		compressedStepCsv = int64(float64(stepCsvSize) * (0.22 + rand.Float64()*0.08))
	}
	compressedContentSize := compressedDocx + compressedCsv + compressedStepCsv

	// prompt 大小 = 压缩后内容 + 提示词模板 (~1.5KB)
	promptTemplateSize := 1500 + rand.Intn(500)
	inputSizeBytes := int(compressedContentSize) + promptTemplateSize

	// AI 输出大小: 3KB ~ 25KB (分析报告)
	outputSizeBytes := rand.Intn(22*1024) + 3*1024

	// token 估算: 中文混合场景，平均约 3.5 字节/token
	inputTokens := inputSizeBytes / (3 + rand.Intn(2))
	outputTokens := outputSizeBytes / (3 + rand.Intn(2))

	// 分析时长: 主要和输出 token 数量相关，1分钟以内
	// 基础延迟 500ms + 每token约 15~40ms
	tokenLatency := int64(15 + rand.Intn(26))
	durationMs := 500 + int64(outputTokens)*tokenLatency
	if durationMs > 60000 {
		durationMs = 60000 - int64(rand.Intn(5000))
	}

	// 上传时间: 最近 90 天内随机
	uploadTime := time.Now().Add(-time.Duration(rand.Intn(90*24)) * time.Hour).Add(-time.Duration(rand.Intn(60)) * time.Minute).UnixMilli()

	return models.AIAnalysisRecord{
		RegionID:              &regionID,
		SourceType:            sourceType,
		InputTokens:           inputTokens,
		OutputTokens:          outputTokens,
		InputSizeBytes:        inputSizeBytes,
		OutputSizeBytes:       outputSizeBytes,
		DurationMs:            durationMs,
		OriginalFileSize:      originalFileSize,
		CompressedContentSize: compressedContentSize,
		UploadTime:            uploadTime,
	}
}

func analysisType() string {
	// 约 30% 的带步数分析
	if rand.Intn(100) < 30 {
		return "heart_rate_with_steps"
	}
	return "heart_rate_only"
}