genbrief/main.go

package main

import (
	"bytes"
	"context"
	_ "embed"
	"encoding/json"
	"fmt"
	"genBrief/db"
	"genBrief/util"
	"strings"

	"github.com/PuerkitoBio/goquery"
	openai "github.com/sashabaranov/go-openai"
)

//go:embed dom.json
var domJson string
var domMap map[string]string
var maxId int

func init() {
	domMap = map[string]string{}
	err := json.Unmarshal([]byte(domJson), &domMap)
	if err != nil {
		panic(err)
	}
	db.Init()
	db.New("t_config").Attr("value").Where("name", "prev_loop_new_id").GetRow().Scan(&maxId)
	fmt.Println("startTask at maxId ", maxId)
}

func main() {
	for {
		newMaxId := loopMaxId(maxId)
		if maxId >= newMaxId {
			fmt.Println("finishTask at maxId ", maxId)
			db.Pool().Update("t_config", map[string]interface{}{"value": maxId}, map[string]interface{}{"name": "prev_loop_new_id"})
			return
		}
		maxId = newMaxId
		loopMaxId(maxId)
	}
}

func loopMaxId(maxId int) int {
	list, err := db.New("t_news").Attr("id,content,url").
		Where("status", "Brief").WhereF("id > ?", 4000).Order("ORDER BY id asc").Limit(0, 100).GetAll()
	fmt.Println("loopLen", len(list), err)
	for _, record := range list {
		update := map[string]interface{}{}
		id := record["id"].(int64)
		content := record["content"].(string)
		url := record["url"].(string)
		brief := ""
		if len(content) >= 1024 {
			brief = genBrief(content)
			update["brief"] = brief
		} else {
			res, err := util.GetHtml(url)
			if err != nil {
				fmt.Println("get Html false", err.Error())
			} else {
				content = getContent(res, "body")
				update["content"] = content
				if len(content) >= 1024 {
					brief = genBrief(content)
					update["brief"] = brief
				}
			}
		}
		//
		if brief == "" {
			update["status"] = "Delete"
		} else {
			update["status"] = "Picture"
			tags := genTag(brief)
			if tags != "" {
				alltag := do_insert_tag(id, tags)
				update["tag"] = alltag
				update["pics"] = get_pics(alltag)
			}
		}
		fmt.Println("finish", id, brief)
		if _, err := db.Pool().Update("t_news", update, map[string]interface{}{"id": id}); err != nil {
			fmt.Println("save", err.Error())
		}
	}
	return maxId
}

func genBrief(content string) string {
	client := openai.NewClient("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")
	resp, err := client.CreateChatCompletion(
		context.Background(),
		openai.ChatCompletionRequest{
			Model: openai.GPT3Dot5Turbo,
			Messages: []openai.ChatCompletionMessage{
				{
					Role:    openai.ChatMessageRoleUser,
					Content: content + "\r\n通过以上内容，生成中文概要，文字控制500字以内。",
				},
			},
		},
	)
	if err != nil {
		fmt.Printf("ChatCompletion error: %v\n", err)
		return ""
	}
	return resp.Choices[0].Message.Content
}

func genTag(content string) string {
	client := openai.NewClient("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")
	resp, err := client.CreateChatCompletion(
		context.Background(),
		openai.ChatCompletionRequest{
			Model: openai.GPT3Dot5Turbo,
			Messages: []openai.ChatCompletionMessage{
				{
					Role:    openai.ChatMessageRoleUser,
					Content: content + "\r\n以上新闻内容属于哪一类新闻 A居民 B商业 C金融 D建筑 E屋内装饰 \r\n可以选一项或者两项",
				},
			},
		},
	)
	if err != nil {
		fmt.Printf("ChatCompletion error: %v\n", err)
		return ""
	}
	return resp.Choices[0].Message.Content
}

func get_pics(tags string) string {
	tag := strings.Split(tags, ",")[0]
	if tag == "" || tag == "other" {
		tag = "gpt"
	}
	var url string
	err := db.New("t_news_img").Attr("url").Where("tag", tag).Order("RDER BY RAND()").GetRow().Scan(&url)
	if err != nil {
		fmt.Println("get_pics err", tag, err.Error())
	}
	return url
}

func do_insert_tag(id int64, tags string) string {
	mtag := map[string]string{
		"A": "residential",
		"B": "commercial",
		"C": "financial",
		"D": "construction",
		"E": "indoor",
	}
	mtags := []string{}
	for opt, tag := range mtag {
		if strings.Contains(tags, opt) {
			record := map[string]interface{}{
				"new_id": id,
				"tag":    tag,
			}
			db.Pool().Insert("t_news_tag", record)
			mtags = append(mtags, tag)
		}
	}
	return strings.Join(mtags, ",")
}

func getContent(body []byte, dom string) string {
	doc, _ := goquery.NewDocumentFromReader(bytes.NewReader(body))
	doc.Find("script").Remove()
	doc.Find("noscript").Remove()
	if dom == "" {
		dom = "boby"
	}
	br := doc.Find(dom).Text()
	if br == "" {
		br = doc.Find("body").Text()
	}
	return util.TrimHtml(br)
}