pcgamedb/crawler/goggames.go

package crawler

import (
	"encoding/json"
	"fmt"
	"path"
	"strings"
	"time"

	"pcgamedb/config"
	"pcgamedb/constant"
	"pcgamedb/db"
	"pcgamedb/model"
	"pcgamedb/utils"

	"git.nite07.com/nite/ccs"
	"go.uber.org/zap"
)

type GOGGamesCrawler struct {
	logger *zap.Logger
}

func NewGOGGamesCrawler(logger *zap.Logger) *GOGGamesCrawler {
	return &GOGGamesCrawler{
		logger: logger,
	}
}

func (c *GOGGamesCrawler) Name() string {
	return "GOGGamesCrawler"
}

func (c *GOGGamesCrawler) CrawlByUrl(URL string) (*model.GameItem, error) {
	if !strings.HasPrefix(URL, "https://www.gog-games.to/game/") {
		return nil, fmt.Errorf("invalid url")
	}
	_, slug := path.Split(URL)

	apiUrl := fmt.Sprintf(constant.GOGGamesGameAPIURL, slug)

	token, err := ccs.TurnstileToken(config.Config.CFClearanceScraper.Url, apiUrl, "0x4AAAAAAAfOlgvCKbOdW1zc")
	if err != nil {
		return nil, err
	}
	resp, err := utils.Request().SetHeader("cf-turnstile-response", token).Get(apiUrl)
	if err != nil {
		return nil, err
	}
	data := gameResult{}
	err = json.Unmarshal(resp.Body(), &data)
	if err != nil {
		return nil, err
	}

	name := data.Title

	// find download links
	fileHosters := []string{
		"gofile",
		"fileditch",
		"qiwi",
		"filesfm",
		"pixeldrain",
		"1fichier",
	}
	links := make([]string, 0)
	for _, h := range fileHosters {
		if value, exist := data.Links.Game[h]; exist {
			for _, link := range value.Links {
				links = append(links, link.Link)
			}
		}
		if value, exist := data.Links.Patch[h]; exist {
			for _, link := range value.Links {
				links = append(links, link.Link)
			}
		}
	}

	if len(links) == 0 {
		return nil, fmt.Errorf("no download link found")
	}

	size := uint64(0)
	for _, file := range data.Files.Game {
		s, _ := utils.SizeToBytes(file.Size)
		size += s
	}

	item, err := db.GetGameItemByUrl(URL)
	if err != nil {
		return nil, err
	}
	item.Name = name
	item.RawName = name
	item.Download = strings.Join(links, ",")
	item.Url = URL
	item.Size = utils.BytesToSize(size)
	item.Author = "GOGGames"
	item.Platform = "windows"
	return item, nil
}

func (c *GOGGamesCrawler) Crawl(page int) ([]*model.GameItem, error) {
	resp, err := utils.Request().Get(fmt.Sprintf(constant.GOGGamesURL, page))
	if err != nil {
		return nil, err
	}
	data := searchResult{}
	err = json.Unmarshal(resp.Body(), &data)
	if err != nil {
		return nil, err
	}
	urls := make([]string, 0)
	var updateFlags []string //link+date
	for _, item := range data.Data {
		urls = append(urls, fmt.Sprintf(constant.GOGGamesPageURL, item.Slug))
		updateFlags = append(updateFlags, fmt.Sprintf("%s%s", item.GogURL, item.LastUpdate))
	}
	res := make([]*model.GameItem, 0)
	for i, u := range urls {
		if db.IsGameCrawled(updateFlags[i], "GOGGames") {
			continue
		}
		c.logger.Info("Crawling", zap.String("URL", u))
		item, err := c.CrawlByUrl(u)
		if err != nil {
			c.logger.Warn("Failed to crawl", zap.Error(err), zap.String("URL", u))
			continue
		}
		item.UpdateFlag = updateFlags[i]
		if err := db.SaveGameItem(item); err != nil {
			c.logger.Warn("Failed to save", zap.Error(err), zap.String("URL", u))
			continue
		}
		res = append(res, item)
		if err := OrganizeGameItem(item); err != nil {
			c.logger.Warn("Failed to organize", zap.Error(err), zap.String("URL", u))
			continue
		}
	}
	return res, nil
}

func (c *GOGGamesCrawler) CrawlMulti(pages []int) ([]*model.GameItem, error) {
	res := make([]*model.GameItem, 0)
	for _, page := range pages {
		items, err := c.Crawl(page)
		if err != nil {
			return nil, err
		}
		res = append(res, items...)
	}
	return res, nil
}

func (c *GOGGamesCrawler) CrawlAll() ([]*model.GameItem, error) {
	totalPageNum, err := c.GetTotalPageNum()
	if err != nil {
		return nil, err
	}
	var res []*model.GameItem
	for i := 1; i <= totalPageNum; i++ {
		items, err := c.Crawl(i)
		if err != nil {
			return nil, err
		}
		res = append(res, items...)
	}
	return res, nil
}

func (c *GOGGamesCrawler) GetTotalPageNum() (int, error) {
	resp, err := utils.Request().Get(fmt.Sprintf(constant.GOGGamesURL, 1))
	if err != nil {
		return 0, err
	}
	data := searchResult{}
	err = json.Unmarshal(resp.Body(), &data)
	if err != nil {
		return 0, err
	}
	return data.Meta.LastPage, nil
}

type searchResult struct {
	Data []struct {
		ID          string    `json:"id"`
		Slug        string    `json:"slug"`
		Title       string    `json:"title"`
		Image       string    `json:"image"`
		Background  string    `json:"background"`
		GogURL      string    `json:"gog_url"`
		IsIndev     bool      `json:"is_indev"`
		IsNew       bool      `json:"is_new"`
		IsUpdated   bool      `json:"is_updated"`
		IsQueued    bool      `json:"is_queued"`
		IsUploading bool      `json:"is_uploading"`
		VotedOn     bool      `json:"voted_on"`
		LastUpdate  time.Time `json:"last_update"`
		Md5Filename string    `json:"md5_filename"`
		Infohash    string    `json:"infohash"`
		IsVotable   bool      `json:"is_votable"`
	} `json:"data"`
	Links struct {
		First string `json:"first"`
		Last  string `json:"last"`
		Prev  any    `json:"prev"`
		Next  string `json:"next"`
	} `json:"links"`
	Meta struct {
		CurrentPage int `json:"current_page"`
		From        int `json:"from"`
		LastPage    int `json:"last_page"`
		Links       []struct {
			URL    any    `json:"url"`
			Label  string `json:"label"`
			Active bool   `json:"active"`
		} `json:"links"`
		Path    string `json:"path"`
		PerPage int    `json:"per_page"`
		To      int    `json:"to"`
		Total   int    `json:"total"`
	} `json:"meta"`
}

type gameResult struct {
	ID          string    `json:"id"`
	Slug        string    `json:"slug"`
	Title       string    `json:"title"`
	Image       string    `json:"image"`
	Background  string    `json:"background"`
	GogURL      string    `json:"gog_url"`
	IsIndev     bool      `json:"is_indev"`
	IsNew       bool      `json:"is_new"`
	IsUpdated   bool      `json:"is_updated"`
	IsQueued    bool      `json:"is_queued"`
	IsUploading bool      `json:"is_uploading"`
	VotedOn     bool      `json:"voted_on"`
	LastUpdate  time.Time `json:"last_update"`
	Md5Filename string    `json:"md5_filename"`
	Infohash    string    `json:"infohash"`
	Links       struct {
		Goodie struct {
			OneFichier struct {
				ID    string `json:"id"`
				Name  string `json:"name"`
				Links []struct {
					Label string `json:"label"`
					Link  string `json:"link"`
				} `json:"links"`
			} `json:"1fichier"`
			Vikingfile struct {
				ID    string `json:"id"`
				Name  string `json:"name"`
				Links []struct {
					Label string `json:"label"`
					Link  string `json:"link"`
				} `json:"links"`
			} `json:"vikingfile"`
			Pixeldrain struct {
				ID    string `json:"id"`
				Name  string `json:"name"`
				Links []struct {
					Label string `json:"label"`
					Link  string `json:"link"`
				} `json:"links"`
			} `json:"pixeldrain"`
			Gofile struct {
				ID    string `json:"id"`
				Name  string `json:"name"`
				Links []struct {
					Label string `json:"label"`
					Link  string `json:"link"`
				} `json:"links"`
			} `json:"gofile"`
		} `json:"goodie"`
		Game map[string]struct {
			ID    string `json:"id"`
			Name  string `json:"name"`
			Links []struct {
				Label string `json:"label"`
				Link  string `json:"link"`
			} `json:"links"`
		} `json:"game"`
		Patch map[string]struct {
			ID    string `json:"id"`
			Name  string `json:"name"`
			Links []struct {
				Label string `json:"label"`
				Link  string `json:"link"`
			} `json:"links"`
		} `json:"patch"`
	} `json:"links"`
	Files struct {
		Game []struct {
			Name string `json:"name"`
			Type string `json:"type"`
			Size string `json:"size"`
		} `json:"game"`
		Goodie []struct {
			Name string `json:"name"`
			Type string `json:"type"`
			Size string `json:"size"`
		} `json:"goodie"`
		Patch []struct {
			Name string `json:"name"`
			Type string `json:"type"`
			Size string `json:"size"`
		} `json:"patch"`
	} `json:"files"`
	IsVotable bool `json:"is_votable"`
}