kafka/internal/engines/youtube.go

package engines

import (
	"context"
	"encoding/json"
	"fmt"
	"io"
	"net/http"
	"net/url"
	"os"
	"strings"
	"time"

	"github.com/metamorphosis-dev/kafka/internal/contracts"
)

type YouTubeEngine struct {
	client  *http.Client
	apiKey  string
	baseURL string
}

func (e *YouTubeEngine) Name() string { return "youtube" }

func (e *YouTubeEngine) Search(ctx context.Context, req contracts.SearchRequest) (contracts.SearchResponse, error) {
	if strings.TrimSpace(req.Query) == "" {
		return contracts.SearchResponse{Query: req.Query}, nil
	}

	if e.apiKey == "" {
		e.apiKey = os.Getenv("YOUTUBE_API_KEY")
	}

	maxResults := 10
	if req.Pageno > 1 {
		maxResults = 20
	}

	u := e.baseURL + "/youtube/v3/search?" + url.Values{
		"part":     {"snippet"},
		"q":        {req.Query},
		"type":     {"video"},
		"maxResults": {fmt.Sprintf("%d", maxResults)},
		"key":      {e.apiKey},
	}.Encode()

	if req.Language != "" && req.Language != "auto" {
		lang := strings.Split(strings.ToLower(req.Language), "-")[0]
		u += "&relevanceLanguage=" + lang
	}

	httpReq, err := http.NewRequestWithContext(ctx, http.MethodGet, u, nil)
	if err != nil {
		return contracts.SearchResponse{}, err
	}

	resp, err := e.client.Do(httpReq)
	if err != nil {
		return contracts.SearchResponse{}, err
	}
	defer resp.Body.Close()

	if resp.StatusCode != http.StatusOK {
		body, _ := io.ReadAll(io.LimitReader(resp.Body, 4096))
		return contracts.SearchResponse{}, fmt.Errorf("youtube api error: status=%d body=%q", resp.StatusCode, string(body))
	}

	var apiResp youtubeSearchResponse
	if err := json.NewDecoder(resp.Body).Decode(&apiResp); err != nil {
		return contracts.SearchResponse{}, err
	}

	if apiResp.Error != nil {
		return contracts.SearchResponse{}, fmt.Errorf("youtube api error: %s", apiResp.Error.Message)
	}

	results := make([]contracts.MainResult, 0, len(apiResp.Items))
	for _, item := range apiResp.Items {
		if item.ID.VideoID == "" {
			continue
		}

		videoURL := "https://www.youtube.com/watch?v=" + item.ID.VideoID
		urlPtr := videoURL

		published := ""
		if item.Snippet.PublishedAt != "" {
			if t, err := time.Parse(time.RFC3339, item.Snippet.PublishedAt); err == nil {
				published = t.Format("Jan 2, 2006")
			}
		}

		content := item.Snippet.Description
		if len(content) > 300 {
			content = content[:300] + "..."
		}
		if published != "" {
			content = "Published " + published + " · " + content
		}

		thumbnail := ""
		if item.Snippet.Thumbnails.High.URL != "" {
			thumbnail = item.Snippet.Thumbnails.High.URL
		} else if item.Snippet.Thumbnails.Medium.URL != "" {
			thumbnail = item.Snippet.Thumbnails.Medium.URL
		}

		results = append(results, contracts.MainResult{
			Template:  "videos.html",
			Title:    item.Snippet.Title,
			URL:       &urlPtr,
			Content:   content,
			Thumbnail: thumbnail,
			Engine:    "youtube",
			Score:     1.0,
			Category:  "videos",
			Engines:   []string{"youtube"},
			Metadata: map[string]any{
				"channel":   item.Snippet.ChannelTitle,
				"video_id": item.Snippet.ResourceID.VideoID,
			},
		})
	}

	return contracts.SearchResponse{
		Query:               req.Query,
		NumberOfResults:     len(results),
		Results:             results,
		Answers:             []map[string]any{},
		Corrections:         []string{},
		Infoboxes:           []map[string]any{},
		Suggestions:         []string{},
		UnresponsiveEngines: [][2]string{},
	}, nil
}

// YouTube API response types.

type youtubeSearchResponse struct {
	Items []youtubeSearchItem `json:"items"`
	PageInfo struct {
		TotalResults   int `json:"totalResults"`
		ResultsPerPage int `json:"resultsPerPage"`
	} `json:"pageInfo"`
	NextPageToken string `json:"nextPageToken"`
	Error         *struct {
		Code    int    `json:"code"`
		Message string `json:"message"`
		Errors  []struct {
			Domain       string `json:"domain"`
			Reason       string `json:"reason"`
			Message      string `json:"message"`
		} `json:"errors"`
	} `json:"error"`
}

type youtubeSearchItem struct {
	ID struct {
		VideoID string `json:"videoId"`
	} `json:"id"`
	Snippet struct {
		PublishedAt string `json:"publishedAt"`
		ChannelID   string `json:"channelId"`
		ChannelTitle string `json:"channelTitle"`
		Title       string `json:"title"`
		Description string `json:"description"`
		Thumbnails  struct {
			Default struct {
				URL string `json:"url"`
			} `json:"default"`
			Medium struct {
				URL string `json:"url"`
			} `json:"medium"`
			High struct {
				URL string `json:"url"`
			} `json:"high"`
		} `json:"thumbnails"`
		ResourceID struct {
			VideoID string `json:"videoId"`
		} `json:"resourceId"`
	} `json:"snippet"`
}