samsa/internal/engines/bing_images.go

// kafka — a privacy-respecting metasearch engine
// Copyright (C) 2026-present metamorphosis-dev
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.

package engines

import (
	"context"
	"encoding/xml"
	"errors"
	"fmt"
	"io"
	"net/http"
	"net/url"
	"strings"

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

// BingImagesEngine searches Bing Images via their public RSS endpoint.
type BingImagesEngine struct {
	client *http.Client
}

func (e *BingImagesEngine) Name() string { return "bing_images" }

func (e *BingImagesEngine) Search(ctx context.Context, req contracts.SearchRequest) (contracts.SearchResponse, error) {
	if e == nil || e.client == nil {
		return contracts.SearchResponse{}, errors.New("bing_images engine not initialized")
	}
	q := strings.TrimSpace(req.Query)
	if q == "" {
		return contracts.SearchResponse{Query: req.Query}, nil
	}

	offset := (req.Pageno - 1) * 10
	endpoint := fmt.Sprintf(
		"https://www.bing.com/images/search?q=%s&count=10&offset=%d&format=rss",
		url.QueryEscape(q),
		offset,
	)

	httpReq, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
	if err != nil {
		return contracts.SearchResponse{}, err
	}
	httpReq.Header.Set("User-Agent", "kafka/0.1 (compatible; +https://git.ashisgreat.xyz/penal-colony/kafka)")

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

	if resp.StatusCode != http.StatusOK {
		io.Copy(io.Discard, io.LimitReader(resp.Body, 4096))
		return contracts.SearchResponse{}, fmt.Errorf("bing_images upstream error: status %d", resp.StatusCode)
	}

	return parseBingImagesRSS(resp.Body, req.Query)
}

// parseBingImagesRSS parses Bing's RSS image search results.
// The description field contains HTML with an <img> tag whose src is the
// thumbnail and whose enclosing <a> tag links to the source page.
func parseBingImagesRSS(r io.Reader, query string) (contracts.SearchResponse, error) {
	type bingImageItem struct {
		Title   string `xml:"title"`
		Link    string `xml:"link"`
		Descrip string `xml:"description"`
	}

	type rssFeed struct {
		XMLName xml.Name `xml:"rss"`
		Channel struct {
			Items []bingImageItem `xml:"item"`
		} `xml:"channel"`
	}

	var rss rssFeed
	if err := xml.NewDecoder(r).Decode(&rss); err != nil {
		return contracts.SearchResponse{}, fmt.Errorf("bing_images RSS parse error: %w", err)
	}

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

		// Extract thumbnail URL from the description HTML.
		thumbnail := extractImgSrc(item.Descrip)
		content := stripHTML(item.Descrip)

		linkPtr := item.Link
		results = append(results, contracts.MainResult{
			Template:  "images",
			Title:     item.Title,
			Content:   content,
			URL:       &linkPtr,
			Thumbnail: thumbnail,
			Engine:    "bing_images",
			Score:     0,
			Category:  "images",
			Engines:   []string{"bing_images"},
		})
	}

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