kafka/internal/engines/braveapi.go

package engines

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

	"github.com/ashie/gosearch/internal/contracts"
)

// BraveEngine implements the SearXNG `braveapi` engine (Brave Web Search API).
//
// Config / gating:
// - BRAVE_API_KEY: required to call Brave
// - BRAVE_ACCESS_TOKEN (optional): if set, the request must include a token
//   that matches the env var (via Authorization Bearer, X-Search-Token,
//   X-Brave-Access-Token, or form field `token`).
type BraveEngine struct {
	client          *http.Client
	apiKey          string
	accessGateToken string
	resultsPerPage int
}

func (e *BraveEngine) Name() string { return "braveapi" }

func (e *BraveEngine) Search(ctx context.Context, req contracts.SearchRequest) (contracts.SearchResponse, error) {
	if e == nil || e.client == nil {
		return contracts.SearchResponse{}, errors.New("brave engine not initialized")
	}

	// Gate / config checks should not be treated as fatal errors; SearXNG
	// treats misconfigured engines as unresponsive.
	if strings.TrimSpace(e.apiKey) == "" {
		return contracts.SearchResponse{
			Query:               req.Query,
			NumberOfResults:    0,
			Results:             []contracts.MainResult{},
			Answers:             []map[string]any{},
			Corrections:         []string{},
			Infoboxes:           []map[string]any{},
			Suggestions:         []string{},
			UnresponsiveEngines: [][2]string{{e.Name(), "missing_api_key"}},
		}, nil
	}

	if gate := strings.TrimSpace(e.accessGateToken); gate != "" {
		if strings.TrimSpace(req.AccessToken) == "" || req.AccessToken != gate {
			return contracts.SearchResponse{
				Query:               req.Query,
				NumberOfResults:    0,
				Results:             []contracts.MainResult{},
				Answers:             []map[string]any{},
				Corrections:         []string{},
				Infoboxes:           []map[string]any{},
				Suggestions:         []string{},
				UnresponsiveEngines: [][2]string{{e.Name(), "unauthorized"}},
			}, nil
		}
	}

	q := strings.TrimSpace(req.Query)
	if q == "" {
		return contracts.SearchResponse{Query: req.Query}, nil
	}

	offset := 0
	if req.Pageno > 1 {
		offset = (req.Pageno - 1) * e.resultsPerPage
	}

	args := url.Values{}
	args.Set("q", q)
	args.Set("count", fmt.Sprintf("%d", e.resultsPerPage))
	args.Set("offset", fmt.Sprintf("%d", offset))

	if req.TimeRange != nil {
		switch *req.TimeRange {
		case "day":
			args.Set("time_range", "past_day")
		case "week":
			args.Set("time_range", "past_week")
		case "month":
			args.Set("time_range", "past_month")
		case "year":
			args.Set("time_range", "past_year")
		}
	}

	// SearXNG's python checks `if params["safesearch"]:` which treats any
	// non-zero (moderate/strict) as strict.
	if req.Safesearch > 0 {
		args.Set("safesearch", "strict")
	}

	endpoint := "https://api.search.brave.com/res/v1/web/search?" + args.Encode()

	httpReq, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
	if err != nil {
		return contracts.SearchResponse{}, err
	}
	httpReq.Header.Set("X-Subscription-Token", e.apiKey)

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

	if resp.StatusCode < 200 || resp.StatusCode >= 300 {
		body, _ := io.ReadAll(io.LimitReader(resp.Body, 16*1024))
		return contracts.SearchResponse{}, fmt.Errorf("brave upstream error: status=%d body=%q", resp.StatusCode, string(body))
	}

	var api struct {
		Web struct {
			Results []struct {
				URL         string `json:"url"`
				Title       string `json:"title"`
				Description string `json:"description"`
				Age         string `json:"age"`
				Thumbnail   struct {
					Src string `json:"src"`
				} `json:"thumbnail"`
			} `json:"results"`
		} `json:"web"`
	}

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

	results := make([]contracts.MainResult, 0, len(api.Web.Results))
	for _, r := range api.Web.Results {
		urlPtr := strings.TrimSpace(r.URL)
		if urlPtr == "" {
			continue
		}
		pub := parseBraveAge(r.Age)

		results = append(results, contracts.MainResult{
			Template: "default.html",
			Title:    r.Title,
			Content:  r.Description,
			URL:      &urlPtr,
			Pubdate:  pub,
			Engine:   e.Name(),
			Score:    0,
			Category: "general",
			Priority: "",
			Positions: nil,
			Engines:  []string{e.Name()},
		})
	}

	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
}

func parseBraveAge(ageRaw string) *string {
	ageRaw = strings.TrimSpace(ageRaw)
	if ageRaw == "" {
		return nil
	}

	// Brave sometimes returns RFC3339-like timestamps for `age`.
	layouts := []string{
		time.RFC3339Nano,
		time.RFC3339,
		"2006-01-02T15:04:05Z07:00",
		"2006-01-02",
	}
	for _, layout := range layouts {
		if t, err := time.Parse(layout, ageRaw); err == nil {
			s := t.Format("2006-01-02 15:04:05-0700")
			return &s
		}
	}
	return nil
}