breakpilot-lehrer/edu-search-service/internal/search/search.go

package search

import (
	"context"
	"encoding/json"
	"fmt"
	"strings"

	"github.com/opensearch-project/opensearch-go/v2"
	"github.com/opensearch-project/opensearch-go/v2/opensearchapi"
)

// SearchRequest represents an API search request
type SearchRequest struct {
	Query   string            `json:"q"`
	Mode    string            `json:"mode"` // keyword, semantic, hybrid
	Limit   int               `json:"limit"`
	Offset  int               `json:"offset"`
	Filters SearchFilters     `json:"filters"`
	Rerank  bool              `json:"rerank"`
	Include SearchInclude     `json:"include"`
}

// SearchFilters for narrowing results
type SearchFilters struct {
	Language       []string `json:"language"`
	CountryHint    []string `json:"country_hint"`
	SourceCategory []string `json:"source_category"`
	DocType        []string `json:"doc_type"`
	SchoolLevel    []string `json:"school_level"`
	Subjects       []string `json:"subjects"`
	State          []string `json:"state"`
	MinTrustScore  float64  `json:"min_trust_score"`
	DateFrom       string   `json:"date_from"`
}

// SearchInclude specifies what to include in response
type SearchInclude struct {
	Snippets    bool `json:"snippets"`
	Highlights  bool `json:"highlights"`
	ContentText bool `json:"content_text"`
}

// SearchResult represents a single search result
type SearchResult struct {
	DocID       string   `json:"doc_id"`
	Title       string   `json:"title"`
	URL         string   `json:"url"`
	Domain      string   `json:"domain"`
	Language    string   `json:"language"`
	DocType     string   `json:"doc_type"`
	SchoolLevel string   `json:"school_level"`
	Subjects    []string `json:"subjects"`
	Scores      Scores   `json:"scores"`
	Snippet     string   `json:"snippet,omitempty"`
	Highlights  []string `json:"highlights,omitempty"`
}

// Scores contains all scoring components
type Scores struct {
	BM25     float64 `json:"bm25"`
	Semantic float64 `json:"semantic"`
	Rerank   float64 `json:"rerank"`
	Trust    float64 `json:"trust"`
	Quality  float64 `json:"quality"`
	Final    float64 `json:"final"`
}

// SearchResponse is the API response
type SearchResponse struct {
	QueryID    string         `json:"query_id"`
	Results    []SearchResult `json:"results"`
	Pagination Pagination     `json:"pagination"`
}

// Pagination info
type Pagination struct {
	Limit         int `json:"limit"`
	Offset        int `json:"offset"`
	TotalEstimate int `json:"total_estimate"`
}

// EmbeddingProvider interface for generating embeddings
type EmbeddingProvider interface {
	Embed(ctx context.Context, text string) ([]float32, error)
	IsEnabled() bool
	Dimension() int
}

// Service handles search operations
type Service struct {
	client            *opensearch.Client
	indexName         string
	embeddingProvider EmbeddingProvider
	semanticEnabled   bool
}

// NewService creates a new search service
func NewService(url, username, password, indexName string) (*Service, error) {
	cfg := opensearch.Config{
		Addresses: []string{url},
		Username:  username,
		Password:  password,
	}

	client, err := opensearch.NewClient(cfg)
	if err != nil {
		return nil, err
	}

	return &Service{
		client:          client,
		indexName:       indexName,
		semanticEnabled: false,
	}, nil
}

// SetEmbeddingProvider configures the embedding provider for semantic search
func (s *Service) SetEmbeddingProvider(provider EmbeddingProvider) {
	if provider != nil && provider.IsEnabled() {
		s.embeddingProvider = provider
		s.semanticEnabled = true
	}
}

// IsSemanticEnabled returns true if semantic search is available
func (s *Service) IsSemanticEnabled() bool {
	return s.semanticEnabled && s.embeddingProvider != nil
}

// Search performs a search query
func (s *Service) Search(ctx context.Context, req *SearchRequest) (*SearchResponse, error) {
	// Determine search mode
	mode := req.Mode
	if mode == "" {
		mode = "keyword" // Default to keyword search
	}

	// For semantic/hybrid modes, generate query embedding
	var queryEmbedding []float32
	var embErr error
	if (mode == "semantic" || mode == "hybrid") && s.IsSemanticEnabled() {
		queryEmbedding, embErr = s.embeddingProvider.Embed(ctx, req.Query)
		if embErr != nil {
			// Fall back to keyword search if embedding fails
			mode = "keyword"
		}
	} else if mode == "semantic" || mode == "hybrid" {
		// Semantic requested but not enabled, fall back
		mode = "keyword"
	}

	// Build OpenSearch query based on mode
	var query map[string]interface{}
	switch mode {
	case "semantic":
		query = s.buildSemanticQuery(req, queryEmbedding)
	case "hybrid":
		query = s.buildHybridQuery(req, queryEmbedding)
	default:
		query = s.buildQuery(req)
	}

	queryJSON, err := json.Marshal(query)
	if err != nil {
		return nil, err
	}

	searchReq := opensearchapi.SearchRequest{
		Index: []string{s.indexName},
		Body:  strings.NewReader(string(queryJSON)),
	}

	res, err := searchReq.Do(ctx, s.client)
	if err != nil {
		return nil, err
	}
	defer res.Body.Close()

	// Parse response
	var osResponse struct {
		Hits struct {
			Total struct {
				Value int `json:"value"`
			} `json:"total"`
			Hits []struct {
				ID     string                 `json:"_id"`
				Score  float64                `json:"_score"`
				Source map[string]interface{} `json:"_source"`
				Highlight map[string][]string `json:"highlight,omitempty"`
			} `json:"hits"`
		} `json:"hits"`
	}

	if err := json.NewDecoder(res.Body).Decode(&osResponse); err != nil {
		return nil, err
	}

	// Convert to SearchResults
	results := make([]SearchResult, 0, len(osResponse.Hits.Hits))
	for _, hit := range osResponse.Hits.Hits {
		result := s.hitToResult(hit.Source, hit.Score, hit.Highlight, req.Include)
		results = append(results, result)
	}

	return &SearchResponse{
		QueryID: fmt.Sprintf("q-%d", ctx.Value("request_id")),
		Results: results,
		Pagination: Pagination{
			Limit:         req.Limit,
			Offset:        req.Offset,
			TotalEstimate: osResponse.Hits.Total.Value,
		},
	}, nil
}

// Helper functions
func getString(m map[string]interface{}, key string) string {
	if v, ok := m[key].(string); ok {
		return v
	}
	return ""
}

func getFloat(m map[string]interface{}, key string) float64 {
	if v, ok := m[key].(float64); ok {
		return v
	}
	return 0.0
}

func getStringArray(m map[string]interface{}, key string) []string {
	if v, ok := m[key].([]interface{}); ok {
		result := make([]string, 0, len(v))
		for _, item := range v {
			if s, ok := item.(string); ok {
				result = append(result, s)
			}
		}
		return result
	}
	return nil
}