searcherside/core/services/bleve_indexer.go

package services

import (
	"context"
	"encoding/json"
	"fmt"
	"io"
	"os"
	"path/filepath"

	"github.com/blevesearch/bleve/v2"
	"github.com/blevesearch/bleve/v2/mapping"

	"code.icb4dc0.de/prskr/searcherside/core/ports"
	"code.icb4dc0.de/prskr/searcherside/infrastructure/archive"
)

var _ ports.Indexer = (*BleveIndexer)(nil)

type BleveIndexer struct {
	DataDirectory string
}

func (b BleveIndexer) IngestIndex(ctx context.Context, request ports.IngestIndexRequest) (res ports.IngestIndexResult, err error) {
	res.Type = ports.IndexTypeBleve

	indexMapping := bleve.NewIndexMapping()
	indexMapping.AddDocumentMapping("page", pageDocumentMapping())

	if err := os.MkdirAll(b.DataDirectory, 0o750); err != nil {
		return res, fmt.Errorf("failed to create data directory: %w", err)
	}

	indexDirName := fmt.Sprintf("%s.bleve", request.Hash)
	res.Path = filepath.Join(b.DataDirectory, indexDirName)

	if info, err := os.Stat(res.Path); err == nil && info.IsDir() {
		return res, nil
	}

	index, err := bleve.New(res.Path, indexMapping)
	if err != nil {
		return res, err
	}

	indexErr := archive.WalkZipFile(request.FilePath, func(name string, r io.Reader) error {
		var page pageToIndex
		if err := json.NewDecoder(r).Decode(&page); err != nil {
			return fmt.Errorf("failed to decode page: %w", err)
		}

		return index.Index(page.ObjectID, page)
	})

	if indexErr != nil {
		return res, fmt.Errorf("failed to index pages: %w", indexErr)
	}

	if closeErr := index.Close(); closeErr != nil {
		return res, fmt.Errorf("failed to close index: %w", closeErr)
	}

	return res, nil
}

var _ mapping.Classifier = (*pageToIndex)(nil)

type pageToIndex struct {
	ObjectID        string `json:"objectID"`
	MainTitle       string `json:"mainTitle"`
	PageTitle       string `json:"pageTitle"`
	Url             string `json:"url"`
	Headings        string `json:"headings"`
	Content         string `json:"content"`
	MetaDescription string `json:"metaDescription"`
	Breadcrumbs     string `json:"breadcrumbs"`
}

func (b pageToIndex) Type() string {
	return "page"
}

func pageDocumentMapping() *mapping.DocumentMapping {
	docMapping := mapping.NewDocumentMapping()

	storeOnlyMapping := &mapping.FieldMapping{
		Type:  "text",
		Store: true,
	}

	docMapping.AddFieldMappingsAt("objectID", storeOnlyMapping)
	docMapping.AddFieldMappingsAt("url", storeOnlyMapping)
	docMapping.AddFieldMappingsAt("breadcrumbs", storeOnlyMapping)

	return docMapping
}