query-orchestration/internal/document/clean/content.go

package documentclean

import (
	"bytes"
	"context"
	"errors"
	"io"

	"github.com/aws/aws-sdk-go-v2/service/s3"
)

type Content interface {
	isCorrupt(context.Context) *InvalidDocumentReason
}

func (s *Service) getContent(ctx context.Context, params *CleanParams, mimeType MimeType) (Content, error) {
	resp, err := s.cfg.GetStoreClient().GetObject(ctx, &s3.GetObjectInput{
		Bucket:  &params.Location.Bucket,
		Key:     &params.Location.Key,
		IfMatch: &params.Hash,
	})
	if err != nil {
		return nil, err
	}
	defer resp.Body.Close()

	seek, err := ReaderToSeeker(resp.Body)
	if err != nil {
		return nil, err
	}

	var content Content
	switch mimeType {
	case MimeTypePDF:
		content = NewPDF(seek)
	default:
		return nil, errors.New("invalid mimetype")
	}

	return content, nil
}

func ReaderToSeeker(readCloser io.ReadCloser) (io.ReadSeeker, error) {
	data, err := io.ReadAll(readCloser)
	if err != nil {
		return nil, err
	}

	readCloser.Close()

	return bytes.NewReader(data), nil
}