initial commit

lib/extract.go

@@ -0,0 +1,77 @@
+package lib
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+
+	"github.com/pdfcpu/pdfcpu/pkg/api"
+	"github.com/pdfcpu/pdfcpu/pkg/pdfcpu/model"
+)
+
+func ExtractFromPdf(inputFilePath string, outputFilePath string) error {
+	logger := GetDebugLogger()
+	logger.Log(fmt.Sprintf("enter ExtractFromPdf(%s, %s)", inputFilePath, outputFilePath))
+	defer func() {
+		logger.Log("leave extractFromPdf")
+	}()
+
+	if inputFilePath == "" {
+		return errors.New("no input file specified")
+	}
+
+	if outputFilePath == "" {
+		outputFilePath = strings.TrimSuffix(inputFilePath, filepath.Ext(inputFilePath)) + ".xml"
+	}
+
+	logger.Log(fmt.Sprintf("pdfInput: '%s' output: '%s'", inputFilePath, outputFilePath))
+
+	inputFile, err := os.Open(inputFilePath)
+	if err != nil {
+		return err
+	}
+	defer inputFile.Close()
+
+	var (
+		attachments []model.Attachment
+		rawData     []byte
+		ctx         *model.Context
+		attachment  *model.Attachment
+	)
+
+	ctx, err = api.ReadContextFile(inputFilePath)
+	if err != nil {
+		return err
+	}
+	attachments, err = api.Attachments(inputFile, model.NewDefaultConfiguration())
+	if err != nil {
+		return err
+	}
+
+	attachment, err = ctx.ExtractAttachment(attachments[0])
+	if err != nil {
+		return err
+	}
+
+	rawData, err = io.ReadAll(attachment)
+	if err != nil {
+		return err
+	}
+
+	var outputFile *os.File
+	outputFile, err = os.Create(outputFilePath)
+	if err != nil {
+		return err
+	}
+	defer outputFile.Close()
+
+	_, err = outputFile.Write(rawData)
+	if err != nil {
+		return err
+	}
+
+	return nil
+}