Merge branch 'development' of https://github.com/unidoc/unipdf into cmap

model/annotations.go

@@ -343,6 +343,12 @@ type PdfAnnotationWidget struct {
 	processing bool // Used in ToPdfObject serialization to avoid infinite loops for merged-in annots.
 }
 
+// Field returns the parent form field of the widget annotation, if one exists.
+// NOTE: the method returns nil if the parent form field has not been parsed.
+func (widget *PdfAnnotationWidget) Field() *PdfField {
+	return widget.parent
+}
+
 // PdfAnnotationWatermark represents Watermark annotations.
 // (Section 12.5.6.22).
 type PdfAnnotationWatermark struct {

model/form_test.go

@@ -6,6 +6,7 @@
 package model
 
 import (
+	"os"
 	"testing"
 
 	"github.com/stretchr/testify/require"
@@ -160,3 +161,18 @@ endobj
 	_ = raw
 	t.Skip("Not implemented yet")
 }
+
+func TestRepairAcroForm(t *testing.T) {
+	f, err := os.Open("./testdata/OoPdfFormExample.pdf")
+	require.NoError(t, err)
+	defer f.Close()
+
+	reader, err := NewPdfReader(f)
+	require.NoError(t, err)
+
+	original := *reader.AcroForm.Fields
+	reader.AcroForm.Fields = nil
+	require.NoError(t, reader.RepairAcroForm(nil))
+	repaired := *reader.AcroForm.Fields
+	require.ElementsMatch(t, original, repaired)
+}

model/reader.go

@@ -507,6 +507,69 @@ func (r *PdfReader) GetOutlines() (*Outline, error) {
 	return outline, nil
 }
 
+// AcroFormRepairOptions contains options for rebuilding the AcroForm.
+type AcroFormRepairOptions struct {
+}
+
+// RepairAcroForm attempts to rebuild the AcroForm fields using the widget
+// annotations present in the document pages. Pass nil for the opts parameter
+// in order to use the default options.
+// NOTE: Currently, the opts parameter is declared in order to enable adding
+// future options, but passing nil will always result in the default options
+// being used.
+func (r *PdfReader) RepairAcroForm(opts *AcroFormRepairOptions) error {
+	var fields []*PdfField
+	fieldCache := map[*core.PdfIndirectObject]struct{}{}
+	for _, page := range r.PageList {
+		annotations, err := page.GetAnnotations()
+		if err != nil {
+			return err
+		}
+
+		for _, annotation := range annotations {
+			var field *PdfField
+			switch t := annotation.GetContext().(type) {
+			case *PdfAnnotationWidget:
+				if t.parent != nil {
+					field = t.parent
+					break
+				}
+				if parentObj, ok := core.GetIndirect(t.Parent); ok {
+					field, err = r.newPdfFieldFromIndirectObject(parentObj, nil)
+					if err == nil {
+						break
+					}
+					common.Log.Debug("WARN: could not parse form field %+v: %v", parentObj, err)
+				}
+				if t.container != nil {
+					field, err = r.newPdfFieldFromIndirectObject(t.container, nil)
+					if err == nil {
+						break
+					}
+					common.Log.Debug("WARN: could not parse form field %+v: %v", t.container, err)
+				}
+			}
+			if field == nil {
+				continue
+			}
+			if _, ok := fieldCache[field.container]; ok {
+				continue
+			}
+			fieldCache[field.container] = struct{}{}
+			fields = append(fields, field)
+		}
+	}
+
+	if len(fields) == 0 {
+		return nil
+	}
+	if r.AcroForm == nil {
+		r.AcroForm = NewPdfAcroForm()
+	}
+	r.AcroForm.Fields = &fields
+	return nil
+}
+
 // loadForms loads the AcroForm.
 func (r *PdfReader) loadForms() (*PdfAcroForm, error) {
 	if r.parser.GetCrypter() != nil && !r.parser.IsAuthenticated() {