PDFBox 是否允许从 AcroForm 中删除一个字段？

Question

我正在使用 Apache PDFBox 2.0.8 并尝试删除一个字段。但是找不到方法，就像我可以用 iText 做的那样：PdfStamper.getAcroFields().removeField("signature3").

我要做什么。最初我有带有 3 个数字签名的模板 PDF。在某些情况下，我只需要 2 个签名，所以在这种情况下，我需要从模板中删除第三个签名。似乎我不能用 PDFBox 做到这一点，我发现的最接近的东西是展平这个领域，但问题是如果展平特定的 PDField（不是整个表格，而只是一个领域）——所有其他签名都失去了它们的功能，看起来就像他们也变得扁平一样。这是执行此操作的代码：

PDDocument document = PDDocument.load(file);
PDDocumentCatalog documentCatalog = document.getDocumentCatalog();
PDAcroForm acroForm = documentCatalog.getAcroForm();

List<PDField> flattenList = new ArrayList<>();
for (PDField field : acroForm.getFieldTree()) {
    if (field instanceof PDSignatureField && "signature3".equals(field.getFullyQualifiedName())) {
        flattenList.add(field);
    }
}

acroForm.flatten(flattenList, true);
document.save(dest);        
document.close();

Answer 1

正如 Tilman 在评论中提到的，PDFBox 没有从字段树中删除字段的方法。尽管如此，它有操作底层 PDF 结构的方法，因此可以自己编写这样的方法，例如像这样：

PDField removeField(PDDocument document, String fullFieldName) throws IOException {
    PDDocumentCatalog documentCatalog = document.getDocumentCatalog();
    PDAcroForm acroForm = documentCatalog.getAcroForm();

    if (acroForm == null) {
        System.out.println("No form defined.");
        return null;
    }

    PDField targetField = null;

    for (PDField field : acroForm.getFieldTree()) {
        if (fullFieldName.equals(field.getFullyQualifiedName())) {
            targetField = field;
            break;
        }
    }
    if (targetField == null) {
        System.out.println("Form does not contain field with given name.");
        return null;
    }

    PDNonTerminalField parentField = targetField.getParent();
    if (parentField != null) {
        List<PDField> childFields = parentField.getChildren();
        boolean removed = false;
        for (PDField field : childFields)
        {
            if (field.getCOSObject().equals(targetField.getCOSObject())) {
                removed = childFields.remove(field);
                parentField.setChildren(childFields);
                break;
            }
        }
        if (!removed)
            System.out.println("Inconsistent form definition: Parent field does not reference the target field.");
    } else {
        List<PDField> rootFields = acroForm.getFields();
        boolean removed = false;
        for (PDField field : rootFields)
        {
            if (field.getCOSObject().equals(targetField.getCOSObject())) {
                removed = rootFields.remove(field);
                break;
            }
        }
        if (!removed)
            System.out.println("Inconsistent form definition: Root fields do not include the target field.");
    }

    removeWidgets(targetField);

    return targetField;
}

void removeWidgets(PDField targetField) throws IOException {
    if (targetField instanceof PDTerminalField) {
        List<PDAnnotationWidget> widgets = ((PDTerminalField)targetField).getWidgets();
        for (PDAnnotationWidget widget : widgets) {
            PDPage page = widget.getPage();
            if (page != null) {
                List<PDAnnotation> annotations = page.getAnnotations();
                boolean removed = false;
                for (PDAnnotation annotation : annotations) {
                    if (annotation.getCOSObject().equals(widget.getCOSObject()))
                    {
                        removed = annotations.remove(annotation);
                        break;
                    }
                }
                if (!removed)
                    System.out.println("Inconsistent annotation definition: Page annotations do not include the target widget.");
            } else {
                System.out.println("Widget annotation does not have an associated page; cannot remove widget.");
                // TODO: In this case iterate all pages and try to find and remove widget in all of them
            }
        }
    } else if (targetField instanceof PDNonTerminalField) {
        List<PDField> childFields = ((PDNonTerminalField)targetField).getChildren();
        for (PDField field : childFields)
            removeWidgets(field);
    } else {
        System.out.println("Target field is neither terminal nor non-terminal; cannot remove widgets.");
    }
}

(RemoveField 辅助方法 removeField 和 removeWidgets)

可以像这样将其应用于文档和字段：

PDDocument document = PDDocument.load(SOURCE_PDF);

PDField field = removeField(document, "Signature1");
Assert.assertNotNull("Field not found", field);

document.save(TARGET_PDF);        
document.close();

(RemoveField 测试 testRemoveInvisibleSignature)

PS: 我不确定 PDFBox 实际上缓存了多少表单相关信息。因此，我建议不要在同一个文档操作会话中进一步操作表单信息，至少在没有测试的情况下是这样。

PPS: 您在 removeWidgets 辅助方法中找到了一个 TODO。如果该方法输出 "Widget annotation does not have an associated page; cannot remove widget"，您必须添加缺少的代码。

Answer 2

感谢@mkl，我设法通过使用 pdfbox-3.0.0-RC1 版本的较短实现来做到这一点。在这种情况下，隐藏按钮（选中）：

        var check = (PDPushButton) pdAcroForm.getField(name);
    List<PDField> fields = pdAcroForm.getFields();
    fields.removeIf(x -> x.getCOSObject().equals(check.getCOSObject()));
    pdAcroForm.setFields(fields);
    check.getWidgets().forEach(widget -> widget.setNoView(true));

PDFBox 是否允许从 AcroForm 中删除一个字段？

Does PDFBox allow to remove one field from AcroForm?

java

pdfbox