org.sejda.impl.sambox.component.AcroFormsMerger Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of sejda-sambox Show documentation
Show all versions of sejda-sambox Show documentation
Package containing tasks implemented using sambox.
/*
* Created on 09 set 2015
* Copyright 2015 by Andrea Vacondio ([email protected]).
* This file is part of Sejda.
*
* Sejda is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* Sejda is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with Sejda. If not, see , PDTerminalField> createOrReuseTerminalField = (
PDTerminalField existing, LookupTable fieldsLookup) -> {
PDField previouslyCreated = ofNullable(getMergedField(existing.getFullyQualifiedName()))
.orElseGet(() -> fieldsLookup.lookup(existing));
boolean shouldCreateNew = isNull(previouslyCreated);
boolean shouldCreateNewAndRename = previouslyCreated != null &&
(
// different types (eg: checkbox vs text)
!previouslyCreated.getClass().equals(existing.getClass()) ||
// different values (eg: john vs jack)
!previouslyCreated.getValueAsString().equals(existing.getValueAsString())
);
if (shouldCreateNew || shouldCreateNewAndRename) {
previouslyCreated = PDFieldFactory.createFieldAddingChildToParent(this.form,
existing.getCOSObject().duplicate(),
(PDNonTerminalField) fieldsLookup.lookup(existing.getParent()));
if(shouldCreateNewAndRename) {
LOG.warn("Cannot merge terminal field because another field with the same name but different value already exists: {}",
existing.getFullyQualifiedName());
previouslyCreated.setPartialName(String.format("%s%s%d", removeDotsIfAny(existing.getPartialName()),
random, ++counter));
}
previouslyCreated.getCOSObject().removeItem(COSName.KIDS);
fieldsLookup.addLookupEntry(existing, previouslyCreated);
}
if (!previouslyCreated.isTerminal()) {
LOG.warn("Cannot merge terminal field because a non terminal field with the same name already exists: {}",
existing.getFullyQualifiedName());
return null;
}
return (PDTerminalField) previouslyCreated;
};
private final BiFunction, PDTerminalField> createRenamingTerminalField = (
PDTerminalField existing, LookupTable fieldsLookup) -> {
PDTerminalField newField = (PDTerminalField) PDFieldFactory.createFieldAddingChildToParent(this.form,
existing.getCOSObject().duplicate(), (PDNonTerminalField) fieldsLookup.lookup(existing.getParent()));
if (nonNull(getMergedField(existing.getFullyQualifiedName())) || fieldsLookup.hasLookupFor(existing)) {
String partialName = removeDotsIfAny(existing.getPartialName());
newField.setPartialName(String.format("%s%s%d", partialName, random, ++counter));
LOG.info("Existing terminal field renamed from {} to {}", partialName, newField.getPartialName());
}
newField.getCOSObject().removeItem(COSName.KIDS);
fieldsLookup.addLookupEntry(existing, newField);
return newField;
};
// dots are not allowed in the partial name, but still broken documents may exist
private static String removeDotsIfAny(String s) {
if(s == null) {
return null;
}
return s.replace(".", "");
}
private final BiConsumer> createOrReuseNonTerminalField = (PDField existing,
LookupTable fieldsLookup) -> {
// do we aready have a lookup for this?
if (!fieldsLookup.hasLookupFor(existing)) {
PDField mergedField = getMergedField(existing.getFullyQualifiedName());
// we don't have a lookup but do we have a merged field with the same name (from a previous document)?
if (isNull(mergedField)) {
mergedField = PDFieldFactory.createFieldAddingChildToParent(this.form,
existing.getCOSObject().duplicate(),
(PDNonTerminalField) fieldsLookup.lookup(existing.getParent()));
mergedField.getCOSObject().removeItem(COSName.KIDS);
} else if (mergedField.isTerminal()) {
mergedField = PDFieldFactory.createFieldAddingChildToParent(this.form,
existing.getCOSObject().duplicate(),
(PDNonTerminalField) fieldsLookup.lookup(existing.getParent()));
mergedField.getCOSObject().removeItem(COSName.KIDS);
mergedField.setPartialName(String.format("%s%s%d", removeDotsIfAny(existing.getPartialName()),
random, ++counter));
LOG.warn("Cannot reuse merged terminal field {} as a non terminal field, renaming it to {}",
existing.getPartialName(), mergedField.getPartialName());
}
fieldsLookup.addLookupEntry(existing, mergedField);
}
};
private PDField getMergedField(String fullyQualifiedName) {
return ofNullable(fullyQualifiedName).map(form::getField).orElse(null);
}
private final BiConsumer> createRenamingNonTerminalField = (PDField field,
LookupTable fieldsLookup) -> {
PDField newField = PDFieldFactory.createFieldAddingChildToParent(this.form, field.getCOSObject().duplicate(),
(PDNonTerminalField) fieldsLookup.lookup(field.getParent()));
if (getMergedField(field.getFullyQualifiedName()) != null || fieldsLookup.hasLookupFor(field)) {
newField.setPartialName(String.format("%s%s%d", removeDotsIfAny(field.getPartialName()), random, ++counter));
LOG.info("Existing non terminal field renamed from {} to {}", field.getPartialName(),
newField.getPartialName());
}
newField.getCOSObject().removeItem(COSName.KIDS);
fieldsLookup.addLookupEntry(field, newField);
};
public AcroFormsMerger(AcroFormPolicy policy, PDDocument destination) {
this.policy = policy;
this.form = new PDAcroForm(destination);
}
/**
* Merge the original form to the current one, considering only fields whose widgets are available in the given lookup table.
*
* @param originalForm
* the form to merge
* @param annotationsLookup
* lookup for relevant annotations
*/
public void mergeForm(PDAcroForm originalForm, LookupTable annotationsLookup) {
if (nonNull(originalForm)) {
if (originalForm.hasXFA()) {
LOG.warn("The AcroForm has XFA resources which will be ignored");
}
LOG.debug("Merging acroforms with policy {}", policy);
switch (policy) {
case MERGE_RENAMING_EXISTING_FIELDS:
updateForm(originalForm, annotationsLookup, createRenamingTerminalField,
createRenamingNonTerminalField);
break;
case MERGE:
updateForm(originalForm, annotationsLookup, createOrReuseTerminalField, createOrReuseNonTerminalField);
break;
case FLATTEN:
updateForm(originalForm, annotationsLookup, createRenamingTerminalField,
createRenamingNonTerminalField);
flatten();
break;
default:
LOG.debug("Discarding acroform");
}
} else {
LOG.debug("Skipped acroform merge, nothing to merge");
}
}
/**
* For each new widget annotation in the lookup table removes all the Field keys.
*
* @param annotations
*/
private void removeFieldKeysFromWidgets(Collection annotations) {
annotations.stream().map(PDAnnotation::getCOSObject).forEach(a -> {
Arrays.stream(FIELD_KEYS).forEach(key -> {
// special case: if policy is MERGE_RENAME, then we don't remove /Q from the widget, to prevent Mac Preview
// from losing field formatting in the merged document
if (policy == AcroFormPolicy.MERGE_RENAMING_EXISTING_FIELDS) {
if (key == Q) {
return;
}
}
a.removeItem(key);
});
// if multiple kids we preserve their /DA even if Widget shouldn't have DA according to specs because if there is one Acrobat honors it.
// if only one kid then /DA is in the field. see for a live example PDFBOX-3687
// special case: if policy is MERGE_RENAME, then we don't remove /DA from the Widget, to prevent Mac Preview
// from losing field formatting in the merged document
if (annotations.size() == 1) {
if (policy != AcroFormPolicy.MERGE_RENAMING_EXISTING_FIELDS) {
a.removeItem(COSName.DA);
}
}
});
LOG.trace("Removed fields keys from widget annotations");
}
private void updateForm(PDAcroForm originalForm, LookupTable annotationsLookup,
BiFunction, PDTerminalField> getTerminalField,
BiConsumer> createNonTerminalField) {
AcroFormUtils.mergeDefaults(originalForm, form);
LookupTable fieldsLookup = new LookupTable<>();
Set allRelevantWidgets = annotationsLookup.keys().stream()
.filter(a -> a instanceof PDAnnotationWidget).map(a -> (PDAnnotationWidget) a).collect(toSet());
Set rootFields = new HashSet<>();
// it must be a pre order visit because we have to process non terminal first otherwise terminal ones won't get a parent
// every widget we meet is removed from the allRelevantWidgets so we can identify widgets not referenced by the originalForm
originalForm.getFieldTree().stream().forEach(
f -> mergeField(f, annotationsLookup, getTerminalField, createNonTerminalField, fieldsLookup,
of(allRelevantWidgets::remove)));
// keep track of the root fields
originalForm.getFields().stream().map(fieldsLookup::lookup).filter(Objects::nonNull).forEach(rootFields::add);
if (!allRelevantWidgets.isEmpty()) {
LOG.info("Found relevant widget annotations ({}) not linked to the form", allRelevantWidgets.size());
// we process those widget annotations not referenced by the originalForm acroform (ex. empty fields array)
PDAcroForm dummy = new PDAcroForm(null);
allRelevantWidgets.forEach(w -> {
COSDictionary currentField = w.getCOSObject();
// if there's a hierarchy we walk up to find the root
while (nonNull(currentField.getDictionaryObject(COSName.PARENT, COSDictionary.class))) {
currentField = currentField.getDictionaryObject(COSName.PARENT, COSDictionary.class);
}
// we add it as root to a dummy form so we can reuse its tree visit logic
dummy.addFields(List.of(createField(originalForm, currentField, null)));
});
dummy.getFieldTree().stream().forEach(f -> mergeField(f, annotationsLookup, getTerminalField,
createNonTerminalField, fieldsLookup, empty()));
// keep track of the root fields
dummy.getFields().stream().map(fieldsLookup::lookup).filter(Objects::nonNull).forEach(rootFields::add);
}
List currentRoots = this.form.getFields();
// add only if not there already
this.form.addFields(rootFields.stream().filter(f -> !currentRoots.contains(f)).collect(toList()));
mergeCalculationOrder(originalForm, fieldsLookup);
}
private void mergeField(PDField field, LookupTable annotationsLookup,
BiFunction, PDTerminalField> getTerminalField,
BiConsumer> createNonTerminalField, LookupTable fieldsLookup,
Optional> onProcessedWidget) {
if (!field.isTerminal()) {
createNonTerminalField.accept(field, fieldsLookup);
} else {
List relevantWidgets = findMappedWidgetsFor((PDTerminalField) field, annotationsLookup);
if (!relevantWidgets.isEmpty()) {
PDTerminalField terminalField = getTerminalField.apply((PDTerminalField) field, fieldsLookup);
if (nonNull(terminalField)) {
removeFieldKeysFromWidgets(relevantWidgets);
for (PDAnnotationWidget widget : relevantWidgets) {
terminalField.addWidgetIfMissing(widget);
onProcessedWidget.ifPresent(c -> field.getWidgets().forEach(c));
}
terminalField.getCOSObject().removeItems(WIDGET_KEYS);
}
} else {
LOG.debug("Discarded not relevant field {}", field.getFullyQualifiedName());
}
}
}
private void mergeCalculationOrder(PDAcroForm originalForm, LookupTable fieldsLookup) {
List co = originalForm.getCalculationOrder().stream().map(fieldsLookup::lookup)
.filter(Objects::nonNull).toList();
if (co.size() > 0) {
COSArray formCo = ofNullable(
this.form.getCOSObject().getDictionaryObject(COSName.CO, COSArray.class)).orElseGet(COSArray::new);
for (PDField field : co) {
formCo.add(field);
}
this.form.setCalculationOrder(formCo);
}
}
/**
* @param field
* @param annotationsLookup
* @return the list of relevant widgets for the given field.
*/
private List findMappedWidgetsFor(PDTerminalField field,
LookupTable annotationsLookup) {
return field.getWidgets().stream().map(annotationsLookup::lookup).filter(w -> w instanceof PDAnnotationWidget)
.map(w -> (PDAnnotationWidget) w).collect(toList());
}
private void flatten() {
try {
List fields = new ArrayList<>();
for (PDField field : form.getFieldTree()) {
fields.add(field);
if (field instanceof PDVariableText variableText) {
ensureValueCanBeDisplayed(variableText, form.getDocument());
}
}
form.flatten(fields, form.isNeedAppearances());
} catch (IOException | UnsupportedOperationException ex) {
LOG.warn("Failed to flatten form", ex);
}
}
/**
*
* @return Performs some cleanup task on the resulting {@link PDAcroForm} and then returns it
*/
public PDAcroForm getForm() {
for (PDField current : form.getFieldTree()) {
if (!current.isTerminal() && !((PDNonTerminalField) current).hasChildren()) {
LOG.info("Removing non terminal field with no child {}", current.getFullyQualifiedName());
if (nonNull(current.getParent())) {
current.getParent().removeChild(current);
} else {
// it's a root field
form.removeField(current);
}
} else if (clipSignature(current)) {
form.setSignaturesExist(true);
}
}
if (isBlank(form.getDefaultAppearance())) {
form.setDefaultAppearance("/Helv 0 Tf 0 g ");
}
return form;
}
}