/**
* Applies this transform to a PDF page.
*
* <p>Using the configured PDF operator processors, a PDF stream engine ({@link PDFStreamEngine}
* parent class) iterates through the page's token stream, aggregating operands and invoking the
* appropriate PDF operator processor for each PDF operator encountered. An output list of tokens
* is created. At the end of the iteration, if any PDF operator processor has indicated to this
* transform that it has changed the output list compared to the original stream, the PDF page's
* contents are replaced with those of the output list (otherwise the PDF page is unchanged).
*
* @param pdfPage A PDF page (belonging to the PDF document).
* @throws PageTransformException if the output list stack does not have exactly one output list
* at the end of the transform.
* @throws IOException if any other processing error occurs.
* @see PDFStreamEngine#processStream
* @see #reset
*/
public synchronized boolean transform(PdfPage pdfPage) throws IOException {
logger.debug3("Begin page stream transform");
// Iterate over stream
reset();
this.currentPdfPage = pdfPage;
processStream(pdfPage.getPdPage(), pdfPage.findResources(), pdfPage.getContentStream());
// Sanity check
if (listStack.size() != 1) {
String logMessage =
"Split/merge mismatch: after processing stream, list stack has size " + listStack.size();
logger.error(logMessage);
throw new PageTransformException(logMessage);
}
writeResult(pdfPage);
logger.debug2("Page stream transform result: " + atLeastOneChange);
return atLeastOneChange;
}
/** Pushes an empty list onto the output list stack. */
public synchronized void splitOutputList() {
logger.debug3("Splitting");
listStack.push(new ArrayList());
}
/**
* Resets the state of this transform; this clears the stack and pushes an empty list onto it, and
* clears the flag that indicates there has been at least one change.
*
* <p>This method is called at the beginning of {@link #transform}.
*
* <p>This method is <em>not</em> called at the end of {@link #transform}. Though the contents of
* the resulting output list may have been written to the PDF page already, clients may wish to
* inspect the final result without having to re-parse the token stream of the PDF page. However,
* the result list may be quite large and the stack will hold on to it until the next call to this
* method, which may represent a memory issue. Thus clients <em>may</em> call {@link #reset} after
* a call to {@link #transform} to clear the output list stack.
*/
public synchronized void reset() {
logger.debug3("Resetting the page stream transform");
atLeastOneChange = false;
listStack.clear();
listStack.push(new ArrayList()); // FIXME: initial capacity?
}
/**
* Pops the output list currently at the top of the output list stack and discards it, and appends
* the given replacement list to the output list immediately underneath.
*
* @param replacement A list of tokens to be appended to the output list currently immediately
* under the top of the stack.
* @throws EmptyStackException if there is currently only one output list on the stack.
*/
public synchronized void mergeOutputList(List replacement) {
logger.debug3("Merging with replacement");
listStack.pop(); // discard result
List newTop = (List) listStack.peek();
newTop.addAll(replacement);
}
/**
* Pops the output list currently at the top of the output list stack and appends it to the end of
* the output list immediately underneath.
*
* @throws EmptyStackException if there is currently only one output list on the stack.
* @see #mergeOutputList(List)
*/
public synchronized void mergeOutputList() {
logger.debug3("Merging");
List oldTop = (List) listStack.pop();
List newTop = (List) listStack.peek();
newTop.addAll(oldTop);
}
/**
* Gets the output list currently at the top of the output list stack.
*
* @return The list of tokens currently at the top of the list stack.
* @see #splitOutputList
* @see #mergeOutputList()
* @see #mergeOutputList(List)
*/
public synchronized List getOutputList() {
return (List) listStack.peek();
}