/**
* Returns the annotations to which should be displayed together with their namespace.
*
* <p>This will check the "show_ns" paramter for determining. the annotations to display. It also
* iterates over all nodes of the graph matching the type.
*
* @param input The input for the visualizer.
* @param types Which types of nodes to include
* @return
*/
public static Set<String> computeDisplayedNamespace(
VisualizerInput input, List<Class<? extends SNode>> types) {
if (input == null) {
return new HashSet<>();
}
String showNamespaceConfig =
input.getMappings().getProperty(GridComponent.MAPPING_SHOW_NAMESPACE);
if (showNamespaceConfig != null) {
SDocumentGraph graph = input.getDocument().getDocumentGraph();
Set<String> annoPool = new LinkedHashSet<>();
for (Class<? extends SNode> t : types) {
annoPool.addAll(
SToken.class.isAssignableFrom(t)
? getAnnotationLevelSet(graph, null, t)
: getAnnotationLevelSet(graph, input.getNamespace(), t));
}
if ("true".equalsIgnoreCase(showNamespaceConfig)) {
// all annotations should be displayed with a namespace
return annoPool;
} else if ("false".equalsIgnoreCase(showNamespaceConfig)) {
return new LinkedHashSet<>();
} else {
Set<String> annos = new LinkedHashSet<>();
List<String> defs =
Splitter.on(',').omitEmptyStrings().trimResults().splitToList(showNamespaceConfig);
for (String s : defs) {
// is regular expression?
if (s.startsWith("/") && s.endsWith("/")) {
// go over all remaining items in our pool of all annotations and
// check if they match
Pattern regex = Pattern.compile(StringUtils.strip(s, "/"));
LinkedList<String> matchingAnnos = new LinkedList<>();
for (String a : annoPool) {
if (regex.matcher(a).matches()) {
matchingAnnos.add(a);
}
}
annos.addAll(matchingAnnos);
annoPool.removeAll(matchingAnnos);
} else {
annos.add(s);
annoPool.remove(s);
}
}
return annos;
}
}
return new LinkedHashSet<>();
}
/**
* Converts Salt document graph to rows.
*
* @param input
* @param showSpanAnnos
* @param showTokenAnnos
* @param mediaLayer A set of all annotation layers which should be treated as special media
* layer.
* @param annotationNames
* @param replaceValueWithMediaIcon If true the actual value is removed and an icon for playing
* the media file is shown instead.
* @param startTokenIndex token index of the first token in the match
* @param endTokenIndex token index of the last token in the match
* @param pdfController makes status of all pdfviewer available for the events.
* @param text If non-null only include annotations for nodes of the specified text.
* @return
*/
public static LinkedHashMap<String, ArrayList<Row>> parseSalt(
VisualizerInput input,
boolean showSpanAnnos,
boolean showTokenAnnos,
List<String> annotationNames,
Set<String> mediaLayer,
boolean replaceValueWithMediaIcon,
long startTokenIndex,
long endTokenIndex,
PDFController pdfController,
STextualDS text) {
SDocumentGraph graph = input.getDocument().getDocumentGraph();
// only look at annotations which were defined by the user
LinkedHashMap<String, ArrayList<Row>> rowsByAnnotation = new LinkedHashMap<>();
for (String anno : annotationNames) {
rowsByAnnotation.put(anno, new ArrayList<Row>());
}
AtomicInteger eventCounter = new AtomicInteger();
PDFPageHelper pageNumberHelper = new PDFPageHelper(input);
if (showSpanAnnos) {
for (SSpan span : graph.getSpans()) {
if (text == null || text == CommonHelper.getTextualDSForNode(span, graph)) {
addAnnotationsForNode(
span,
graph,
startTokenIndex,
endTokenIndex,
pdfController,
pageNumberHelper,
eventCounter,
rowsByAnnotation,
true,
mediaLayer,
replaceValueWithMediaIcon);
}
} // end for each span
}
if (showTokenAnnos) {
for (SToken tok : graph.getTokens()) {
if (text == null || text == CommonHelper.getTextualDSForNode(tok, graph)) {
addAnnotationsForNode(
tok,
graph,
startTokenIndex,
endTokenIndex,
pdfController,
pageNumberHelper,
eventCounter,
rowsByAnnotation,
false,
mediaLayer,
replaceValueWithMediaIcon);
}
}
}
// 2. merge rows when possible
for (Map.Entry<String, ArrayList<Row>> e : rowsByAnnotation.entrySet()) {
mergeAllRowsIfPossible(e.getValue());
}
// 3. sort events on one row by left token index
for (Map.Entry<String, ArrayList<Row>> e : rowsByAnnotation.entrySet()) {
for (Row r : e.getValue()) {
sortEventsByTokenIndex(r);
}
}
// 4. split up events if they cover islands
for (Map.Entry<String, ArrayList<Row>> e : rowsByAnnotation.entrySet()) {
for (Row r : e.getValue()) {
splitRowsOnIslands(r, graph, text, startTokenIndex, endTokenIndex);
}
}
// 5. split up events if they have gaps
for (Map.Entry<String, ArrayList<Row>> e : rowsByAnnotation.entrySet()) {
for (Row r : e.getValue()) {
splitRowsOnGaps(r, graph, startTokenIndex, endTokenIndex);
}
}
return rowsByAnnotation;
}
/**
* Returns the annotations to display according to the mappings configuration.
*
* <p>This will check the "annos" and "annos_regex" paramters for determining. the annotations to
* display. It also iterates over all nodes of the graph matching the type.
*
* @param input The input for the visualizer.
* @param type Which type of nodes to include
* @return
*/
public static List<String> computeDisplayAnnotations(
VisualizerInput input, Class<? extends SNode> type) {
if (input == null) {
return new LinkedList<>();
}
SDocumentGraph graph = input.getDocument().getDocumentGraph();
Set<String> annoPool =
SToken.class.isAssignableFrom(type)
? getAnnotationLevelSet(graph, null, type)
: getAnnotationLevelSet(graph, input.getNamespace(), type);
List<String> annos = new LinkedList<>(annoPool);
String annosConfiguration = input.getMappings().getProperty(MAPPING_ANNOS_KEY);
if (annosConfiguration != null && annosConfiguration.trim().length() > 0) {
String[] split = annosConfiguration.split(",");
annos.clear();
for (String s : split) {
s = s.trim();
// is regular expression?
if (s.startsWith("/") && s.endsWith("/")) {
// go over all remaining items in our pool of all annotations and
// check if they match
Pattern regex = Pattern.compile(StringUtils.strip(s, "/"));
LinkedList<String> matchingAnnos = new LinkedList<>();
for (String a : annoPool) {
if (regex.matcher(a).matches()) {
matchingAnnos.add(a);
}
}
annos.addAll(matchingAnnos);
annoPool.removeAll(matchingAnnos);
} else {
annos.add(s);
annoPool.remove(s);
}
}
}
// filter already found annotation names by regular expression
// if this was given as mapping
String regexFilterRaw = input.getMappings().getProperty(MAPPING_ANNO_REGEX_KEY);
if (regexFilterRaw != null) {
try {
Pattern regexFilter = Pattern.compile(regexFilterRaw);
ListIterator<String> itAnnos = annos.listIterator();
while (itAnnos.hasNext()) {
String a = itAnnos.next();
// remove entry if not matching
if (!regexFilter.matcher(a).matches()) {
itAnnos.remove();
}
}
} catch (PatternSyntaxException ex) {
log.warn("invalid regular expression in mapping for grid visualizer", ex);
}
}
return annos;
}