public static Header consumeHeaders(List<Line> remainingMessage) {
final int headerLineNumber =
remainingMessage.size() != 0 ? remainingMessage.get(0).getLineNumber() : 0;
final Header headers = new Header(headerLineNumber);
final Iterator<Line> iter = remainingMessage.iterator();
Line currentLine;
boolean isHeader = true;
while (iter.hasNext() && isHeader) {
currentLine = iter.next();
final Matcher headerMatcher = PATTERN_HEADER_LINE.matcher(currentLine.toString());
if (headerMatcher.matches() && headerMatcher.groupCount() == 2) {
iter.remove();
String headerName = headerMatcher.group(1).trim();
String headerValue = headerMatcher.group(2).trim();
headers.addHeader(
headerName, Header.splitValuesByComma(headerValue), currentLine.getLineNumber());
} else {
isHeader = false;
}
}
return headers;
}
public static Line removeEndingCRLF(final Line line) {
Matcher matcher = PATTERN_LAST_CRLF.matcher(line.toString());
if (matcher.matches()) {
return new Line(matcher.group(1), line.getLineNumber());
} else {
return line;
}
}
/**
* This method is used for extraction of tables with lot of empty cells in it. It is required for
* the successful extraction of most Matrix tables.
*/
private void fillBlankCells() {
// We say: cells get a line number. If a column does not contain a cell on a certain line, add a
// whitespace.
// Any cell that is not filled must be empty:
for (Line line : data) {
int lineNumber = line.getLineNumber();
COLUMNLOOP:
for (Column2 column : dataInColumns) {
for (Cell cell : column.getCellObjects()) {
if (cell.getLineNumber() == lineNumber) {
break;
}
if (cell.getLineNumber() > line.getLineNumber()) { // the last cell?
// Add a blank cell to this column.
// System.out.println("Add line to :" + column + " in line: " +
// line.getLineNumber());
// <span class='ocrx_word' id='word_9' title="bbox 2175 514 2346 555">were</span>
Tag t = Tag.valueOf("span");
Attributes attributes = new Attributes();
attributes.put("class", "ocrx_word");
attributes.put("id", "word_ADDEDBYTEA");
attributes.put(
"title",
"bbox "
+ column.getAverageX1()
+ " "
+ (int) line.getAverageY1()
+ " "
+ column.getAverageX2()
+ " "
+ (int) line.getAverageY2());
Element newElement = new Element(t, "localhost:8080", attributes);
newElement.text(" ");
ArrayList<Element> newCell = new ArrayList<Element>();
newCell.add(newElement);
// System.out.println("adding: " +newElement.text());
column.addCell(newCell);
break COLUMNLOOP;
}
}
}
}
}
/**
* This method adds lines with missing partitions to the current columns. It loops trough lines
* that were flagged for containing missing data and then adds the ones to columns that they fit
* in (or to columns that fit in the partition). If this fails it will also try to check if the
* partition merely touches a column, although this will return a lower validation score if
* successful.
*/
private void addLinesWithMissingDataToColumns() {
ArrayList<Cell> cellsWithMissingDataAdded = new ArrayList<Cell>();
ArrayList<Column2> newDataInColumns = new ArrayList<Column2>();
for (Line line : linesWithMissingData) {
ArrayList<ArrayList<Element>> clusters = line.getClusters();
for (ArrayList<Element> cluster : clusters) {
for (Column2 column : dataInColumns) {
if (column.fitsInColumn(Line.getClusterX1(cluster), Line.getClusterX2(cluster))
&& column.columnFitsIn(Line.getClusterX1(cluster), Line.getClusterX2(cluster))) {
// then we need to add this cluster to that column:
newDataInColumns.remove(column);
column.addCell(cluster);
newDataInColumns.add(column);
Cell cell = new Cell(cluster, 3, line.getLineNumber());
cellsWithMissingDataAdded.add(cell);
} else if (column.fitsInColumn(Line.getClusterX1(cluster), Line.getClusterX2(cluster))
|| column.columnFitsIn(Line.getClusterX1(cluster), Line.getClusterX2(cluster))) {
// then we need to add this cluster to that column:
newDataInColumns.remove(column);
column.addCell(cluster);
newDataInColumns.add(column);
Cell cell = new Cell(cluster, 2, line.getLineNumber());
cellsWithMissingDataAdded.add(cell);
} else if (column.touchesColumn(Line.getClusterX1(cluster), Line.getClusterX2(cluster))) {
// System.out.println("CLUSTER: " + cluster + " touches: " +
// column);
newDataInColumns.remove(column);
column.addCell(cluster);
newDataInColumns.add(column);
Cell cell = new Cell(cluster, 1, line.getLineNumber());
cellsWithMissingDataAdded.add(cell);
}
}
}
}
validation.setCellsWithMissingDataAdded(cellsWithMissingDataAdded.size());
validation.setCellsWithMissingDataAddedScores(cellsWithMissingDataAdded);
}
/**
* This method creates a list of cell objects derived from the partitions in this line.
*
* @return A list containing the cell objects in this line.
*/
public ArrayList<Cell> getCellObjects() {
ArrayList<Cell> cellObjects = new ArrayList<Cell>();
for (ArrayList<Element> cell : cells) {
// Now we need to find out in which line this cell is:
String pos;
String[] positions;
pos = cell.get(0).attr("title");
positions = pos.split("\\s+");
int cellY1 = Integer.parseInt(positions[4]);
int lineNumber = 0;
for (Line line : data) {
if (line.getAverageY1() > cellY1) {
break;
}
lineNumber = line.getLineNumber();
}
Cell currentCell = new Cell(cell, lineNumber);
cellObjects.add(currentCell);
}
return cellObjects;
}