public BlockPage Process(BlockPage page)
{
IBlock last = null;
BlockSet <IBlock> currentBlockSet = null;
var result = new BlockPage();
foreach (var block in page.AllBlocks)
{
bool shouldBreak = false;
if (last != null)
{
// expect: previous >~ next
float previous = last.GetH();
float next = block.GetH();
// previous >> next
if (previous > next + statDownInTheBottom)
{
shouldBreak = true;
}
// previous < next
if (previous < next - statGoingUp)
{
shouldBreak = true;
}
}
// check for superscript font
if ((shouldBreak) && (Block.IsSuperscriptFont((Block)last, (Block)block)))
{
shouldBreak = false;
}
if (shouldBreak && currentBlockSet.Count() > 1)
{
var tableline = currentBlockSet.TakeLast(2).First();
if (Block.AreSameLine(tableline, block))
{
shouldBreak = false;
}
}
if ((currentBlockSet == null) || shouldBreak)
{
currentBlockSet = new BlockSet <IBlock>();
result.Add(currentBlockSet);
}
currentBlockSet.Add(block);
last = block;
}
return(result);
}
IBlock[] CreateNewBlocks(BlockSet <IBlock> blocks, int middle)
{
int total = blocks.Count();
var blockA = new BlockSet <IBlock>();
var blockB = new BlockSet <IBlock>();
blockA.AddRange(blocks.Take(middle));
blockB.AddRange(blocks.TakeLast(total - middle));
return(new IBlock[] { blockA, blockB });
}
BlockSet <IBlock> GetCleanBlockSet(Block overlap, BlockSet <IBlock> block1)
{
var blockSet = new BlockSet <IBlock>();
var b1 = block1.Where(b => !HasOverlapY(b, overlap));
blockSet.AddRange(b1);
if (blockSet.Count() == 0)
{
return(null);
}
return(blockSet);
}
bool HasLineOverlap(BlockSet <IBlock> a, BlockSet <IBlock> b)
{
// why it would happen?
if (a.Count() < 2)
{
PdfReaderException.AlwaysThrow("a.Count() < 2");
}
var lastLines = a.TakeLast(2);
var firstLine = b.Take(1).First();
var last = lastLines.First();
if (Block.AreSameLine(last, firstLine))
{
return(true);
}
return(false);
}
BlockSet <IBlock>[] SplitBlock(BlockSet <IBlock> blockset)
{
var blocks = blockset.ToList();
int total = blocks.Count - 1;
float limit = blockset.GetWidth() / 2;
int start = ScanBlock(i => blocks[i], blockset.GetX() + limit);
int end = ScanBlock(i => blocks[total - i], blockset.GetX() + limit);
// no split
if (start == 0 && end == 0)
{
// VALIDATE
//System.Diagnostics.Debugger.Break();
return(new BlockSet <IBlock>[] { blockset });
}
// split into 2 pieces
int middle = -1;
// split into 2 pieces: there is a clear division
middle = (start + end > total) ? (start) : middle;
middle = (start == 0) ? (total - end + 1) : middle;
middle = (end == 0) ? (start) : middle;
if (middle > 0)
{
var blockA = new BlockSet <IBlock>();
var blockB = new BlockSet <IBlock>();
blockA.AddRange(blocks.Take(middle));
blockB.AddRange(blocks.TakeLast(total - middle + 1));
int count2 = blockA.Count() + blockB.Count();
if (count2 != blocks.Count)
{
PdfReaderException.AlwaysThrow("count2 != blocks.Count");
}
// VALIDATE
//System.Diagnostics.Debugger.Break();
return(new BlockSet <IBlock>[] { blockA, blockB });
}
// split into 3 pieces
var topBlock = new BlockSet <IBlock>();
var coreBlock = new BlockSet <IBlock>();
var bottomBlock = new BlockSet <IBlock>();
topBlock.AddRange(blocks.Take(start));
for (int i = start; i <= total - end; i++)
{
coreBlock.Add(blocks[i]);
}
bottomBlock.AddRange(blocks.TakeLast(end));
int count3 = topBlock.Count() + coreBlock.Count() + bottomBlock.Count();
if (count3 != blocks.Count)
{
PdfReaderException.AlwaysThrow("count3 != blocks.Count");
}
// VALIDATE
//System.Diagnostics.Debugger.Break();
return(new BlockSet <IBlock>[] { topBlock, coreBlock, bottomBlock });
}