public override string Query(string word)
{
// Log($"-- Query : '{word}'");
var idx = findInLevel1(word);
if (idx < 0)
{
return(string.Empty);
}
const int bufSize = 32768;
using var fs = new FileStream(this.DictFileName, FileMode.Open, FileAccess.Read, FileShare.Read, bufSize);
var w = word;
if (!DictHeader.KeyCaseSensitiveBool)
{
w = word.ToUpperInvariant();
}
var kw1 = _seg1.IndexList[idx];
var l2 = readLevel2Block(idx, kw1, fs);
KwIndex2 result = null; // = l2.Find(kw2 => kw2.Keyword == word);
var size = -1;
for (var x = 0; x < l2.Count; x++)
{
if (l2[x].Keyword.ToUpperInvariant() != w)
{
continue;
}
result = l2[x];
if (x < l2.Count - 1)
{
var nextOne = l2[x + 1];
size = (int)(nextOne.RelOffsetUL - result.RelOffsetUL);
}
break;
}
if (result == null)
{
return(string.Empty);
}
Console.WriteLine($"<!-- kw1 block #{idx}: {kw1} -->");
Console.WriteLine($"<!-- kw2 found: {result} @ kw1 block #{idx} -->");
Console.WriteLine($"<!-- unzipped offset: 0x{result.RelOffsetUL:X8}|{result.RelOffsetUL} -->");
foreach (var ci in DictLargeContentIndexTable.Indexes)
{
if (result.RelOffsetUL > ci.OffsetUncomp + ci.UncompressedSize)
{
continue;
}
Console.WriteLine(
$"<!-- ctt: block found ({ci}) | Seg2ContentBlockOffset = 0x{_seg2.Seg2ContentBlockOffset:X8}, ContentIndexTableEndPos = 0x{ContentIndexTableEndPos:X8} -->");
var block = seekAndLoadCttBlock(fs, ci);
var ofs = (int)(result.RelOffsetUL - ci.OffsetUncomp);
var data = new byte[size > 0 ? size : block.Length - ofs];
Array.Copy(block, ofs, data, 0, data.Length);
Console.WriteLine(
$"<!-- ctt entry '{word}': ofs = 0x{ofs:X8}/{ofs}, len = 0x{data.Length:X8}/{data.Length} -->");
var html = DictHeader.LanguageMode.GetString(data, 0, data.Length);
html = Normalize(html);
return(html);
}
return(string.Empty);
}
public virtual string LoadContentByKeyword(KwIndex2 kwi2)
{
throw new NotImplementedException();
}
private KwIndex2List readLevel2Block(int idx, KwIndex1 kw1, Stream fs)
{
var pos = Seg2BeginPos;
// ulong unzipped = 0;
for (var kwiIdx = 0; kwiIdx < idx; kwiIdx++)
{
var kwi = _seg1.IndexList[kwiIdx];
pos += kwi.CompressedSize;
// unzipped += kwi.UncompressedSize;
// Console.WriteLine(
// $"seg 1, block {kwiIdx}, end pos = {pos}/0x{pos:X8} | unzipped offset: {unzipped} | kwi = {kwi}");
}
// Console.WriteLine($"seg 1, block {idx}, start pos = {pos}/0x{pos:X8} | kwi = {kw1}");
var list2 = new KwIndex2List();
fs.Seek((long)pos, SeekOrigin.Begin);
var magicNum = readUInt32(fs);
var j2 = readUInt32(fs);
var rawData = new byte[kw1.CompressedSize - 8];
fs.Read(rawData, 0, rawData.Length);
if (magicNum == 0x02000000)
{
#region InflaterDecompress
try
{
// var txt = CompressUtil.InflaterDecompress(rawData, 0, rawData.Length, false);
// TODO need review and debug
var txt = Zipper.InflateBufferWithPureZlib(rawData, rawData.Length);
#region log & trying to parse 2nd level indexes
var ofs = 0;
while (ofs < txt.Length)
{
var kwi2 = new KwIndex2 {
RelOffsetUL = readUInt64(txt, ofs)
};
ofs += 8;
var ofs0 = ofs;
if (Equals(DictHeader.LanguageMode, Encoding.Unicode))
{
uint x9 = 1;
while (x9 != 0)
{
x9 = readUInt16(txt, ofs);
ofs += 2;
}
}
else
{
while (txt[ofs] != 0)
{
ofs++;
}
ofs++;
}
kwi2.Keyword = DictHeader.LanguageMode.GetString(txt, ofs0, ofs - ofs0).TrimEnd('\0');
list2.Add(kwi2);
}
#endregion
}
catch (Exception ex)
{
ErrorLog(ex.ToString());
throw;
}
#endregion
}
else if (magicNum == 0x01000000)
{
#region LZO.Decompress (V2.0 & V1.2)
#region v1.2: 提取二级索引表
var ofs = 0;
var cZipped = rawData[ofs]; //预览一个word
var zipped = cZipped != (byte)0; //如果头一个byte为0，则实际上该块未压缩。
byte[] decompressedData;
if (!zipped)
{
decompressedData = rawData;
}
else
{
#region TODO: unzip
decompressedData = new byte[kw1.UncompressedSize];
var ok = false;
// var in_len = rawData.Length - 3;
// int out_len = BitConverter.ToInt16(rawData, 1);
// if (!ok)
{
try
{
Zipper.MiniLzoDecompress(rawData, 0, (int)kw1.CompressedSize, decompressedData);
ok = decompressedData.Length == (int)kw1.UncompressedSize;
// Debug.Assert(ok, "M-dict 1.2 lzo decompressed failed.");
}
catch (Exception ex)
{
Debug.WriteLine(ex.ToString());
}
}
if (!ok)
{
ErrorLog("BAD, 1.1");
}
#endregion
}
#region 解释二级索引表
while (ofs < (int)kw1.UncompressedSize)
{
var kwi2 = new KwIndex2 {
RelOffsetUL = readUInt32(decompressedData, ofs)
};
ofs += 4; // if (x1 < 0) { Debug.WriteLine(">>> ???"); }
var ofs0 = ofs;
if (Equals(DictHeader.LanguageMode, Encoding.Unicode))
{
uint x9 = 1;
while (x9 != 0)
{
x9 = readUInt16(decompressedData, ofs);
ofs += 2;
}
}
else
{
while (decompressedData[ofs] != 0)
{
ofs++;
}
ofs++;
}
kwi2.Keyword = DictHeader.LanguageMode.GetString(decompressedData, ofs0, ofs - ofs0)
.TrimEnd('\0');
list2.Add(kwi2);
}
#endregion
#endregion
#endregion
}
else if (magicNum == 0)
{
#region log & trying to parse 2nd level indexes
var ofs = 0;
while (ofs < rawData.Length)
{
var kwi2 = new KwIndex2 {
RelOffsetUL = readUInt32(rawData, ofs)
};
ofs += 4;
var ofs0 = ofs;
if (Equals(DictHeader.LanguageMode, Encoding.Unicode))
{
uint x9 = 1;
while (x9 != 0)
{
x9 = readUInt16(rawData, ofs);
ofs += 2;
}
}
else
{
while (rawData[ofs] != 0)
{
ofs++;
}
ofs++;
}
kwi2.Keyword = DictHeader.LanguageMode.GetString(rawData, ofs0, ofs - ofs0).TrimEnd('\0');
list2.Add(kwi2);
Log($" > {kwi2}");
}
#endregion
}
else
{
throw new Exception(
$"提取KWIndex2[]时，期望正确的算法标志0x2000000/0x1000000，然而遇到了{magicNum}/0x{magicNum:X}");
}
return(list2);
}
public virtual byte[] LoadContentBytesByKeyword(KwIndex2 kwi2)
{
throw new NotImplementedException();
}