/// <summary> Initializes the data structure for segment positions with given string.</summary>
/// <param name="str">- the plain string to analyze
/// </param>
/// <param name="simti">- SIMple Trie Index
/// </param>
public virtual void init(System.String str, Simti simti)
{
int prevIndex = 0;
int nextIndex = 0;
positionEnd = 0;
prevIndex = addPosition(POSITION_START_KEY);
position[prevIndex].state = SP_STATE_M;
System.String rev = "";
for (int i = str.Length - 1; i >= 0; i--)
{
rev += str[i];
}
for (int i = 0; i < str.Length; i++)
{
char c = str[i];
nextIndex = addPosition(c);
setPositionLink(prevIndex, nextIndex);
prevIndex = nextIndex;
simti.insert(rev.Substring(0, (str.Length - i) - (0)).ToCharArray(), nextIndex);
}
/* for marking the end of the eojeol */
setPositionLink(prevIndex, 0);
}
/// <summary> Initializes the Chart-based Morphological Analyzer plug-in.</summary>
/// <param name="baseDir">- the path for base directory, which should have the 'conf' and 'data' directory
/// </param>
/// <param name="configFile">- the path for the configuration file (relative path to the base directory)
/// </param>
public virtual void initialize(System.String baseDir, System.String configFile)
{
JSONReader json = new JSONReader(configFile);
fileDicSystem = baseDir + "/" + json.getValue("dic_system");
fileDicUser = baseDir + "/" + json.getValue("dic_user");
fileConnections = baseDir + "/" + json.getValue("connections");
fileConnectionsNot = baseDir + "/" + json.getValue("connections_not");
fileDicAnalyzed = baseDir + "/" + json.getValue("dic_analyzed");
fileTagSet = baseDir + "/" + json.getValue("tagset");
tagSet = new TagSet();
tagSet.init(fileTagSet, TagSet.TAG_SET_KAIST);
connection = new Connection();
connection.init(fileConnections, tagSet.TagCount, tagSet);
connectionNot = new ConnectionNot();
connectionNot.init(fileConnectionsNot, tagSet);
analyzedDic = new AnalyzedDic();
analyzedDic.readDic(fileDicAnalyzed);
systemDic = new Trie(Trie.DEFAULT_TRIE_BUF_SIZE_SYS);
systemDic.read_dic(fileDicSystem, tagSet);
userDic = new Trie(Trie.DEFAULT_TRIE_BUF_SIZE_USER);
userDic.read_dic(fileDicUser, tagSet);
numDic = new NumberDic();
simti = new Simti();
simti.init();
eojeolList = new LinkedList <Eojeol>();
chart = new MorphemeChart(tagSet, connection, systemDic, userDic, numDic, simti, eojeolList);
postProc = new PostProcessor();
}
/// <summary> Constructor.</summary>
/// <param name="tagSet">- the morpheme tag set
/// </param>
/// <param name="connection">- the morpheme connection rules
/// </param>
/// <param name="systemDic">- the system morpheme dictionary
/// </param>
/// <param name="userDic">- the user morpheme dictionary
/// </param>
/// <param name="numDic">- the number dictionary
/// </param>
/// <param name="simti">- the SIMple Trie Index
/// </param>
/// <param name="resEojeolList">- the list of eojeols to store the analysis result
/// </param>
public MorphemeChart(TagSet tagSet, Connection connection, Trie systemDic, Trie userDic, NumberDic numDic, Simti simti, LinkedList <Eojeol> resEojeolList)
{
chart = new Morpheme[MAX_MORPHEME_CHART];
for (int i = 0; i < MAX_MORPHEME_CHART; i++)
{
chart[i] = new Morpheme(this);
}
this.sp = new SegmentPosition();
this.tagSet = tagSet;
this.connection = connection;
this.exp = new Exp(this, tagSet);
this.systemDic = systemDic;
this.userDic = userDic;
this.numDic = numDic;
this.simti = simti;
this.resEojeols = resEojeolList;
resMorphemes = new List <String>();
resTags = new List <String>();
chiReplacementList = new LinkedList <String>();
engReplacementList = new LinkedList <String>();
}
private void InitBlock(Simti enclosingInstance)
{
this.enclosingInstance = enclosingInstance;
}
public ST_NODE(Simti enclosingInstance)
{
InitBlock(enclosingInstance);
}
private void InitBlock(Simti enclosingInstance)
{
this.enclosingInstance = enclosingInstance;
node = new ST_NODE(enclosingInstance);
free = new ST_FREE(enclosingInstance);
}
public HEADI(Simti enclosingInstance)
{
InitBlock(enclosingInstance);
}
/// <summary> Initializes the data structure for segment positions with given string.</summary>
/// <param name="str">- the plain string to analyze
/// </param>
/// <param name="simti">- SIMple Trie Index
/// </param>
public virtual void init(System.String str, Simti simti)
{
int prevIndex = 0;
int nextIndex = 0;
positionEnd = 0;
prevIndex = addPosition(POSITION_START_KEY);
position[prevIndex].state = SP_STATE_M;
System.String rev = "";
for (int i = str.Length - 1; i >= 0; i--)
{
rev += str[i];
}
for (int i = 0; i < str.Length; i++)
{
char c = str[i];
nextIndex = addPosition(c);
setPositionLink(prevIndex, nextIndex);
prevIndex = nextIndex;
simti.insert(rev.Substring(0, (str.Length - i) - (0)).ToCharArray(), nextIndex);
}
/* for marking the end of the eojeol */
setPositionLink(prevIndex, 0);
}
/// <summary> Initializes the Chart-based Morphological Analyzer plug-in.</summary>
/// <param name="baseDir">- the path for base directory, which should have the 'conf' and 'data' directory
/// </param>
/// <param name="configFile">- the path for the configuration file (relative path to the base directory)
/// </param>
public virtual void initialize(System.String baseDir, System.String configFile)
{
JSONReader json = new JSONReader(configFile);
fileDicSystem = baseDir + "/" + json.getValue("dic_system");
fileDicUser = baseDir + "/" + json.getValue("dic_user");
fileConnections = baseDir + "/" + json.getValue("connections");
fileConnectionsNot = baseDir + "/" + json.getValue("connections_not");
fileDicAnalyzed = baseDir + "/" + json.getValue("dic_analyzed");
fileTagSet = baseDir + "/" + json.getValue("tagset");
tagSet = new TagSet();
tagSet.init(fileTagSet, TagSet.TAG_SET_KAIST);
connection = new Connection();
connection.init(fileConnections, tagSet.TagCount, tagSet);
connectionNot = new ConnectionNot();
connectionNot.init(fileConnectionsNot, tagSet);
analyzedDic = new AnalyzedDic();
analyzedDic.readDic(fileDicAnalyzed);
systemDic = new Trie(Trie.DEFAULT_TRIE_BUF_SIZE_SYS);
systemDic.read_dic(fileDicSystem, tagSet);
userDic = new Trie(Trie.DEFAULT_TRIE_BUF_SIZE_USER);
userDic.read_dic(fileDicUser, tagSet);
numDic = new NumberDic();
simti = new Simti();
simti.init();
eojeolList = new LinkedList < Eojeol >();
chart = new MorphemeChart(tagSet, connection, systemDic, userDic, numDic, simti, eojeolList);
postProc = new PostProcessor();
}
private void InitBlock(Simti enclosingInstance)
{
this.enclosingInstance = enclosingInstance;
}
public ST_FREE(Simti enclosingInstance)
{
InitBlock(enclosingInstance);
}
private void InitBlock(Simti enclosingInstance)
{
this.enclosingInstance = enclosingInstance;
s_free = new ST_FREE(enclosingInstance);
s_node = new ST_NODE(enclosingInstance);
}
public HEADI(Simti enclosingInstance)
{
InitBlock(enclosingInstance);
}
/// <summary> Constructor.</summary>
/// <param name="tagSet">- the morpheme tag set
/// </param>
/// <param name="connection">- the morpheme connection rules
/// </param>
/// <param name="systemDic">- the system morpheme dictionary
/// </param>
/// <param name="userDic">- the user morpheme dictionary
/// </param>
/// <param name="numDic">- the number dictionary
/// </param>
/// <param name="simti">- the SIMple Trie Index
/// </param>
/// <param name="resEojeolList">- the list of eojeols to store the analysis result
/// </param>
public MorphemeChart(TagSet tagSet, Connection connection, Trie systemDic, Trie userDic, NumberDic numDic, Simti simti, LinkedList<Eojeol> resEojeolList)
{
chart = new Morpheme[MAX_MORPHEME_CHART];
for (int i = 0; i < MAX_MORPHEME_CHART; i++)
{
chart[i] = new Morpheme(this);
}
this.sp = new SegmentPosition();
this.tagSet = tagSet;
this.connection = connection;
this.exp = new Exp(this, tagSet);
this.systemDic = systemDic;
this.userDic = userDic;
this.numDic = numDic;
this.simti = simti;
this.resEojeols = resEojeolList;
resMorphemes = new List<String>();
resTags = new List<String>();
chiReplacementList = new LinkedList<String>();
engReplacementList = new LinkedList<String>();
}
