/// <summary>
/// Loads all proximity classes from the hard disk The WORDS map must be
/// created before!
/// </summary>
/// <exception cref="System.IO.IOException"/>
public static void LoadProximityClasses(string proxFileName)
{
log.Info("Loading proximity classes...");
BufferedReader @in = null;
try
{
@in = new BufferedReader(new FileReader(proxFileName));
}
catch (IOException)
{
log.Info("Warning: no proximity database found.");
return;
}
string line;
while ((line = @in.ReadLine()) != null)
{
List <string> tokens = SimpleTokenize.Tokenize(line);
if (tokens.Count > 0)
{
int key = Words.Get(tokens[0]);
List <int> value = new List <int>();
for (int i = 0; i < tokens.Count && i < ProximityClassSize; i++)
{
int word = Words.Get(tokens[i]);
value.Add(word);
}
ProxClasses[key] = value;
}
}
@in.Close();
log.Info("Finished loading proximity classes.");
}
/// <summary>Loads one dictionary from disk</summary>
/// <exception cref="Java.IO.FileNotFoundException"/>
/// <exception cref="System.IO.IOException"/>
private static void LoadDictionary(IDictionary <string, string> dict, string file)
{
BufferedReader @in = new BufferedReader(new FileReader(file));
string line;
while ((line = @in.ReadLine()) != null)
{
List <string> tokens = SimpleTokenize.Tokenize(line);
if (tokens.Count > 0)
{
string lower = tokens[0].ToLower();
if (tokens.Count == 1)
{
dict[lower] = "true";
}
else
{
dict[lower] = tokens[1];
}
}
}
}
/// <exception cref="System.IO.IOException"/>
private void ReadPredictedEntityBoundaries(BufferedReader @is)
{
// System.out.println("Reading boundaries from file: " + mPrefix);
//
// read Massi's B-ENT, I-ENT, or O labels
//
List <string> labels = new List <string>();
string line;
while ((line = @is.ReadLine()) != null)
{
List <string> tokens = SimpleTokenize.Tokenize(line);
if (tokens.IsEmpty() == false)
{
labels.Add(tokens[0]);
}
}
System.Diagnostics.Debug.Assert((labels.Count == mTokens.Count));
int entityId = 1;
//
// traverse the label array and create entities as needed
//
for (int i = 0; i < labels.Count; i++)
{
// System.out.println(labels.get(i));
if (labels[i].StartsWith("B-") || labels[i].StartsWith("I-"))
{
// Massi's
// ents
// may
// start
// with
// I-ENT
int startToken = i;
int endToken = i + 1;
while (endToken < labels.Count && labels[endToken].StartsWith("I-"))
{
endToken++;
}
//
// Set the type/subtype to whatever Massi predicted
// This is not directly used in this system. It is needed only
// to generate the APF files with Massi info, which are needed
// by Edgar. Otherwise type/subtype could be safely set to "none".
//
string label = labels[startToken];
int dash = label.IndexOf("-", 2);
if (dash <= 2 || dash >= label.Length)
{
throw new Exception(label);
}
System.Diagnostics.Debug.Assert((dash > 2 && dash < label.Length - 1));
string type = Sharpen.Runtime.Substring(label, 2, dash);
string subtype = Sharpen.Runtime.Substring(label, dash + 1);
/*
* String type = "none"; String subtype = "none";
*/
// create a new entity between [startToken, endToken)
MakeEntity(startToken, endToken, entityId, type, subtype);
// skip over this entity
i = endToken - 1;
entityId++;
}
else
{
System.Diagnostics.Debug.Assert((labels[i].Equals("O")));
}
}
}
