public static InflModel getInstance()
{
if (ins == null)
{
ins = new InflModel();
}
return(ins);
}
public static List <Dictionary <int, double> > StrFreq2IdFreq(List <Dictionary <string, double> > strFeqList, FeatureType featureType, int pos, ref int count)
{
Dictionary <string, int> dic = new Dictionary <string, int>();
switch (featureType)
{
case FeatureType.root:
{
RootModel ins = RootModel.getInstance();
dic = ins.dic;
count = dic.Count;
break;
}
case FeatureType.infl:
{
InflModel ins = InflModel.getInstance();
dic = ins.dic;
count = dic.Count;
break;
}
}
List <Dictionary <int, double> > ans = new List <Dictionary <int, double> >(strFeqList.Count);
int id;
foreach (var inpDict in strFeqList)
{
Dictionary <int, double> dict = new Dictionary <int, double>();
foreach (var kvp in inpDict)
{
if (dic.ContainsKey(kvp.Key) && (id = dic[kvp.Key]) >= 0)
{
dict[id + pos] = kvp.Value;
}
}
ans.Add(dict);
}
return(ans);
}
/// <summary>
/// convert input string to letter-n-gram sequence, each word is a letter-n-gram vector
/// </summary>
/// <param name="s">input string</param>
/// <param name="v">vocab</param>
/// <param name="N">ngram</param>
/// <param name="nMaxLength">max length</param>
/// <returns></returns>
/// <summary>
/// convert input string to letter-n-gram sequence, each word is a letter-n-gram vector
/// </summary>
/// <param name="s">input string</param>
/// <param name="v">vocab</param>
/// <param name="N">ngram</param>
/// <param name="nMaxLength">max length</param>
/// <returns></returns>
public static List <Dictionary <string, double> > String2FeatStrSeq(string s, int N, int nMaxLength, FeatureType feaType)
{
List <Dictionary <string, double> > rgWfs = new List <Dictionary <string, double> >();
string[] rgw = TokenizeToArray(s);
for (int i = 0; i < Math.Min(rgw.Length, nMaxLength - 1); ++i)
{
switch (feaType)
{
case FeatureType.l3g:
{
rgWfs.Add(String2L3g(rgw[i], N));
break;
}
case FeatureType.root:
{
RootModel rootModelIns = RootModel.getInstance();
rgWfs.Add(String2Root(rgw[i], rootModelIns.dicWord2Roots));
break;
}
case FeatureType.infl:
{
InflModel inflModelIns = InflModel.getInstance();
rgWfs.Add(String2Root(rgw[i], inflModelIns.dicInfl2Ori));
break;
}
}
}
Dictionary <string, double> dict = new Dictionary <string, double>();
for (int i = nMaxLength - 1; i < rgw.Length; ++i)
{
Dictionary <string, double> tmp_dict = null;
switch (feaType)
{
case FeatureType.l3g:
{
tmp_dict = String2L3g(rgw[i], N);
break;
}
case FeatureType.root:
{
RootModel rootModelIns = RootModel.getInstance();
tmp_dict = String2Root(rgw[i], rootModelIns.dicWord2Roots);
break;
}
case FeatureType.infl:
{
InflModel inflModelIns = InflModel.getInstance();
tmp_dict = String2Root(rgw[i], inflModelIns.dicInfl2Ori);
break;
}
}
foreach (KeyValuePair <string, double> kv in tmp_dict)
{
if (dict.ContainsKey(kv.Key))
{
dict[kv.Key] += kv.Value;
}
else
{
dict.Add(kv.Key, kv.Value);
}
}
}
if (dict.Count > 0)
{
rgWfs.Add(dict);
}
return(rgWfs);
}