/// <summary>
/// <seealso cref="FieldFragList.Add(int, int, IList{WeightedPhraseInfo})"/>
/// </summary>
public override void Add(int startOffset, int endOffset, IList <WeightedPhraseInfo> phraseInfoList)
{
float totalBoost = 0;
List <SubInfo> subInfos = new List <SubInfo>();
foreach (WeightedPhraseInfo phraseInfo in phraseInfoList)
{
subInfos.Add(new SubInfo(phraseInfo.GetText(), phraseInfo.TermsOffsets, phraseInfo.Seqnum, phraseInfo.Boost));
totalBoost += phraseInfo.Boost;
}
FragInfos.Add(new WeightedFragInfo(startOffset, endOffset, subInfos, totalBoost));
}
/// <summary>
/// <seealso cref="FieldFragList.Add(int, int, IList{WeightedPhraseInfo})"/>.
/// </summary>
public override void Add(int startOffset, int endOffset, IList <WeightedPhraseInfo> phraseInfoList)
{
List <SubInfo> tempSubInfos = new List <SubInfo>();
List <SubInfo> realSubInfos = new List <SubInfo>();
HashSet <string> distinctTerms = new HashSet <string>();
int length = 0;
foreach (WeightedPhraseInfo phraseInfo in phraseInfoList)
{
float phraseTotalBoost = 0;
foreach (TermInfo ti in phraseInfo.TermsInfos)
{
if (distinctTerms.Add(ti.Text))
{
phraseTotalBoost += ti.Weight * phraseInfo.Boost;
}
length++;
}
tempSubInfos.Add(new SubInfo(phraseInfo.GetText(), phraseInfo.TermsOffsets,
phraseInfo.Seqnum, phraseTotalBoost));
}
// We want that terms per fragment (length) is included into the weight. Otherwise a one-word-query
// would cause an equal weight for all fragments regardless of how much words they contain.
// To avoid that fragments containing a high number of words possibly "outrank" more relevant fragments
// we "bend" the length with a standard-normalization a little bit.
float norm = length * (1 / (float)Math.Sqrt(length));
float totalBoost = 0;
foreach (SubInfo tempSubInfo in tempSubInfos)
{
float subInfoBoost = tempSubInfo.Boost * norm;
realSubInfos.Add(new SubInfo(tempSubInfo.Text, tempSubInfo.TermsOffsets,
tempSubInfo.Seqnum, subInfoBoost));
totalBoost += subInfoBoost;
}
FragInfos.Add(new WeightedFragInfo(startOffset, endOffset, realSubInfos, totalBoost));
}
