Class: Ferret::Search::Similarity

Inherits:
Object
  • Object
show all
Defined in:
lib/ferret/search/similarity.rb,
lib/ferret/search/similarity.rb

Overview

Expert: Scoring API. Subclasses implement search scoring.

The score of query q for document d is defined in terms of these methods as follows:

<table cellpadding=“0” cellspacing=“0” border=“0”>

<tr>
  <td valign="middle" align="right" rowspan="2">score(q,d) =<br></td>
  <td valign="middle" align="center">
  <big><big><big><big><big>&Sigma</big></big></big></big></big></td>
  <td valign="middle"><small>
  #tf(int) tf(t in d)#
  #idf_term(Term,Searcher) idf(t)#
  Field#getBoost getBoost(t.field in d)#
  #length_norm(String,int) length_norm(t.field in d)
  </small></td>
  <td valign="middle" rowspan="2">&nbsp*
  #coord(int,int) coord(q,d)#
  #query_norm(float) query_norm(q)
  </td>
</tr>
<tr>
 <td valign="top" align="right">
  <small>t in q</small>
  </td>
</tr>

</table>

See #set_default See IndexWriter#set_similarity See Searcher#set_similarity

Direct Known Subclasses

DefaultSimilarity

Constant Summary collapse

NORM_TABLE =

Cache of decoded bytes

Array.new(256) { |i| Similarity.byte_to_float(i) }
@@default =

The Similarity implementation used by default.

DefaultSimilarity.new()

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.byte_to_float(b) ⇒ Object 



36
37
38
39
40
41
42
43
 
# File 'lib/ferret/search/similarity.rb', line 36

def Similarity.byte_to_float(b)
  if (b == 0)
    return 0.0
  end
  mantissa = b & 0x07           # 0x07 =  7 = 0b00000111
  exponent = (b >> 3) & 0x1F    # 0x1f = 31 = 0b00011111
  return [0,0,(mantissa << 5),(exponent+48)].pack("cccc").unpack("f")[0]
end

.decode_norm(b) ⇒ Object

Decodes a normalization factor stored in an index. See Similarity#encode_norm(float)



70
71
72
 
# File 'lib/ferret/search/similarity.rb', line 70

def Similarity.decode_norm(b) 
  return NORM_TABLE[b & 0xFF]
end

.default ⇒ Object 



270
271
272
 
# File 'lib/ferret/search/similarity.rb', line 270

def Similarity.default
  return @@default
end

.default=(default) ⇒ Object 



274
275
276
 
# File 'lib/ferret/search/similarity.rb', line 274

def Similarity.default=(default)
  @@default = default
end

.encode_norm(f) ⇒ Object

Encodes a normalization factor for storage in an index.

The encoding uses a five-bit exponent and three-bit mantissa, thus representing values from around 7x10^9 to 2x10^-9 with about one significant decimal digit of accuracy. Zero is also represented. Negative numbers are rounded up to zero. Values too large to represent are rounded down to the largest representable value. Positive values too small to represent are rounded up to the smallest positive representable value.

See Field#boost=



127
128
129
 
# File 'lib/ferret/search/similarity.rb', line 127

def Similarity.encode_norm(f) 
  return Similarity.float_to_byte(f)
end

.float_to_byte(f) ⇒ Object 



45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
 
# File 'lib/ferret/search/similarity.rb', line 45

def Similarity.float_to_byte(f) 
  if (f <= 0.0) then return 0 end

  bits = [f].pack("f").unpack("cccc")
  mantissa = (bits[2] & 0xEf) >> 5 
  exponent = (bits[3] - 48)

  if (exponent > 0x1f)
    exponent = 0x1f   # 0x1f = 31 = 0b00011111
    mantissa = 0x07   # 0x07 =  7 = 0b00000111
  end

  if (exponent < 0)
    exponent = 0
    mantissa = 1
  end

  return ((exponent<<3) | mantissa)
end

Instance Method Details

#coord ⇒ Object

Computes a score factor based on the fraction of all query terms that a document contains. This value is multiplied into scores.

The presence of a large portion of the query terms indicates a better match with the query, so implemenations of this method usually return larger values when the ratio between these parameters is large and smaller values when the ratio between them is small.

overlap

the number of query terms matched in the document

max_overlap

the total number of terms in the query

Return

a score factor based on term overlap with the query

Raises:

  • (NotImplementedError) 


227
228
229
 
# File 'lib/ferret/search/similarity.rb', line 227

def coord
  raise NotImplementedError
end

#decode_norm(b) ⇒ Object

Decodes a normalization factor stored in an index. See Similarity#encode_norm(float)



76
77
78
 
# File 'lib/ferret/search/similarity.rb', line 76

def decode_norm(b)
  return self.class.decode_norm(b)
end

#encode_norm(f) ⇒ Object 



131
132
133
 
# File 'lib/ferret/search/similarity.rb', line 131

def encode_norm(f) 
  return self.class.float_to_byte(f)
end

#idf ⇒ Object

Computes a score factor based on a term’s document frequency (the number of documents which contain the term). This value is multiplied by the #tf(int) factor for each term in the query and these products are then summed to form the initial score for a document.

Terms that occur in fewer documents are better indicators of topic, so implemenations of this method usually return larger values for rare terms, and smaller values for common terms.

doc_freq

the number of documents which contain the term

num_docs

the total number of documents in the collection

Return

a score factor based on the term’s document frequency

Raises:

  • (NotImplementedError) 


212
213
214
 
# File 'lib/ferret/search/similarity.rb', line 212

def idf
  raise NotImplementedError
end

#idf_phrase(terms, searcher) ⇒ Object

Computes a score factor for a phrase.

The default implementation sums the #idf(Term,Searcher) factor for each term in the phrase.

terms

the terms in the phrase

searcher

the document collection being searched

Return

a score factor for the phrase



194
195
196
197
198
 
# File 'lib/ferret/search/similarity.rb', line 194

def idf_phrase(terms, searcher)
  idf = 0.0
  terms.each { |term| idf += idf_term(term, searcher) }
  return idf
end

#idf_term(term, searcher) ⇒ Object

Computes a score factor for a simple term.

The default implementation is:

return idf(searcher.doc_freq(term), searcher.max_doc())

Note that Searcher#max_doc() is used instead of IndexReader#num_docs() because it is proportional to Searcher#doc_freq(Term) , i.e., when one is inaccurate, so is the other, and in the same direction.

term

the term in question

searcher

the document collection being searched

Return

a score factor for the term



182
183
184
 
# File 'lib/ferret/search/similarity.rb', line 182

def idf_term(term, searcher)
  return idf(searcher.doc_freq(term), searcher.max_doc())
end

#length_norm ⇒ Object

Computes the normalization value for a field given the total number of terms contained in a field. These values, together with field boosts, are stored in an index and multipled into scores for hits on each field by the search code.

Matches in longer fields are less precise, so implemenations of this method usually return smaller values when num_tokens is large, and larger values when num_tokens is small.

That these values are computed under IndexWriter#add_document and stored then using #encode_norm(float). Thus they have limited precision, and documents must be re-indexed if this method is altered.

field

the name of the field

num_tokens

the total number of tokens contained in fields named field of doc.

See Field#set_boost

Raises:

  • (NotImplementedError) 


99
100
101
 
# File 'lib/ferret/search/similarity.rb', line 99

def length_norm
  raise NotImplementedError
end

#query_norm ⇒ Object

Computes the normalization value for a query given the sum of the squared weights of each of the query terms. This value is then multipled into the weight of each query term.

This does not affect ranking, but rather just attempts to make scores from different queries comparable.

sum_of_squared_weights

the sum of the squares of query term weights

Return

a normalization factor for query weights

Raises:

  • (NotImplementedError) 


112
113
114
 
# File 'lib/ferret/search/similarity.rb', line 112

def query_norm
  raise NotImplementedError
end

#sloppy_freq ⇒ Object

Computes the amount of a sloppy phrase match, based on an edit distance. This value is summed for each sloppy phrase match in a document to form the frequency that is passed to #tf(float).

A phrase match with a small edit distance to a document passage more closely matches the document, so implementations of this method usually return larger values when the edit distance is small and smaller values when it is large.

See PhraseQuery#slop(int)

distance

the edit distance of this sloppy phrase match

Return

the frequency increment for this match

Raises:

  • (NotImplementedError) 


165
166
167
 
# File 'lib/ferret/search/similarity.rb', line 165

def sloppy_freq
  raise NotImplementedError
end

#tf ⇒ Object

Computes a score factor based on a term or phrase’s frequency in a document. This value is multiplied by the #idf_term(Term, Searcher) factor for each term in the query and these products are then summed to form the initial score for a document.

Terms and phrases repeated in a document indicate the topic of the document, so implementations of this method usually return larger values when freq is large, and smaller values when freq is small.

The default implementation calls #tf(float)

freq

the frequency of a term within a document

Return

a score factor based on a term’s within-document frequency

Raises:

  • (NotImplementedError) 


149
150
151
 
# File 'lib/ferret/search/similarity.rb', line 149

def tf
  raise NotImplementedError
end