documentation cleanup

Author: tiagolv

docs-site/content/docs/Documentation/core/Levenshtein.mdx

@@ -154,7 +154,6 @@ The `Levenshtein` class offers methods to measure the difference between two str
           if len1 <= 3:
               return Levenshtein._simple_distance(seq1, seq2)
 
-          # Optimized algorithm with only two rows (memory efficient)
           # Instead of full matrix, we only keep previous and current row
           previous_row = list(range(len2 + 1))
           current_row = [0] * (len2 + 1)

docs-site/content/docs/Documentation/data/composedword.mdx

@@ -73,7 +73,6 @@ The `ComposedWord` class stores and aggregates information about multi-word keyw
           self._h = 1.0
 
           # Check if the candidate starts or ends with stopwords
-          # Optimized: use truthiness instead of len() > 0
           if self._terms:
               self._start_or_end_stopwords = (
                   self._terms[0].stopword or self._terms[-1].stopword
@@ -203,7 +202,6 @@ The class provides property accessors for backward compatibility:
               tf_used = self.tf
 
           # For virtual candidates, use mean frequency of constituent terms
-          # Optimized: use built-in sum/len instead of numpy for small lists
           if is_virtual:
               tfs = [term_obj.tf for term_obj in self.terms]
               tf_used = sum(tfs) / len(tfs) if tfs else 1.0
@@ -476,7 +474,6 @@ The class provides property accessors for backward compatibility:
               tf_used = self.tf
 
           # For virtual candidates, use mean frequency of constituent terms
-          # Optimized: use built-in sum/len instead of numpy for small lists
           if is_virtual:
               tfs = [term_obj.tf for term_obj in self.terms]
               tf_used = sum(tfs) / len(tfs) if tfs else 1.0

yake/core/Levenshtein.py

@@ -111,7 +111,6 @@ def distance(seq1: str, seq2: str) -> int:
         if len1 <= 3:
             return Levenshtein._simple_distance(seq1, seq2)
 
-        # Optimized algorithm with only two rows (memory efficient)
         previous_row = list(range(len2 + 1))
         current_row = [0] * (len2 + 1)
 

yake/data/composed_word.py

@@ -70,7 +70,6 @@ def __init__(self, terms: Optional[List[Tuple[str, str, Any]]]):
         self._h = 1.0
 
         # Check if the candidate starts or ends with stopwords
-        # Optimized: use truthiness instead of len() > 0
         if self._terms:
             self._start_or_end_stopwords = (
                 self._terms[0].stopword or self._terms[-1].stopword
@@ -389,7 +388,6 @@ def update_h(self, features=None, is_virtual=False):
             tf_used = self.tf
 
         # For virtual candidates, use mean frequency of constituent terms
-        # Optimized: use built-in sum/len instead of numpy for small lists
         if is_virtual:
             tfs = [term_obj.tf for term_obj in self.terms]
             tf_used = sum(tfs) / len(tfs) if tfs else 1.0
@@ -450,7 +448,6 @@ def update_h_old(self, features=None, is_virtual=False):
             tf_used = self.tf
 
         # For virtual candidates, use mean frequency of constituent terms
-        # Optimized: use built-in sum/len instead of numpy for small lists
         if is_virtual:
             tfs = [term_obj.tf for term_obj in self.terms]
             tf_used = sum(tfs) / len(tfs) if tfs else 1.0

yake/data/core.py

@@ -247,10 +247,8 @@ def _process_sentence(
         # Process each word in the sentence
         for pos_sent, word in enumerate(sentence):
             # Check if the word is just punctuation (all characters are excluded)
-            # Optimized: use all() instead of creating a list
             if all(c in self.exclude for c in word):
                 # If we have a block of words, save it and start a new block
-                # Optimized: use truthiness instead of len() > 0
                 if block_of_word_obj:
                     sentence_obj_aux.append(block_of_word_obj)
                     block_of_word_obj = []
@@ -266,12 +264,10 @@ def _process_sentence(
                 )
 
         # Save any remaining word block
-        # Optimized: use truthiness instead of len() > 0
         if block_of_word_obj:
             sentence_obj_aux.append(block_of_word_obj)
 
         # Add processed sentence to collection if not empty
-        # Optimized: use truthiness instead of len() > 0
         if sentence_obj_aux:
             self.sentences_obj.append(sentence_obj_aux)
 
@@ -466,7 +462,6 @@ def build_single_terms_features(self, features: Optional[List[str]] = None) -> N
         valid_tfs = np.array([x.tf for x in valid_terms])
 
         # Skip if no valid terms
-        # Optimized: use 'not' instead of len() == 0
         if not valid_tfs.size:
             return
 

yake/data/utils.py

@@ -120,7 +120,6 @@ def get_tag(word: str, i: int, exclude: frozenset) -> str:
         return "d"
 
     # Count character types for classification
-    # Optimized: single pass through word instead of multiple
     cdigit = calpha = cexclude = 0
     for c in word:
         if c.isdigit():