Syntactic Uninvert

• Short Description: Syntactic uninvert the input phrase.
  
  
• Full Description:

  This flow strips out non-information words (such as NOS, NEC); and then uninvert phrase around commas if the words after comma is not a conjunction words (such as about, beneath, where, yet).

  Non-information words are defined in file "data/misc/nonInfoWords.data". Conjunction words are defined in file "data/misc/conjunctionWord.data". Both non-information words and conjunction words are configurable.

  No effect on the -m option. "none" is added at the end of the output.
  
  

• Difference:
  • The Java version drops comma if comma is at the end of the word after stripping the non-information words, such as "Kyphosis, NOS".
  • The Java version does not uninvert words if it is a conjunction word shows up after a comma at any position of the words; while the "C" version does not uninvert only if a conjunction word show up after the first comma.
  
  
  
• Features:
  1. Strip non-information words.
  2. Tokenize phrase by using commas as delimiters.
  3. Check if the word after commas is a conjunction word.
  4. Uninvert input phrases around commas if above condition is false.
  
  
  
• Symbol: S
  
  
• Examples:

  shell> lvg -f:S

  
  Angioplasty, Transluminal, Percutaneous Coronary
  Angioplasty, Transluminal, Percutaneous Coronary|Percutaneous Coronary Transluminal Angioplasty|2047|16777215|S|1|
  
  Kyphosis, NOS
  Kyphosis, NOS|Kyphosis|2047|16777215|S|1|
  
  Sedative, hypnotic, or anxiolytic amnestic disorder
  Sedative, hypnotic, or anxiolytic amnestic disorder|Sedative, hypnotic, or anxiolytic amnestic disorder|2047|16777215|S|1|
  

  More examples
  
  

• Implementation Logic:
  1. Strip non-information words.
  2. Tokenize phrases by using commas as delimiters.
  3. Compose tokens in the reverse order if the word after comma is not a conjunction words.
  
  
• Source Code: ToSyntacticUninvert.java
  
  
• Hierarchy: Object -> Transformation -> ToSyntacticUninvert