UMLS-Core: Corpus Tree

• Descriptions:
  • Load all normalized synonyms (key in normTerm-synonyms table) into a trie tree

• Examples - Test Cases:

  • Terms:

    Terms
    dog
    dog and cat
    dog and poppy
    canine
    cat
    feline
    k9
    bull dog
    pets

  • Trie Tree (from Synonym Terms):

    

• Algorithm:
  • Trie Node:
    • Trie nodes are used to compose the Trie tree.
    • Each node has a key, which is a (single) word
    • Each node has a level, the top (root) level is 0
    • Each node has a child, which is a Vector<TrieNode>

    • Use "$_ROOT" for the word of root node
    • Use "$_END" for the word of end node

  • Trie Tree:
    • Each branch (path) starts from root node to the end node is a term
    • All terms are stored in the branch of the Trie tree

    • Tokenize input term into words
    • Go through each word:
      • curNode starts from the top (root) node
      • Init curWordNode by curWord
      • Check if Childs contains curWordNode
        • if yes, update curNode to the matched node in Childs
        • If no, add curWordNode to Childs and update curNode to curWordNode