Steps

1. Get a few text docs

2. Use WordNet to parse:

• terms, frequency in a document
• insert into database: new document, update TermFrequencyInDoc, TermFrequencyOverall

3. Until the above steps done, process the tfidf for each pair of term-document, store in the score column

4. Begin to search:

• SELECT document_id, sum(score) as T from tfidf_termfrequencyindoc WHERE term IN %s GROUPBY document_id ORDER BY T desc LIMIT 100

• From retrieved document_id, get the real docs