HTML/XML tags are dropped from the Postgres full text index and from search queries.

Example:

    launchpad_dev=# select to_tsvector('aaa <div>bbb</div> ccc');
           to_tsvector
    -------------------------
     'aaa':1 'bbb':2 'ccc':3
    (1 row)

    launchpad_dev=# select to_tsquery('aaa & <div>bbb</div> & ccc');
          to_tsquery
    -----------------------
     'aaa' & 'bbb' & 'ccc'
    (1 row)

    launchpad_dev=# select to_tsquery('aaa & <div> & bbb & </div> & ccc');
          to_tsquery
    -----------------------
     'aaa' & 'bbb' & 'ccc'
    (1 row)

    launchpad_dev=# select ts_debug('aaa <div>bbb</div> ccc');
                                  ts_debug
    ---------------------------------------------------------------------
     (asciiword,"Word, all ASCII",aaa,{english_stem},english_stem,{aaa})
     (blank,"Space symbols"," ",{},,)
     (tag,"XML tag",<div>,{},,)
     (asciiword,"Word, all ASCII",bbb,{english_stem},english_stem,{bbb})
     (tag,"XML tag",</div>,{},,)
     (blank,"Space symbols"," ",{},,)
     (asciiword,"Word, all ASCII",ccc,{english_stem},english_stem,{ccc})
    (7 rows)

So, strings like '<div>' are treated as tokens of type "tag" -- but these
tokens do not appear in the FTI data, and they do not appear the the
result of to_tsquery().