#!/usr/bin/env python
import os, sys, re
oj = os.path.join
from glob import glob

SQL_START = re.compile('^\s*SELECT|UPDATE|CREATE|DELETE|INSERT|BEGIN', re.I|re.M)
TAGS_FORMAT_TMPL = '%s\t%s\t%s\n'

def _notworthit(directory):
    if directory.endswith('CVS'):
        return True
    return False

def _findlinenumber(code, charstart):
    t = 0
    for i, c in enumerate(code):
        if c == '\n':
            t += 1
        if i >= charstart:
            return t + 1
    return t
    

def append(t, d):
    fa = open(t, 'a')
    for sqlfile in glob(oj(d, '*.sql')):
        tagname = os.path.basename(sqlfile).replace('.sql','')
        tagfile = sqlfile.replace(os.path.dirname(t),'')[1:]
        sql = open(sqlfile).read()
        match = SQL_START.search(sql)
        if match:
            #print sqlfile, match.start()
            linenumber = _findlinenumber(sql, match.start())
            line = TAGS_FORMAT_TMPL % (tagname, tagfile, linenumber)
            fa.write(line)
        else:
            print "PROB ", repr(sqlfile)
    fa.close()
    for f in os.listdir(d):
        if os.path.isdir(oj(d, f)):
            if _notworthit(oj(d, f)):
                continue
            append(t, oj(d, f))
    
    
def run(t, d):
    """ index all .sql files recursively and enter into 
    the tags file.
    For more info on the TAGS format, see
    http://ctags.sourceforge.net/FORMAT
    """
    append(t, d)
    if not os.path.isfile(t):
        raise OSError, "File %r does not exist" % t
    
    return 0

if __name__=='__main__':
    tagsfile = os.path.abspath(sys.argv[1])
    directory = os.path.dirname(tagsfile)
    sys.exit(run(tagsfile, directory))