generate method list from javadoc

gen_methodlist.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import os
import os.path
import sys

import lxml.html
import xlwt

def is_method(method):
    if method.find('(') == -1:
        return False
    if method.find(')') == -1:
        return False
    return True

def write_on_sheet(doc_file, sheet, start, xpath_exp='//a/@name', debug=True):
    root = lxml.html.parse(doc_file)
    methods = [method for method in root.xpath(xpath_exp) if is_method(method)]
    methods.sort()

    # write
    sheet.write(start, 0, os.path.basename(doc_file.split('.')[0])) # remove extention
    for method in methods:
        if debug:
            print method
        sheet.write(start, 1, method)
        start += 1
    return start
    
def main():
    if len(sys.argv) != 2:
        print 'Error'
        sys.exit(1)
    doc_files = []
    extention = 'html'
    exclude = 'package-'
    for path, dirs, files in os.walk(sys.argv[1]):
        doc_files.extend([os.path.join(path, f) for f in files if (f.find(exclude) == -1) and (f.find(extention) != -1)])

    # excel file
    book_name = 'method_list.xls'
    sheet_name = 'All Methods'
    if os.path.exists(book_name):
        try:
            os.remove(book_name)
        except ex:
            print "Sheet can't be removed"
            print ex
            sys.exit(1)

    wb = xlwt.Workbook()
    ws = wb.add_sheet(sheet_name)

    start = 0
    print doc_files
    for doc_file in doc_files:
        start = write_on_sheet(doc_file, ws, start)

    wb.save(book_name)

if __name__ == '__main__':
    main()