#!/usr/bin/env python

"""
concatenates JavaScript files in the order specified in an HTML file


Usage:
    $ concat_js.py <input> [output]

input is name of the respective HTML file, output that of the resulting
JavaScript file (STDOUT by default)
"""

import sys

import html5lib


def main(args):
    args = [unicode(arg, "utf-8") for arg in args]
    html = args[1]
    try:
        outfile = args[2]
    except IndexError, exc:
        outfile = None

    with open(html, "r") as f:
        doc = html5lib.parse(f, treebuilder="dom")
        els = doc.getElementsByTagName("script")

        code = []
        for el in els:
            filepath = el.getAttribute("src")
            if filepath:
                with open(filepath, "r") as f:
                    code.append(f.read().decode("UTF-8"))
            else:
                nodes = [node.data for node in el.childNodes]
                code.append(u"".join(nodes))

    code = u"\n".join(code).encode("UTF-8")
    if outfile:
        with open(outfile, "w") as f:
            f.write(code)
    else:
        print code

    return True


if __name__ == "__main__":
    status = not main(sys.argv)
    sys.exit(status)