tornado_static.py

"""
tornado_static is a module for displaying static resources in a Tornado web
application.

It can take care of merging, compressing and giving URLs ideal renamings
suitable for aggressive HTTP caching.

(c) [email protected]
"""

__version__ = '1.3'

import os
import cPickle
import re
import stat
import marshal
from time import time
from tempfile import gettempdir
from base64 import encodestring
from subprocess import Popen, PIPE
import tornado.web

try:
    import cssmin
except ImportError:
    cssmin = None

def mkdir(newdir):
    """works the way a good mkdir should :)
        - already exists, silently complete
        - regular file in the way, raise an exception
        - parent directory(ies) does not exist, make them as well
    """
    if os.path.isdir(newdir):
        pass
    elif os.path.isfile(newdir):
        raise OSError("a file with the same name as the desired " \
                    "dir, '%s', already exists." % newdir)
    else:
        head, tail = os.path.split(newdir)
        if head and not os.path.isdir(head):
            _mkdir(head)
        if tail:
            os.mkdir(newdir)

################################################################################
# Global variable where we store the conversions so we don't have to do them
# again every time the UI module is rendered with the same input

out_file = os.path.join(os.path.abspath(os.curdir), '.static_name_conversion')
def _delete_old_static_name_conversion():
    """In this app we marshal all static file conversion into a file called
    '.static_name_conversion' located here in the working directory.
    The reason we're doing this is so that when you start multiple Python
    interpreters of the app (e.g. production environment) you only need to
    work out which name conversions have been done once.

    When you do a new deployment it's perfectly natural that this name
    conversion should be invalidated since there are now potentially new static
    resources so it needs to have different static names.

    So delete the file if it's older than a small amount of time in a human
    sense.
    """
    if os.path.isfile(out_file):
        mtime = os.stat(out_file)[stat.ST_MTIME]
        age = time() - mtime
        if age >= 60:
            os.remove(out_file)

def load_name_conversion():
    try:
        return marshal.load(file(out_file))
    except IOError:
        return dict()

_delete_old_static_name_conversion()
_name_conversion = load_name_conversion()

def save_name_conversion():
    marshal.dump(_name_conversion, file(out_file, 'w'))

class StaticURL(tornado.web.UIModule):

    def render(self, *static_urls):
        # the following 4 lines will have to be run for every request. Since
        # it's just a basic lookup on a dict it's going to be uber fast.
        basic_name = ''.join(static_urls)
        already = _name_conversion.get(basic_name)
        if already:
            cdn_prefix = self.handler.get_cdn_prefix()
            if cdn_prefix:
                already = cdn_prefix + already
            return already

        new_name = self._combine_filename(static_urls)
        # If you run multiple tornados (on different ports) it's possible
        # that another process has already dealt with this static URL.
        # Therefore we now first of all need to figure out what the final name
        # is going to be
        youngest = 0
        full_paths = []
        old_paths = {} # maintain a map of what the filenames where before
        for path in static_urls:
            full_path = os.path.join(
              self.handler.settings['static_path'], path)
            #f = open(full_path)
            mtime = os.stat(full_path)[stat.ST_MTIME]
            if mtime > youngest:
                youngest = mtime
            full_paths.append(full_path)
            old_paths[full_path] = path

        n, ext = os.path.splitext(new_name)
        new_name = "%s.%s%s" % (n, youngest, ext)
        if os.path.isfile(new_name):
            # conversion and preparation has already been done!
            # No point doing it again, so just exit here
            pass
        else:
            destination = file(new_name, 'w')
            do_optimize_static_content = self.handler.settings\
              .get('optimize_static_content', True)

            if do_optimize_static_content:
                uglifyjs_location = self.handler\
                  .settings.get('UGLIFYJS_LOCATION')
                closure_location = self.handler\
                  .settings.get('CLOSURE_LOCATION')
                yui_location = self.handler\
                  .settings.get('YUI_LOCATION')

            for full_path in full_paths:
                f = open(full_path)
                code = f.read()
                if full_path.endswith('.js'):
                    if len(full_paths) > 1:
                        destination.write('/* %s */\n' % os.path.basename(full_path))
                    if do_optimize_static_content and not self._already_optimized_filename(full_path):
                        if uglifyjs_location:
                            code = run_uglify_js_compiler(code, uglifyjs_location,
                              verbose=self.handler.settings.get('debug', False))
                        elif closure_location:
                        #if closure_location:
                            orig_code = code
                            code = run_closure_compiler(code, closure_location,
                              verbose=self.handler.settings.get('debug', False))
                        elif yui_location:
                            code = run_yui_compressor(code, 'js', yui_location,
                          verbose=self.handler.settings.get('debug', False))
                elif full_path.endswith('.css'):
                    if len(full_paths) > 1:
                        destination.write('/* %s */\n' % os.path.basename(full_path))
                    if do_optimize_static_content and not self._already_optimized_filename(full_path):
                        if cssmin is not None:
                            code = cssmin.cssmin(code)
                        elif yui_location:
                            code = run_yui_compressor(code, 'css', yui_location,
                              verbose=self.handler.settings.get('debug', False))
                    # do run this after the run_yui_compressor() has been used so that
                    # code that is commented out doesn't affect
                    code = self._replace_css_images_with_static_urls(
                      code,
                      os.path.dirname(old_paths[full_path])
                      )
                else:
                    # this just copies the file
                    pass
                    #raise ValueError("Unknown extension %s" % full_path)
                destination.write(code)
                destination.write("\n")

            destination.close()
        prefix = self.handler.settings.get('combined_static_url_prefix', '/combined/')
        new_name = os.path.join(prefix, os.path.basename(new_name))
        _name_conversion[basic_name] = new_name
        save_name_conversion()

        ## Commented out, because I don't want to use CDN when it might take 5 seconds
        # to generate the new file.
        #cdn_prefix = self.handler.get_cdn_prefix()
        #if cdn_prefix:
        #    new_name = cdn_prefix + new_name
        return new_name


    def _combine_filename(self, names, max_length=60):
        # expect the parameter 'names' be something like this:
        # ['css/foo.css', 'css/jquery/datepicker.css']
        # The combined filename is then going to be
        # "/tmp/foo.datepicker.css"
        first_ext = os.path.splitext(names[0])[-1]
        save_dir = self.handler.application.settings.get('combined_static_dir')
        if save_dir is None:
            save_dir = os.environ.get('TMP_DIR')
            if not save_dir:
                save_dir = gettempdir()
        save_dir = os.path.join(save_dir, 'combined')
        mkdir(save_dir)
        combined_name = []
        for name in names:
            name, ext = os.path.splitext(os.path.basename(name))
            if ext != first_ext:
                raise ValueError("Mixed file extensions (%s, %s)" %\
                 (first_ext, ext))
            combined_name.append(name)
        if sum(len(x) for x in combined_name) > max_length:
            combined_name = [x.replace('.min','.m').replace('.pack','.p')
                             for x in combined_name]
            combined_name = [re.sub(r'-[\d\.]+', '', x) for x in combined_name]
            while sum(len(x) for x in combined_name) > max_length:
                try:
                    combined_name = [x[-2] == '.' and x[:-2] or x[:-1]
                                 for x in combined_name]
                except IndexError:
                    break

        combined_name.append(first_ext[1:])
        return os.path.join(save_dir, '.'.join(combined_name))

    def _replace_css_images_with_static_urls(self, css_code, rel_dir):
        def replacer(match):
            filename = match.groups()[0]
            if (filename.startswith('"') and filename.endswith('"')) or \
              (filename.startswith("'") and filename.endswith("'")):
                filename = filename[1:-1]
            if 'data:image' in filename or filename.startswith('http://'):
                return filename
            if filename == '.':
                # this is a known IE hack in CSS
                return filename
            # It's really quite common that the CSS file refers to the file
            # that doesn't exist because if you refer to an image in CSS for
            # a selector you never use you simply don't suffer.
            # That's why we say not to warn on nonexisting files
            new_filename = self.handler.static_url(os.path.join(rel_dir, filename))
            return match.group().replace(filename, new_filename)
        _regex = re.compile('url\(([^\)]+)\)')
        css_code = _regex.sub(replacer, css_code)

        return css_code

    def _already_optimized_filename(self, file_path):
        file_name = os.path.basename(file_path)
        for part in ('.min.', '.minified.', '.pack.', '-jsmin.'):
            if part in file_name:
                return True
        return False


class Static(StaticURL):
    """given a list of static resources, return the whole HTML tag"""
    def render(self, *static_urls, **options):
        extension = static_urls[0].split('.')[-1]
        if extension == 'css':
            template = '<link rel="stylesheet" type="text/css" href="%(url)s">'
        elif extension == 'js':
            template = '<script '
            if 'defer' in options:
                template += 'defer '
            elif 'async' in options:
                template += 'async '
            template += 'src="%(url)s"></script>'
        else:
            raise NotImplementedError
        url = super(Static, self).render(*static_urls)
        return template % dict(url=url)


def run_closure_compiler(code, jar_location, verbose=False): # pragma: no cover
    if verbose:
        t0 = time()
    r = _run_closure_compiler(code, jar_location)
    if verbose:
        t1 = time()
        a, b = len(code), len(r)
        c = round(100 * float(b) / a, 1)
        print "Closure took", round(t1 - t0, 4),
        print "seconds to compress %d bytes into %d (%s%%)" % (a, b, c)
    return r

def _run_closure_compiler(jscode, jar_location, advanced_optmization=False): # pragma: no cover
    cmd = "java -jar %s " % jar_location
    if advanced_optmization:
        cmd += " --compilation_level ADVANCED_OPTIMIZATIONS "
    proc = Popen(cmd, shell=True, stdout=PIPE, stdin=PIPE, stderr=PIPE)
    try:
        (stdoutdata, stderrdata) = proc.communicate(jscode)
    except OSError, msg:
        # see comment on OSErrors inside _run_yui_compressor()
        stderrdata = \
          "OSError: %s. Try again by making a small change and reload" % msg
    if stderrdata:
        return "/* ERRORS WHEN RUNNING CLOSURE COMPILER\n" + stderrdata + '\n*/\n' + jscode
    return stdoutdata

def run_uglify_js_compiler(code, location, verbose=False): # pragma: no cover
    if verbose:
        t0 = time()
    r = _run_uglify_js_compiler(code, location)
    if verbose:
        t1 = time()
        a, b = len(code), len(r)
        c = round(100 * float(b) / a, 1)
        print "UglifyJS took", round(t1 - t0, 4),
        print "seconds to compress %d bytes into %d (%s%%)" % (a, b, c)
    return r

def _run_uglify_js_compiler(jscode, location, options=''): # pragma: no cover
    cmd = "%s %s" % (location, options)
    proc = Popen(cmd, shell=True, stdout=PIPE, stdin=PIPE, stderr=PIPE)
    try:
        (stdoutdata, stderrdata) = proc.communicate(jscode)
    except OSError, msg:
        # see comment on OSErrors inside _run_yui_compressor()
        stderrdata = \
          "OSError: %s. Try again by making a small change and reload" % msg
    if stderrdata:
        return "/* ERRORS WHEN RUNNING UGLIFYJS COMPILER\n" + stderrdata + '\n*/\n' + jscode
    return stdoutdata

def run_yui_compressor(code, type_, jar_location, verbose=False): # pragma: no cover
    if verbose:
        t0 = time()
    r = _run_yui_compressor(code, type_, jar_location)
    if verbose:
        t1 = time()
        a, b = len(code), len(r)
        c = round(100 * float(b) / a, 1)
        print "YUI took", round(t1 - t0, 4),
        print "seconds to compress %d bytes into %d (%s%%)" % (a, b, c)
    return r

def _run_yui_compressor(code, type_, jar_location):
    cmd = "java -jar %s --type=%s" % (jar_location, type_)
    proc = Popen(cmd, shell=True, stdout=PIPE, stdin=PIPE, stderr=PIPE)
    try:
        (stdoutdata, stderrdata) = proc.communicate(code)
    except OSError, msg:
        # Sometimes, for unexplicable reasons, you get a Broken pipe when
        # running the popen instance. It's always non-deterministic problem
        # so it probably has something to do with concurrency or something
        # really low level.
        stderrdata = \
          "OSError: %s. Try again by making a small change and reload" % msg

    if stderrdata:
        return "/* ERRORS WHEN RUNNING YUI COMPRESSOR\n" + stderrdata + '\n*/\n' + code

    return stdoutdata


class PlainStaticURL(tornado.web.UIModule):
    def render(self, url):
        return self.handler.static_url(url)

class PlainStatic(tornado.web.UIModule):
    """Render the HTML that displays a static resource without any optimization
    or combing.
    """

    def render(self, *static_urls, **options):
        extension = static_urls[0].split('.')[-1]
        if extension == 'css':
            template = '<link rel="stylesheet" type="text/css" href="%(url)s">'
        elif extension == 'js':
            template = '<script '
            if 'defer' in options:
                template += 'defer '
            elif 'async' in options:
                template += 'async '
            template += 'src="%(url)s"></script>'
        else:
            raise NotImplementedError

        html = []
        for each in static_urls:
            url = self.handler.static_url(each)
            html.append(template % dict(url=url))
        return "\n".join(html)


_base64_conversion_file = '.base64-image-conversions.pickle'
try:
    _base64_conversions = cPickle.load(file(_base64_conversion_file))
    #raise IOError
except IOError:
    _base64_conversions = {}

class Static64(tornado.web.UIModule):
    def render(self, image_path):
        already = _base64_conversions.get(image_path)
        if already:
            return already

        template = 'data:image/%s;base64,%s'
        extension = os.path.splitext(os.path.basename(image_path))
        extension = extension[-1][1:]
        assert extension in ('gif','png'), extension
        full_path = os.path.join(
              self.handler.settings['static_path'], image_path)
        data = encodestring(file(full_path,'rb').read()).replace('\n','')#.replace('\n','\\n')
        result = template % (extension, data)

        _base64_conversions[image_path] = result
        cPickle.dump(_base64_conversions, file(_base64_conversion_file, 'wb'))
        return result