More https fixes

[personal/website.git] / source / blog / blog.cgi

#!/usr/bin/env python
#coding: utf8

# blitiri - A single-file blog engine.
# Alberto Bertogli (albertito@gmail.com)

#
# Configuration section
#
# You can edit these values, or create a file named "config.py" and put them
# there to make updating easier. The ones in config.py take precedence.
#

# Directory where entries are stored
data_path = "/tmp/blog/data"

# Are comments allowed? (if False, comments_path option is not used)
enable_comments = False

# Directory where comments are stored (must be writeable by the web server)
comments_path = "/tmp/blog/comments"

# Path where templates are stored. Use an empty string for the built-in
# default templates. If they're not found, the built-in ones will be used.
templates_path = "/tmp/blog/templates"

# Path where the cache is stored (must be writeable by the web server);
# set to None to disable. When enabled, you must take care of cleaning it up
# every once in a while.
#cache_path = "/tmp/blog/cache"
cache_path = None

# URL to the blog, including the name. Can be a full URL or just the path.
blog_url = "/blog/blitiri.cgi"

# Style sheet (CSS) URL. Can be relative or absolute. To use the built-in
# default, set it to blog_url + "/style".
css_url = blog_url + "/style"

# Blog title
title = "I don't like blogs"

# Default author
author = "Hartmut Kegan"

# Article encoding
encoding = "utf8"

# Captcha method to use. At the moment only "title" is supported, but if you
# are keen with Python you can provide your own captcha implementation, see
# below for details.
captcha_method = "title"

# How many articles to show in the index
index_articles = 10


#
# End of configuration
# DO *NOT* EDIT ANYTHING PAST HERE
#


import sys
import os
import errno
import shutil
import time
import datetime
import calendar
import zlib
import urllib
import cgi
from docutils.core import publish_parts
from docutils.utils import SystemMessage

# Before importing the config, add our cwd to the Python path
sys.path.append(os.getcwd())

# Load the config file, if there is one
try:
        from config import *
except:
        pass


# Pimp *_path config variables to support relative paths
data_path = os.path.realpath(data_path)
templates_path = os.path.realpath(templates_path)


#
# Captcha classes
#
# They must follow the interface described below.
#
# Constructor:
#       Captcha(article) -> constructor, takes an article[1] as argument
# Attributes:
#       puzzle -> a string with the puzzle the user must solve to prove he is
#                 not a bot (can be raw HTML)
#       help -> a string with extra instructions, shown only when the user
#               failed to solve the puzzle
# Methods:
#       validate(form_data) -> based on the form data[2],  returns True if
#                              the user has solved the puzzle uccessfully
#                              (False otherwise).
#
# Note you must ensure that the puzzle attribute and validate() method can
# "communicate" because they are executed in different requests. You can pass a
# cookie or just calculate the answer based on the article's data, for example.
#
# [1] article is an object with all the article's information:
#       path -> string
#       created -> datetime
#       updated -> datetime
#       uuid -> string (unique ID)
#       title -> string
#       author -> string
#       tags -> list of strings
#       raw_contents -> string in rst format
#       comments -> list of Comment objects (not too relevant here)
# [2] form_data is an object with the form fields (all strings):
#       author, author_error
#       link, link_error
#       catpcha, captcha_error
#       body, body_error
#       action, method

class TitleCaptcha (object):
        "Captcha that uses the article's title for the puzzle"
        def __init__(self, article):
                self.article = article
                words = article.title.split()
                self.nword = hash(article.title) % len(words) % 5
                self.answer = words[self.nword]
                self.help = 'gotcha, damn spam bot!'

        @property
        def puzzle(self):
                nword = self.nword + 1
                if nword == 1:
                        n = '1st'
                elif nword == 2:
                        n = '2nd'
                elif nword == 3:
                        n = '3rd'
                else:
                        n = str(nword) + 'th'
                return "enter the %s word of the article's title" % n

        def validate(self, form_data):
                if form_data.captcha.lower() == self.answer.lower():
                        return True
                return False

known_captcha_methods = {
        'title': TitleCaptcha,
}

# If the configured captcha method was a known string, replace it by the
# matching class; otherwise assume it's already a class and leave it
# alone. This way the user can either use one of our methods, or provide one
# of his/her own.
if captcha_method in known_captcha_methods:
        captcha_method = known_captcha_methods[captcha_method]


# Default template

default_main_header = """\
<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
          "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">

<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<link rel="alternate" title="%(title)s" href="%(fullurl)s/atom"
        type="application/atom+xml" />
<link href="%(css_url)s" rel="stylesheet" type="text/css" />
<title>%(title)s</title>
</head>

<body>

<h1><a href="%(url)s">%(title)s</a></h1>

<div class="content">
"""

default_main_footer = """
</div>
<div class="footer">
  %(showyear)s: %(monthlinks)s<br/>
  years: %(yearlinks)s<br/>
  subscribe: <a href="%(url)s/atom">atom</a><br/>
  views: <a href="%(url)s/">blog</a> <a href="%(url)s/list">list</a><br/>
  tags: %(taglinks)s<br/>
</div>

</body>
</html>
"""

default_article_header = """
<div class="article">
<h2><a href="%(url)s/post/%(uuid)s">%(arttitle)s</a></h2>
<span class="artinfo">
  by %(author)s on <span class="date">

<a class="date" href="%(url)s/%(cyear)d/">%(cyear)04d</a>-\
<a class="date" href="%(url)s/%(cyear)d/%(cmonth)d/">%(cmonth)02d</a>-\
<a class="date" href="%(url)s/%(cyear)d/%(cmonth)d/%(cday)d/">%(cday)02d</a>\
    %(chour)02d:%(cminute)02d</span>
  (updated on <span class="date">
<a class="date" href="%(url)s/%(uyear)d/">%(uyear)04d</a>-\
<a class="date" href="%(url)s/%(uyear)d/%(umonth)d/">%(umonth)02d</a>-\
<a class="date" href="%(url)s/%(uyear)d/%(umonth)d/%(uday)d/">%(uday)02d</a>\
    %(uhour)02d:%(uminute)02d)</span><br/>
  <span class="tags">tagged %(tags)s</span> -
  <span class="comments">with %(comments)s
    <a href="%(url)s/post/%(uuid)s#comments">comment(s)</a></span>
</span><br/>
<p/>
<div class="artbody">
"""

default_article_footer = """
<p/>
</div>
</div>
"""

default_comment_header = """
<div class="comment">
<a name="comment-%(number)d" />
<h3><a href="#comment-%(number)d">Comment #%(number)d</a></h3>
<span class="cominfo">by %(linked_author)s
  on %(year)04d-%(month)02d-%(day)02d %(hour)02d:%(minute)02d</span>
<p/>
<div class="combody">
"""

default_comment_footer = """
<p/>
</div>
</div>
"""

default_comment_form = """
<div class="comform">
<a name="comment" />
<h3 class="comform"><a href="#comment">Your comment</a></h3>
<div class="comforminner">
<form method="%(form_method)s" action="%(form_action)s">
<div class="comformauthor">
  <label for="comformauthor">Your name %(form_author_error)s</label>
  <input type="text" class="comformauthor" id="comformauthor"
         name="comformauthor" value="%(form_author)s" />
</div>
<div class="comformlink">
  <label for="comformlink">Your link
    <span class="comformoptional">(optional, will be published)</span>
      %(form_link_error)s</label>
  <input type="text" class="comformlink" id="comformlink"
         name="comformlink" value="%(form_link)s" />
  <div class="comformhelp">
    like <span class="formurlexample">http://www.example.com/</span>
    or <span class="formurlexample">mailto:you@example.com</span>
  </div>
</div>
<div class="comformcaptcha">
  <label for="comformcaptcha">Your humanity proof %(form_captcha_error)s</label>
  <input type="text" class="comformcaptcha" id="comformcaptcha"
         name="comformcaptcha" value="%(form_captcha)s" />
  <div class="comformhelp">%(captcha_puzzle)s</div>
</div>
<div class="comformbody">
  <label for="comformbody" class="comformbody">The comment
    %(form_body_error)s</label>
  <textarea class="comformbody" id="comformbody" name="comformbody" rows="15"
            cols="80">%(form_body)s</textarea>
  <div class="comformhelp">
    in
    <a href="http://docutils.sourceforge.net/docs/user/rst/quickref.html">\
RestructuredText</a> format, please
  </div>
</div>
<div class="comformsend">
  <button type="submit" class="comformsend" id="comformsend" name="comformsend">
    Send comment
  </button>
</div>
</form>
</div>
</div>
"""

default_comment_error = '<span class="comformerror">(%(error)s)</span>'


# Default CSS
default_css = """
body {
        font-family: sans-serif;
        font-size: small;
        width: 52em;
}

div.content {
        width: 96%;
}

h1 {
        font-size: large;
        border-bottom: 2px solid #99F;
        width: 100%;
        margin-bottom: 1em;
}

h2 {
        font-size: medium;
        font-weigth: none;
        margin-bottom: 1pt;
        border-bottom: 1px solid #99C;
}

h3 {
        font-size: small;
        font-weigth: none;
        margin-bottom: 1pt;
        border-bottom: 1px solid #99C;
}

h1 a, h2 a, h3 a {
        text-decoration: none;
        color: black;
}

span.artinfo {
        font-size: xx-small;
}

span.artinfo a {
        text-decoration: none;
        color: #339;
}

span.artinfo a:hover {
        text-decoration: none;
        color: blue;
}

div.artbody {
        margin-left: 1em;
}

div.article {
        margin-bottom: 2em;
}

span.cominfo {
        font-size: xx-small;
}

span.cominfo a {
        text-decoration: none;
        color: #339;
}

span.cominfo a:hover {
        text-decoration: none;
        color: blue;
}

div.combody {
        margin-left: 2em;
}

div.comment {
        margin-left: 1em;
        margin-bottom: 1em;
}

div.comforminner {
        margin-left: 2em;
}

div.comform {
        margin-left: 1em;
        margin-bottom: 1em;
}

div.comform label {
        display: block;
        border-bottom: 1px solid #99C;
        margin-top: 0.5em;
        clear: both;
}

div.comform span.comformoptional {
        font-size: xx-small;
        color: #666;
}

div.comform input {
        font-size: small;
        width: 99%;
}

div.comformhelp {
        font-size: xx-small;
        text-align: right;
        float: right;
}

span.formurlexample {
        color: #111;
        background-color: #EEF;
        font-family: monospace;
        padding-left: 0.2em;
        padding-right: 0.2em;
}

textarea.comformbody {
        font-family: monospace;
        font-size: small;
        width: 99%;
        height: 15em;
}

button.comformsend {
        margin-top: 0.5em;
}

span.comformerror {
        color: #900;
        font-size: xx-small;
        margin-left: 0.5em;
}

hr {
        float: left;
        height: 2px;
        border: 0;
        background-color: #99F;
        width: 60%;
}

div.footer {
        margin-top: 1em;
        padding-top: 0.4em;
        width: 100%;
        border-top: 2px solid #99F;
        font-size: x-small;
}

div.footer a {
        text-decoration: none;
}

/* Articles are enclosed in <div class="section"> */
div.section h1 {
        font-size: small;
        font-weigth: none;
        width: 100%;
        margin-bottom: 1pt;
        border-bottom: 1px dotted #99C;
}

"""


# Cache decorator
# It only works if the function is pure (that is, its return value depends
# only on its arguments), and if all the arguments are hash()eable.
def cached(f):
        # do not decorate if the cache is disabled
        if cache_path is None:
                return f

        def decorate(*args, **kwargs):
                hashes = '-'.join( str(hash(x)) for x in args +
                                tuple(kwargs.items()) )
                fname = 'blitiri.%s.%s.cache' % (f.__name__, hashes)
                cache_file = os.path.join(cache_path, fname)
                try:
                        s = open(cache_file).read()
                except:
                        s = f(*args, **kwargs)
                        open(cache_file, 'w').write(s)
                return s

        return decorate


# helper functions
@cached
def rst_to_html(rst, secure = True):
        settings = {
                'input_encoding': encoding,
                'output_encoding': 'utf8',
                'halt_level': 1,
                'traceback':  1,
                'file_insertion_enabled': secure,
                'raw_enabled': secure,
        }
        parts = publish_parts(rst, settings_overrides = settings,
                                writer_name = "html")
        return parts['body'].encode('utf8')

def validate_rst(rst, secure = True):
        try:
                rst_to_html(rst, secure)
                return None
        except SystemMessage, e:
                desc = e.args[0].encode('utf-8') # the error string
                desc = desc[9:] # remove "<string>:"
                line = int(desc[:desc.find(':')] or 0) # get the line number
                desc = desc[desc.find(')')+2:-1] # remove (LEVEL/N)
                try:
                        desc, context = desc.split('\n', 1)
                except ValueError:
                        context = ''
                if desc.endswith('.'):
                        desc = desc[:-1]
                return (line, desc, context)

def valid_link(link):
        import re
        scheme_re = r'^[a-zA-Z]+:'
        mail_re = r"^[^ \t\n\r@<>()]+@[a-z0-9][a-z0-9\.\-_]*\.[a-z]+$"
        url_re = r'^(?:[a-z0-9\-]+|[a-z0-9][a-z0-9\-\.\_]*\.[a-z]+)' \
                        r'(?::[0-9]+)?(?:/.*)?$'

        if re.match(scheme_re, link, re.I):
                scheme, rest = link.split(':', 1)
                # if we have an scheme and a rest, assume the link is valid
                # and return it as-is; otherwise (having just the scheme) is
                # invalid
                if rest:
                        return link
                return None

        # at this point, we don't have a scheme; we will try to recognize some
        # common addresses (mail and http at the moment) and complete them to
        # form a valid link, if we fail we will just claim it's invalid
        if re.match(mail_re, link, re.I):
                return 'mailto:' + link
        elif re.match(url_re, link, re.I):
                return 'https://' + link

        return None

def sanitize(obj):
        return cgi.escape(obj, quote = True)


# find out our URL, needed for syndication
try:
        n = os.environ['SERVER_NAME']
        p = os.environ['SERVER_PORT']
        s = os.environ['SCRIPT_NAME']
        if p == '80': p = ''
        else: p = ':' + p
        full_url = 'https://%s%s%s' % (n, p, s)
except KeyError:
        full_url = 'Not needed'


class Templates (object):
        def __init__(self, tpath, db, showyear = None):
                self.tpath = tpath
                self.db = db
                now = datetime.datetime.now()
                if not showyear:
                        showyear = now.year

                self.vars = {
                        'css_url': css_url,
                        'title': title,
                        'url': blog_url,
                        'fullurl': full_url,
                        'year': now.year,
                        'month': now.month,
                        'day': now.day,
                        'showyear': showyear,
                        'monthlinks': ' '.join(db.get_month_links(showyear)),
                        'yearlinks': ' '.join(db.get_year_links()),
                        'taglinks': ' '.join(db.get_tag_links()),
                }

        def get_template(self, page_name, default_template, extra_vars = None):
                if extra_vars is None:
                        vars = self.vars
                else:
                        vars = self.vars.copy()
                        vars.update(extra_vars)

                p = '%s/%s.html' % (self.tpath, page_name)
                if os.path.isfile(p):
                        return open(p).read() % vars
                return default_template % vars

        def get_main_header(self):
                return self.get_template('header', default_main_header)

        def get_main_footer(self):
                return self.get_template('footer', default_main_footer)

        def get_article_header(self, article):
                return self.get_template(
                        'art_header', default_article_header, article.to_vars())

        def get_article_footer(self, article):
                return self.get_template(
                        'art_footer', default_article_footer, article.to_vars())

        def get_comment_header(self, comment):
                vars = comment.to_vars()
                if comment.link:
                        vars['linked_author'] = '<a href="%s">%s</a>' \
                                        % (vars['link'], vars['author'])
                else:
                        vars['linked_author'] = vars['author']
                return self.get_template(
                        'com_header', default_comment_header, vars)

        def get_comment_footer(self, comment):
                return self.get_template(
                        'com_footer', default_comment_footer, comment.to_vars())

        def get_comment_form(self, article, form_data, captcha_puzzle):
                vars = article.to_vars()
                vars.update(form_data.to_vars(self))
                vars['captcha_puzzle'] = captcha_puzzle
                return self.get_template(
                        'com_form', default_comment_form, vars)

        def get_comment_error(self, error):
                return self.get_template(
                        'com_error', default_comment_error, dict(error=error))


class CommentFormData (object):
        def __init__(self, author = '', link = '', captcha = '', body = ''):
                self.author = author
                self.link = link
                self.captcha = captcha
                self.body = body
                self.author_error = ''
                self.link_error = ''
                self.captcha_error = ''
                self.body_error = ''
                self.action = ''
                self.method = 'post'

        def to_vars(self, template):
                render_error = template.get_comment_error
                a_error = self.author_error and render_error(self.author_error)
                l_error = self.link_error and render_error(self.link_error)
                c_error = self.captcha_error \
                                and render_error(self.captcha_error)
                b_error = self.body_error and render_error(self.body_error)
                return {
                        'form_author': sanitize(self.author),
                        'form_link': sanitize(self.link),
                        'form_captcha': sanitize(self.captcha),
                        'form_body': sanitize(self.body),

                        'form_author_error': a_error,
                        'form_link_error': l_error,
                        'form_captcha_error': c_error,
                        'form_body_error': b_error,

                        'form_action': self.action,
                        'form_method': self.method,
                }


class Comment (object):
        def __init__(self, article, number, created = None):
                self.article = article
                self.number = number
                if created is None:
                        self.created = datetime.datetime.now()
                else:
                        self.created = created

                self.loaded = False

                # loaded on demand
                self._author = author
                self._link = ''
                self._raw_content = 'Removed comment'

        @property
        def author(self):
                if not self.loaded:
                        self.load()
                return self._author

        @property
        def link(self):
                if not self.loaded:
                        self.load()
                return self._link

        @property
        def raw_content(self):
                if not self.loaded:
                        self.load()
                return self._raw_content

        def set(self, author, raw_content, link = '', created = None):
                self.loaded = True
                self._author = author
                self._raw_content = raw_content
                self._link = link
                self.created = created or datetime.datetime.now()


        def load(self):
                filename = os.path.join(comments_path, self.article.uuid,
                                        str(self.number))
                try:
                        raw = open(filename).readlines()
                except:
                        return

                count = 0
                for l in raw:
                        if ':' in l:
                                name, value = l.split(':', 1)
                                if name.lower() == 'author':
                                        self._author = value.strip()
                                elif name.lower() == 'link':
                                        self._link = value.strip()
                        elif l == '\n':
                                # end of header
                                break
                        count += 1
                self._raw_content = ''.join(raw[count + 1:])
                self.loaded = True

        def save(self):
                filename = os.path.join(comments_path, self.article.uuid,
                                        str(self.number))
                try:
                        f = open(filename, 'w')
                        f.write('Author: %s\n' % self.author)
                        f.write('Link: %s\n' % self.link)
                        f.write('\n')
                        f.write(self.raw_content)
                except:
                        return


        def to_html(self):
                return rst_to_html(self.raw_content)

        def to_vars(self):
                return {
                        'number': self.number,
                        'author': sanitize(self.author),
                        'link': sanitize(self.link),
                        'date': self.created.isoformat(' '),
                        'created': self.created.isoformat(' '),

                        'year': self.created.year,
                        'month': self.created.month,
                        'day': self.created.day,
                        'hour': self.created.hour,
                        'minute': self.created.minute,
                        'second': self.created.second,
                }

class CommentDB (object):
        def __init__(self, article):
                self.path = os.path.join(comments_path, article.uuid)
                # if comments were enabled after the article was added, we
                # will need to create the directory
                if not os.path.exists(self.path):
                        os.mkdir(self.path, 0777)

                self.comments = []
                self.load(article)

        def load(self, article):
                try:
                        f = open(os.path.join(self.path, 'db'))
                except:
                        return

                for l in f:
                        # Each line has the following comma separated format:
                        # number, created (epoch)
                        # Empty lines are meaningful and represent removed
                        # comments (so we can preserve the comment number)
                        l = l.split(',')
                        try:
                                n = int(l[0])
                                d = datetime.datetime.fromtimestamp(float(l[1]))
                        except:
                                # Removed/invalid comment
                                self.comments.append(None)
                                continue
                        self.comments.append(Comment(article, n, d))

        def save(self):
                old_db = os.path.join(self.path, 'db')
                new_db = os.path.join(self.path, 'db.tmp')
                f = open(new_db, 'w')
                for c in self.comments:
                        s = ''
                        if c is not None:
                                s = ''
                                s += str(c.number) + ', '
                                s += str(time.mktime(c.created.timetuple()))
                        s += '\n'
                        f.write(s)
                f.close()
                os.rename(new_db, old_db)


class Article (object):
        def __init__(self, path, created = None, updated = None):
                self.path = path
                self.created = created
                self.updated = updated
                self.uuid = "%08x" % zlib.crc32(self.path)

                self.loaded = False

                # loaded on demand
                self._title = 'Removed post'
                self._author = author
                self._tags = []
                self._raw_content = ''
                self._comments = []

        @property
        def title(self):
                if not self.loaded:
                        self.load()
                return self._title

        @property
        def author(self):
                if not self.loaded:
                        self.load()
                return self._author

        @property
        def tags(self):
                if not self.loaded:
                        self.load()
                return self._tags

        @property
        def raw_content(self):
                if not self.loaded:
                        self.load()
                return self._raw_content

        @property
        def comments(self):
                if not self.loaded:
                        self.load()
                return self._comments


        def __eq__(self, other):
                if self.path == other.path:
                        return True
                return False


        def add_comment(self, author, raw_content, link = ''):
                c = Comment(self, len(self.comments))
                c.set(author, raw_content, link)
                self.comments.append(c)
                return c


        def load(self):
                # XXX this tweak is only needed for old DB format, where
                # article's paths started with a slash
                path = self.path
                if path.startswith('/'):
                        path = path[1:]
                filename = os.path.join(data_path, path)
                try:
                        raw = open(filename).readlines()
                except:
                        return

                count = 0
                for l in raw:
                        if ':' in l:
                                name, value = l.split(':', 1)
                                if name.lower() == 'title':
                                        self._title = value.strip()
                                elif name.lower() == 'author':
                                        self._author = value.strip()
                                elif name.lower() == 'tags':
                                        ts = value.split(',')
                                        ts = [t.strip() for t in ts]
                                        self._tags = set(ts)
                        elif l == '\n':
                                # end of header
                                break
                        count += 1
                self._raw_content = ''.join(raw[count + 1:])
                db = CommentDB(self)
                self._comments = db.comments
                self.loaded = True

        def to_html(self):
                dirname = os.path.dirname
                post_url = '/'.join([dirname(full_url), 'posts',
                                dirname(self.path)])
                post_dir = '/'.join([data_path, dirname(self.path)])
                rst = self.raw_content.replace('##POST_URL##', post_url)
                rst = rst.replace('##POST_DIR##', post_dir)
                # TODO: make it better!
                import re
                rst = re.sub(r'.. youtube:: (.*)', r'''.. raw:: html

                        <iframe width="560" height="315"
                                src="https://www.youtube.com/embed/\1"
                                frameborder="0" allowfullscreen></iframe>
                        ''', rst)
                rst = re.sub(r'.. vimeo:: (\w*)', r'''.. raw:: html

                        <iframe src="https://player.vimeo.com/video/\1"
                                width="500" height="281" frameborder="0"
                                webkitallowfullscreen mozallowfullscreen
                                allowfullscreen></iframe>
                        ''', rst)
                rst = re.sub(r'.. grooveshark:: (\w*)', r'''.. raw:: html

                        Grooveshark is no more! This was supposed to show \1.
                        ''', rst)
                return rst_to_html(rst)

        def to_vars(self):
                return {
                        'arttitle': sanitize(self.title),
                        'author': sanitize(self.author),
                        'date': self.created.isoformat(' '),
                        'uuid': self.uuid,
                        'tags': self.get_tags_links(),
                        'comments': len(self.comments),

                        'created': self.created.isoformat(' '),
                        'ciso': self.created.isoformat(),
                        'cyear': self.created.year,
                        'cmonth': self.created.month,
                        'cday': self.created.day,
                        'chour': self.created.hour,
                        'cminute': self.created.minute,
                        'csecond': self.created.second,

                        'updated': self.updated.isoformat(' '),
                        'uiso': self.updated.isoformat(),
                        'uyear': self.updated.year,
                        'umonth': self.updated.month,
                        'uday': self.updated.day,
                        'uhour': self.updated.hour,
                        'uminute': self.updated.minute,
                        'usecond': self.updated.second,
                }

        def get_tags_links(self):
                l = []
                tags = list(self.tags)
                tags.sort()
                for t in tags:
                        l.append('<a class="tag" href="%s/tag/%s">%s</a>' % \
                                (blog_url, urllib.quote(t), sanitize(t) ))
                return ', '.join(l)


class ArticleDB (object):
        def __init__(self, dbpath):
                self.dbpath = dbpath
                self.articles = []
                self.uuids = {}
                self.actyears = set()
                self.actmonths = set()
                self.acttags = set()
                self.load()

        def get_articles(self, year = 0, month = 0, day = 0, tags = None):
                l = []
                for a in self.articles:
                        if year and a.created.year != year: continue
                        if month and a.created.month != month: continue
                        if day and a.created.day != day: continue
                        if tags and not tags.issubset(a.tags): continue

                        l.append(a)

                return l

        def get_article(self, uuid):
                return self.uuids[uuid]

        def load(self):
                try:
                        f = open(self.dbpath)
                except:
                        return

                for l in f:
                        # Each line has the following comma separated format:
                        # path (relative to data_path), \
                        #       created (epoch), \
                        #       updated (epoch)
                        try:
                                l = l.split(',')
                        except:
                                continue

                        a = Article(l[0],
                                datetime.datetime.fromtimestamp(float(l[1])),
                                datetime.datetime.fromtimestamp(float(l[2])))
                        self.uuids[a.uuid] = a
                        self.acttags.update(a.tags)
                        self.actyears.add(a.created.year)
                        self.actmonths.add((a.created.year, a.created.month))
                        self.articles.append(a)

        def save(self):
                f = open(self.dbpath + '.tmp', 'w')
                for a in self.articles:
                        s = ''
                        s += a.path + ', '
                        s += str(time.mktime(a.created.timetuple())) + ', '
                        s += str(time.mktime(a.updated.timetuple())) + '\n'
                        f.write(s)
                f.close()
                os.rename(self.dbpath + '.tmp', self.dbpath)

        def get_year_links(self):
                yl = list(self.actyears)
                yl.sort(reverse = True)
                return [ '<a href="%s/%d/">%d</a>' % (blog_url, y, y)
                                for y in yl ]

        def get_month_links(self, year):
                am = [ i[1] for i in self.actmonths if i[0] == year ]
                ml = []
                for i in range(1, 13):
                        name = calendar.month_name[i][:3]
                        if i in am:
                                s = '<a href="%s/%d/%d/">%s</a>' % \
                                        ( blog_url, year, i, name )
                        else:
                                s = name
                        ml.append(s)
                return ml

        def get_tag_links(self):
                tl = list(self.acttags)
                tl.sort()
                return [ '<a href="%s/tag/%s">%s</a>' % (blog_url,
                                sanitize(t), sanitize(t)) for t in tl ]

#
# Main
#

def render_comments(article, template, form_data):
        print '<a name="comments" />'
        for c in article.comments:
                if c is None:
                        continue
                print template.get_comment_header(c)
                print c.to_html()
                print template.get_comment_footer(c)
        if not form_data:
                form_data = CommentFormData()
        form_data.action = blog_url + '/comment/' + article.uuid + '#comment'
        captcha = captcha_method(article)
        print template.get_comment_form(article, form_data, captcha.puzzle)

def render_html(articles, db, actyear = None, show_comments = False,
                redirect =  None, form_data = None):
        if redirect:
                print 'Status: 303 See Other\r\n',
                print 'Location: %s\r\n' % redirect,
        print 'Content-type: text/html; charset=utf-8\r\n',
        print '\r\n',
        template = Templates(templates_path, db, actyear)
        print template.get_main_header()
        for a in articles:
                print template.get_article_header(a)
                print a.to_html()
                print template.get_article_footer(a)
                if show_comments:
                        render_comments(a, template, form_data)
        print template.get_main_footer()

def render_artlist(articles, db, actyear = None):
        template = Templates(templates_path, db, actyear)
        print 'Content-type: text/html; charset=utf-8\n'
        print template.get_main_header()
        print '<h2>Articles</h2>'
        for a in articles:
                print '<li><a href="%(url)s/post/%(uuid)s">%(title)s</a></li>' \
                        % {     'url': blog_url,
                                'uuid': a.uuid,
                                'title': a.title,
                                'author': a.author,
                        }
        print template.get_main_footer()

def render_atom(articles):
        if len(articles) > 0:
                updated = articles[0].updated.isoformat()
        else:
                updated = datetime.datetime.now().isoformat()

        print 'Content-type: application/atom+xml; charset=utf-8\n'
        print """<?xml version="1.0" encoding="utf-8"?>

<feed xmlns="http://www.w3.org/2005/Atom">
 <title>%(title)s</title>
 <link rel="alternate" type="text/html" href="%(url)s"/>
 <link rel="self" type="application/atom+xml" href="%(url)s/atom"/>
 <id>%(url)s</id> <!-- TODO: find a better <id>, see RFC 4151 -->
 <updated>%(updated)sZ</updated>

        """ % {
                'title': title,
                'url': full_url,
                'updated': updated,
        }

        for a in articles:
                vars = a.to_vars()
                vars.update( {
                        'url': full_url,
                        'contents': a.to_html(),
                } )
                print """
  <entry>
    <title>%(arttitle)s</title>
    <author><name>%(author)s</name></author>
    <link href="%(url)s/post/%(uuid)s" />
    <id>%(url)s/post/%(uuid)s</id>
    <summary>%(arttitle)s</summary>
    <published>%(ciso)sZ</published>
    <updated>%(uiso)sZ</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml">
%(contents)s
      </div>
    </content>
  </entry>
                """ % vars
        print "</feed>"


def render_style():
        print 'Content-type: text/css\r\n\r\n',
        print default_css

# Get a dictionary with sort() arguments (key and reverse) by parsing the sort
# specification format:
# [+-]?<key>?
# Where "-" is used to specify reverse order, while "+" is regular, ascending,
# order (reverse = False). The key value is an Article's attribute name (title,
# author, created, updated and uuid are accepted), and will be used as key for
# sorting. If a value is omitted, that value is taken from the default, which
# should be provided using the same format specification, with the difference
# that all values must be provided for the default.
def get_sort_args(sort_str, default):
        def parse(s):
                d = dict()
                if not s:
                        return d
                key = None
                if len(s) > 0:
                        # accept ' ' as an alias of '+' since '+' is translated
                        # to ' ' in URLs
                        if s[0] in ('+', ' ', '-'):
                                key = s[1:]
                                d['reverse'] = (s[0] == '-')
                        else:
                                key = s
                if key in ('title', 'author', 'created', 'updated', 'uuid'):
                        d['key'] = lambda a: getattr(a, key)
                return d
        args = parse(default)
        assert args['key'] is not None and args['reverse'] is not None
        args.update(parse(sort_str))
        return args

def handle_cgi():
        import cgitb; cgitb.enable()

        form = cgi.FieldStorage()
        year = int(form.getfirst("year", 0))
        month = int(form.getfirst("month", 0))
        day = int(form.getfirst("day", 0))
        tags = set(form.getlist("tag"))
        sort_str = form.getfirst("sort", None)
        uuid = None
        atom = False
        style = False
        post = False
        post_preview = False
        artlist = False
        comment = False

        if os.environ.has_key('PATH_INFO'):
                path_info = os.environ['PATH_INFO']
                style = path_info == '/style'
                atom = path_info == '/atom'
                tag = path_info.startswith('/tag/')
                post = path_info.startswith('/post/')
                post_preview = path_info.startswith('/preview/post/')
                artlist = path_info.startswith('/list')
                comment = path_info.startswith('/comment/') and enable_comments
                if not style and not atom and not post and not post_preview \
                                and not tag and not comment and not artlist:
                        date = path_info.split('/')[1:]
                        try:
                                if len(date) > 1 and date[0]:
                                        year = int(date[0])
                                if len(date) > 2 and date[1]:
                                        month = int(date[1])
                                if len(date) > 3 and date[2]:
                                        day = int(date[2])
                        except ValueError:
                                pass
                elif post:
                        uuid = path_info.replace('/post/', '')
                        uuid = uuid.replace('/', '')
                elif post_preview:
                        art_path = path_info.replace('/preview/post/', '')
                        art_path = urllib.unquote_plus(art_path)
                        art_path = os.path.join(data_path, art_path)
                        art_path = os.path.realpath(art_path)
                        common = os.path.commonprefix([data_path, art_path])
                        if common != data_path: # something nasty happened
                                post_preview = False
                        art_path = art_path[len(data_path)+1:]
                elif tag:
                        t = path_info.replace('/tag/', '')
                        t = t.replace('/', '')
                        t = urllib.unquote_plus(t)
                        tags = set((t,))
                elif comment:
                        uuid = path_info.replace('/comment/', '')
                        uuid = uuid.replace('#comment', '')
                        uuid = uuid.replace('/', '')
                        author = form.getfirst('comformauthor', '')
                        link = form.getfirst('comformlink', '')
                        captcha = form.getfirst('comformcaptcha', '')
                        body = form.getfirst('comformbody', '')

        db = ArticleDB(os.path.join(data_path, 'db'))
        if atom:
                articles = db.get_articles(tags = tags)
                articles.sort(**get_sort_args(sort_str, '-created'))
                render_atom(articles[:index_articles])
        elif style:
                render_style()
        elif post:
                render_html( [db.get_article(uuid)], db, year, enable_comments )
        elif post_preview:
                article = Article(art_path, datetime.datetime.now(),
                                        datetime.datetime.now())
                render_html( [article], db, year, enable_comments )
        elif artlist:
                articles = db.get_articles()
                articles.sort(**get_sort_args(sort_str, '+title'))
                render_artlist(articles, db)
        elif comment and enable_comments:
                form_data = CommentFormData(author.strip().replace('\n', ' '),
                                link.strip().replace('\n', ' '), captcha,
                                body.replace('\r', ''))
                article = db.get_article(uuid)
                captcha = captcha_method(article)
                redirect = False
                valid = True
                if not form_data.author:
                        form_data.author_error = 'please, enter your name'
                        valid = False
                if form_data.link:
                        link = valid_link(form_data.link)
                        if link:
                                form_data.link = link
                        else:
                                form_data.link_error = 'please, enter a ' \
                                                'valid link'
                                valid = False
                if not captcha.validate(form_data):
                        form_data.captcha_error = captcha.help
                        valid = False
                if not form_data.body:
                        form_data.body_error = 'please, write a comment'
                        valid = False
                else:
                        error = validate_rst(form_data.body, secure=False)
                        if error is not None:
                                (line, desc, ctx) = error
                                at = ''
                                if line:
                                        at = ' at line %d' % line
                                form_data.body_error = 'error%s: %s' \
                                                % (at, desc)
                                valid = False
                if valid:
                        c = article.add_comment(form_data.author,
                                        form_data.body, form_data.link)
                        c.save()
                        cdb = CommentDB(article)
                        cdb.comments = article.comments
                        cdb.save()
                        redirect = blog_url + '/post/' + uuid + '#comment-' \
                                        + str(c.number)
                render_html( [article], db, year, enable_comments, redirect,
                                form_data )
        else:
                articles = db.get_articles(year, month, day, tags)
                articles.sort(**get_sort_args(sort_str, '-created'))
                if not year and not month and not day and not tags:
                        articles = articles[:index_articles]
                render_html(articles, db, year)


def usage():
        print 'Usage: %s {add|rm|update} article_path' % sys.argv[0]

def handle_cmd():
        if len(sys.argv) != 3:
                usage()
                return 1

        cmd = sys.argv[1]
        art_path = os.path.realpath(sys.argv[2])

        if os.path.commonprefix([data_path, art_path]) != data_path:
                print "Error: article (%s) must be inside data_path (%s)" % \
                                (art_path, data_path)
                return 1
        art_path = art_path[len(data_path)+1:]

        db_filename = os.path.join(data_path, 'db')
        if not os.path.isfile(db_filename):
                open(db_filename, 'w').write('')
        db = ArticleDB(db_filename)

        if cmd == 'add':
                article = Article(art_path, datetime.datetime.now(),
                                        datetime.datetime.now())
                for a in db.articles:
                        if a == article:
                                print 'Error: article already exists'
                                return 1
                db.articles.append(article)
                db.save()
                if enable_comments:
                        comment_dir = os.path.join(comments_path, article.uuid)
                        try:
                                os.mkdir(comment_dir, 0775)
                        except OSError, e:
                                if e.errno != errno.EEXIST:
                                        print "Error: can't create comments " \
                                                "directory %s (%s)" \
                                                        % (comment_dir, e)
                                # otherwise is probably a removed and re-added
                                # article
        elif cmd == 'rm':
                article = Article(art_path)
                for a in db.articles:
                        if a == article:
                                break
                else:
                        print "Error: no such article"
                        return 1
                if enable_comments:
                        r = raw_input('Remove comments [y/N]? ')
                db.articles.remove(a)
                db.save()
                if enable_comments and r.lower() == 'y':
                        shutil.rmtree(os.path.join(comments_path, a.uuid))
        elif cmd == 'update':
                article = Article(art_path)
                for a in db.articles:
                        if a == article:
                                break
                else:
                        print "Error: no such article"
                        return 1
                a.updated = datetime.datetime.now()
                db.save()
        else:
                usage()
                return 1

        return 0


if os.environ.has_key('GATEWAY_INTERFACE'):
        i = datetime.datetime.now()
        handle_cgi()
        f = datetime.datetime.now()
        print '<!-- render time: %s -->' % (f-i)
else:
        sys.exit(handle_cmd())