#!/usr/bin/python
# -*- coding: utf-8 -*-
#   ~/bin/cron/cron-bookmarks
#   translate the Firefox (once Netscape) bookmarks database (once HTML list)
#      into something public, hiding items whose description begins with
#      the word HIDDEN

# Notes:
#  Python UTF-8 handling is really, really annoying.
#  And don't even get me started about not being able to put it in doc strings.

# Author C. Alex. North-Keys
# Incept Fri Jan 24 18:06:36 CST 1997
# Update 2002-05-11 02:54:45 CDT (May Sat) += mozilla, more perl impl.
# Update 2006-08-03 01:03:16 CDT (Aug Thu) bookmarks via link in .mozilla
# Update 2009-03-21 02:44:35 CDT (Mar Sat) translation to Python

from __future__ import with_statement
import os
import sys
import shutil
import subprocess
import sqlite3
import lxml
import lxml.builder

hide_hiddens = True

query = ('SELECT '
         'id, '
         'main.type AS kind, '
         'main.position AS position, '
         '(SELECT title FROM moz_bookmarks AS other '
         ' WHERE other.id = main.parent) AS parent, '
         'main.parent AS parentid, '
         '(SELECT url FROM moz_places AS places '
         ' WHERE places.id = main.fk) AS uri, '
         'main.title AS name, '
         '(SELECT content FROM moz_items_annos AS items ' # 3 = text, 2 = time?
         ' WHERE items.item_id = main.id AND items.type = 3) AS note '
         'FROM moz_bookmarks AS main;')

# INPUT
home = os.environ['HOME']
bookmarks_sqlite = os.environ['BOOKMARKS'] if os.environ.has_key('BOOKMARKS') \
    else os.path.join(home, '.mozilla', 'places.sqlite')
bookmarks_sqlite_tmp = '/tmp/' + str(os.getpid())
# OUTPUT - full
wwwpriv = 'www/@auth/private'
bookmarks_priv_xml_out = os.path.join(home, wwwpriv, 'bookmarks.xml')
bookmarks_priv_html_out = os.path.join(home, wwwpriv, 'bookmarks-core.shtml')
# OUTPUT - no HIDDENS
bookmarks_pub_xml_out = os.path.join(home, 'www', 'bookmarks.xml')
bookmarks_pub_html_out = os.path.join(home, 'www', 'bookmarks-core.shtml')

shutil.copy(bookmarks_sqlite, bookmarks_sqlite_tmp)

x = lxml.builder.ElementMaker()

xt_top = x('folders', x('title', 'root'))
xt_top.attrib['title'] = 'root'

db  = sqlite3.connect(bookmarks_sqlite_tmp)
cur = db.cursor()
cur.execute(query)

# =0============================================ parsing ======

nodes = {}

#                        none?             none?   none?  none?
for (id, kind, position, folder, folderid, url,    title, note) in cur:
    #(1,    2,        0, None,          0, None,   u'',               None)
    #(11,   2,        1, u'Book...',    3, None,   u'Acts',           None)
    #(55,   1,        5, u'\u30...',   52, u'...', u'About the Logo', None)

    if folder is None: folder = ''
    if url    is None: url    = ''
    if title  is None: title  = ''
    if note   is None: note   = ''

    nodetype = None
    if   1 == kind:  nodetype = 'bookmark'
    elif 2 == kind:  nodetype = 'folder'
    elif 3 == kind:  continue   # nodetype = 'mystery'? just skipping
    else: print 'KIND error: %r' % (kind, )

    xt_node = x(nodetype, x('title', title))

    if 'bookmark' == nodetype:
        xt_url = x('url', url)
        xt_node.append(xt_url)
    if (0 < len(note)):
        xt_note = x('note', note)
        xt_node.append(xt_note)
    xt_node.attrib['title'] = title

    nodes[id] = (xt_node, nodetype, id, folderid, position)

# =1=========================================== foldering =====

bookmarks_by_folder_id = {}   # bookmarks_by_folder_id[id] = [] of bookmarks
folders = {}                  # folders[id] = a folder

folders[0] = xt_top

for (id, (xt_node, nodetype, id, parent_id, position)) in nodes.iteritems():
    if not bookmarks_by_folder_id.has_key(parent_id):
        bookmarks_by_folder_id[parent_id] = list()
    bookmarks_by_folder_id[parent_id].append((xt_node, position))

    if 'folder' == nodetype:
        folders[id] = xt_node

# =2============================================ ordering =====

def sort_on_int2nd_of_tuple(a, b):
    (a_ign, a_sort) = a
    (b_ign, b_sort) = b
    return cmp(int(a_sort), int(b_sort))

for id in bookmarks_by_folder_id.iterkeys():
    bookmarks_by_folder_id[id].sort(sort_on_int2nd_of_tuple)

for (id, stuff) in bookmarks_by_folder_id.iteritems():
    for (xt_node, _) in stuff:
        folders[id].append(xt_node)

xt_top.append(folders[1])

# =3=========================================== stripping =====

def xml2html_note_check(xml_node_check, xml_node_target):
    xt_note = xml_node_check.find('note')
    if xt_note is not None:
        note = xt_note.text
        if 0 < len(note):
            xt_html_note = x('div', note)
            xt_html_note.attrib['class'] = 'note'
            xml_node_target.append(xt_html_note)

def xml2html_folder(xml_folder):
    '''Returns the html_folder node derived from the XML_FOLDER.'''
    title   = xml_folder.find('title').text
    html_xt_folder = x('div', x('h3', title))
    xml2html_note_check(xml_folder, html_xt_folder)
    html_xt_list = x('ul')
    for xml_node in xml_folder:
        html_xt_node = None
        if 'folder' == xml_node.tag:
            html_xt_node = x('li', xml2html_folder(xml_node))
        elif 'bookmark' == xml_node.tag:
            url   = xml_node.find('url').text
            title = xml_node.find('title').text
            xt_note  = xml_node.find('note')
            html_xt_a = x('a', title)
            html_xt_a.attrib['href'] = url
            html_xt_node = x('li', html_xt_a)
            xml2html_note_check(xml_node, html_xt_node)
        elif 'title' == xml_node.tag:
            pass # already handled
        elif 'note' == xml_node.tag:
            pass # not using these
        else:
            sys.stderr.write('oops on type %s\n' % (xml_node.tag,))
        if html_xt_node is not None:
            html_xt_list.append(html_xt_node)
    if 0 < len(html_xt_list):
        html_xt_folder.append(html_xt_list)
    return html_xt_folder

xt_bookmarks_toolbar = xt_top.xpath( \
    './/folder/title[text()="Bookmarks Toolbar"]/..')
if xt_bookmarks_toolbar is not None:
    xt_top = xt_bookmarks_toolbar[0]

# private version with hidden stuff
xml_full = ('<?xml version="1.0" encoding="UTF-8" ?>\n'
       + lxml.etree.tostring(xt_top, pretty_print=True))
html_full = lxml.etree.tostring(xml2html_folder(xt_top), pretty_print=True)

# remove hidden stuff
if hide_hiddens:
    for xt_node in xt_top.xpath('.//note[starts-with(text(),"HIDDEN")]/..'):
        xt_parent = xt_node.getparent()
        if xt_parent is not None:
            xt_parent.remove(xt_node)

# public version
xml = ('<?xml version="1.0" encoding="UTF-8" ?>\n'
       + lxml.etree.tostring(xt_top, pretty_print=True))
html = lxml.etree.tostring(xml2html_folder(xt_top), pretty_print=True)

# create/update the bookmark files
with open(bookmarks_priv_xml_out, 'w') as output:
    output.write(xml_full) # autoconvert

with open(bookmarks_priv_html_out, 'w') as output:
    output.write(html_full)

with open(bookmarks_pub_xml_out, 'w') as output:
    output.write(xml) # autoconvert

with open(bookmarks_pub_html_out, 'w') as output:
    output.write(html)

#------------------------------------------------------------eof