ljmkstatic

#!/usr/bin/python
# -*- encoding: utf-8 -*-
"""
Formats two files, downloaded by ljdump.py into HTML page using
template.

Syntax ljformatxml -t template -o file.html L-nnnn C-nnnn

"""
from ConfigParser import ConfigParser
import xml.dom.minidom, xml.dom
import re
import sys,os,glob
import codecs
# Параметры конфигурации
# Директория для симлинков
# url картинки с человечком
# Директория с результатами дампа
# директория для картинки
# шаблон для страницы поста
# шаблон для блока комментария
template={}
urls={}
dirs={}
def read_templates(config):
    global template
    for i in ['post','comment']:
        with codecs.open(config.get('templates',i),'r','utf-8') as f:
            template[i]=f.read()


def set_parameters(config):
    global urls,dirs
    for i in config.options('urls'):
        urls[i]=config.get('urls',i)
    for i in config.options('directories'):
        dirs[i]=config.get('directories',i)

def process_ljtag(m):
    """
    Receives lj tag match object with lj tag and returns
    html text which should be used as replacement
    Used to pass into re.sub
    """
    tag = m.group(0)
    if tag.find("lj-cut")!=-1:
        return ""
    if tag.find('user=')!=-1:
        name= re.search('user=[\'\"]?(\w+)[\'\"]?',tag).group(1)
        title = re.search('title=[\"\']?([^"\'>]+)[\'\"]?',tag)
        if title:
            title = title.group(1)
        else:
            title=name
        tag= '<a style="color: blue; font-weight: bold;" href="http://www.livejournal.com/users/%s/profile"><img src="%s">%s</a>'%(name,urls['icons']+"/userinfo.gif",title)
        return tag
    print "unknown lj tag: ",tag

def process_text(text):
      # Выделить оттуда текст, распарсить как html, заменяя lj-тэги
      try:
        text = re.sub("</?lj[^>]+>",process_ljtag,text)
      except Exception as e:
        print 'bad text :',text
        raise e
      text = re.sub("\r?\n","<br>",text)
      # и заменяя img на локальные копии, если они есть. Если нет, писать в
      # кеш картинок
      # FIXME post_props[post_text] = re.sub("<img # [^>]+>",process_img.post_text)
      return text
def format_comments(cmt_list):
    out=[]
    for cmt in cmt_list:
        if len(cmt['children']):
            cmt['comments']=format_comments(cmt['children'])
        else:
            cmt['comments']=''
        if 'user' in cmt:
            cmt['userlink']=process_text('<lj user="%s">'%cmt['user'])
        out.append(template['comment'] % cmt)
    return ''.join(out)

def do_post(postfile,commentfile,outputfile):
    """
    Handles one post. Returns post date, url, subject and tag list
    """
# Прочитать L-nnnn
    post_xml = xml.dom.minidom.parse(postfile)
    post_props = {'subject':'','taglist':''}
    for n in post_xml.documentElement.childNodes:
        if n.nodeType == xml.dom.Node.ELEMENT_NODE:
            if n.nodeName == u'event':
                post_props['text']=process_text(n.firstChild.nodeValue)
            elif n.nodeName == u'props':
            # Выделить необходимую метаинформацию
                for n2 in n.childNodes:
                    if n2.nodeType == xml.dom.Node.ELEMENT_NODE:
                        post_props[str(n2.nodeName)] = n2.firstChild.nodeValue
            else:
                post_props[str(n.nodeName)] = n.firstChild.nodeValue

    if not 'text' in post_props:
        raise ValueError("No event node in ths post")
    if 'picture_keyword' in post_props:
        userpic=post_props['picture_keyword']
    else:
        userpic='_'
    for fmt in ('jpg','gif','png'):
        if os.access("%s/%s.%s" % (dirs['archive'],userpic,fmt),os.R_OK):
            post_props['userpic']='%s/userpics/%s.%s'%(urls['images'],userpic,fmt)
            break
    if commentfile:
        comment_xml = xml.dom.minidom.parse(  commentfile)
    # We suppose that comments are already sorted accoridng to post time    
        comment_tree = []
        comment_hash = {}
        comment_count = 0
        for c in comment_xml.documentElement.childNodes:
            if c.nodeType != xml.dom.Node.ELEMENT_NODE or c.nodeName != 'comment':
                continue
            comment={'date':'Unknown','children':[],'subject':'','userlink':'(Anonymous)'}
            for i in c.childNodes:
                if i.nodeType != xml.dom.Node.ELEMENT_NODE:
                    continue
                if i.nodeName == 'body':
                    if i.firstChild is None:
                        comment['body']='<b>Deleted comment</b>' 
                    else:  
                        comment['body']=process_text(i.firstChild.nodeValue)
                else:
                    tx=i.firstChild
                    if tx:
                        comment[str(i.nodeName)]=tx.nodeValue
            comment_hash[comment['id']]=comment
            if 'parentid' in comment and comment['parentid'] in comment_hash:
                comment_hash[comment['parentid']]['children'].append(comment)
                comment_count +=1
            else:
                comment_tree.append(comment)
                
        post_props['comments'] = format_comments(comment_tree)                
        post_props['comment_count'] = comment_count
    else:
        post_props['comments'] = ''
        post_props['comment_count'] = 0
    page = template['post']%post_props

    with codecs.open(outputfile,"w","utf-8") as f :
        f.write(page)
    return (post_props['logtime'],post_props['ditemid'],post_props['subject'],post_props['taglist'])




if __name__ == '__main__':
    config=ConfigParser()
    if config.read(["ljmkstatic.conf"]) < 1:
        raise ValueError("No config file found")
    read_templates(config)
    set_parameters(config)
    for post_file in sorted(glob.glob(dirs['dump']+"/L-*")):
        post_id = re.search("(\d+)$",post_file).group(1)
        comment_file = dirs['dump']+"/C-"+post_id
        outfile=dirs['dump']+"/"+post_id+".html"
        try:
            t1=os.stat(post_file).st_mtime
            try:
                t2=os.stat(comment_file).st_mtime
            except OSError:
                t2=0
                comment_file = None
            t3=os.stat(outfile).st_mtime
            if t3 > t1 and t3 > t2:
                continue
        except OSError:
            pass
        print "Processing post L-%s"%post_id
        (date,post_id,subject,tags) = do_post(post_file,comment_file,outfile)
        # Fix me - update index structures