# builder.py                                    -*- coding: ISO-8859-1 -*-
#
#   Copyright © 2000, 2002, 2004 Carey Evans.
#
#   Permission to use, copy, modify, and distribute this software and
#   its documentation for any purpose and without fee is hereby granted,
#   provided that the above copyright notice appear in all copies and
#   that both that copyright notice and this permission notice appear in
#   supporting documentation.
#
#   $Id: builder.py 26 2004-01-20 23:40:48Z carey $

"""builder.py

For a given input file, convert it to HTML and write it out.
If the document contains stylesheet processing instructions, use
that stylesheet first before converting CWML->HTML.
"""

import os, re, string, datetime
import urlparse

import libxml2, libxslt

libxml2.lineNumbersDefault(1)
libxml2.substituteEntitiesDefault(1)

class PageBuilderError(Exception):
    pass

class PageBuilder:
    def __init__(self):
        # Load the CWML to HTML stylesheet.
        xslpath = os.path.join(os.path.dirname(__file__), 'cwml.xsl')
        self.cwml_style = libxslt.parseStylesheetFile(xslpath)
        if self.cwml_style is None:
            raise PageBuilderError('Error loading default stylesheet.')

    def __del__(self):
        self.cwml_style.freeStylesheet()

    def build(self, page, src, dst):
        """Convert one page to HTML."""
        # Load text or XML document from disk.
        if page.is_text():
            doc = self.load_text(page, src)
        else:
            doc = libxml2.parseFile(src)
        if doc is None:
            raise PageBuilderError('Error loading ' + src)
        # Work out stylesheet parameters.
        mtime = self.get_doc_mtime(doc)
        if mtime is None:
            mtime = self.get_file_mtime(src)
        params = { 'root': "'%s'" % page.rootpath(),
                   'mtime': self.strparam(mtime),
                   'navtitle': self.strparam(page.title),
                   'navbar': self.strparam(self.navbar(page)) }
        # Process with stylesheet from PI, while present.
        pi_style = libxslt.loadStylesheetPI(doc)
        while pi_style is not None:
            newdoc = pi_style.applyStylesheet(doc, params)
            doc.freeDoc()
            pi_style.freeStylesheet()
            if newdoc is None:
                raise PageBuilderError(
                    'Error applying internal stylesheet to ' + src)
            doc = newdoc
            pi_style = libxslt.loadStylesheetPI(doc)
        # Process with CWML to HTML stylesheet.
        res = self.cwml_style.applyStylesheet(doc, params)
        doc.freeDoc()
        if res is None:
            raise PageBuilderError(
                'Error applying default stylesheet to ' + src)
        self.cwml_style.saveResultToFilename(dst, res, False)
        res.freeDoc()

    def load_text(self, page, src):
        """Load a text document and wrap in XML."""
        f = open(src, 'r')
        data = f.read()
        f.close()
        doc = libxml2.newDoc(None)
        top = doc.newChild(None, 'webpage', None)
        top.newTextChild(None, 'title', page.title)
        top.newTextChild(None, 'pre', data)
        return doc

    def strparam(self, s):
        return "'" + s.encode('utf-8').replace("'", "'") + "'"

    def navbar(self, page):
        return '||'.join(['|'.join(t) for t in page.navbar()])

    def get_doc_mtime(self, doc):
        keyword = doc.xpathEval('string(/webpage/modified)')
        tok = keyword.split(' ')
        if len(tok) < 3:
            return None
        return tok[1] + 'T' + tok[2]

    def get_file_mtime(self, src):
        ts = datetime.datetime.fromtimestamp(os.path.getmtime(src))
        return ts.isoformat()