# Copyright (c) 2002-2005 Infrae. All rights reserved.
# See also LICENSE.txt
# $Revision: 1.35 $
# Zope

from StringIO import StringIO
import re
import sys
import traceback
import urlparse

from AccessControl import ClassSecurityInfo
from Products.PageTemplates.PageTemplateFile import PageTemplateFile
from Globals import InitializeClass
from Persistence import Persistent
from zExceptions import InternalError

from webdav.common import Conflict

from Products.ParsedXML.ExtraDOM import writeStream
from Products.ParsedXML.ParsedXML import createDOMDocument
from Products.ParsedXML.PrettyPrinter import _translateCdata
from Products.ParsedXML.ParsedXML import ParsedXML

# Silva
from Products.Silva import SilvaPermissions
from Products.Silva.VersionedContent import CatalogedVersionedContent
from Products.Silva.helpers import add_and_edit, translateCdata
from Products.Silva.Version import CatalogedVersion
from Products.Silva import mangle

from Products.Silva.ImporterRegistry import importer_registry, xml_import_helper, get_xml_id, get_xml_title
from Products.Silva.Metadata import export_metadata
from Products.Silva.i18n import translate as _

from Products.Silva.ContentObjectFactoryRegistry import contentObjectFactoryRegistry

# For XML-Conversions for editors
from transform.Transformer import EditorTransformer
from transform.base import Context

from Products.Silva.interfaces import IVersionedContent, IContainerPolicy
from Products.Silva.interfaces import IVersion
from Products.Silva.Versioning import VersioningError
from Products.SilvaDocument.i18n import translate as _

from Products.SilvaDocument import externalsource

from xml.sax import make_parser
from xml.sax.handler import ContentHandler
from Products.SilvaMetadata.Exceptions import BindingError

icon="www/silvadoc.gif"
addable_priority = -1

class Document(CatalogedVersionedContent):
    __doc__ = _(
    """A Document is the basic unit of information in Silva. A document
        can – much like word processor documents – contain text,
        lists, tables, headers, subheads, images, etc. Documents can have
        two accessible versions, one online for the public, another in
        process (editable or approved/published). Older versions can be rolled 
        forward and made editable.
    """)
    security = ClassSecurityInfo()

    meta_type = "Silva Document"

    __implements__ = IVersionedContent

    # some scary DAV stuff...
    __dav_collection__ = False
    isAnObjectManager = False

    # A hackish way, to get a Silva tab in between the standard ZMI tabs
    inherited_manage_options = CatalogedVersionedContent.manage_options
    manage_options=(
        (inherited_manage_options[0],)+
        ({'label':'Silva /edit...', 'action':'edit'},)+
        inherited_manage_options[1:]
        )

    # ACCESSORS
    security.declareProtected(SilvaPermissions.View, 'is_cacheable')
    def is_cacheable(self):
        """Return true if this document is cacheable.
        That means the document contains no dynamic elements like
        code, datasource or toc.
        """
        non_cacheable_elements = ['toc', 'code',]

        viewable = self.get_viewable()
        if viewable is None:
            return 0

        # It should suffice to test the children of the root element only,
        # since currently the only non-cacheable elements are root elements
        for node in viewable.content.documentElement.childNodes:
            node_name = node.nodeName
            if node_name in non_cacheable_elements:
                return 0
            # FIXME: how can we make this more generic as it is very
            # specific now..?
            if node_name == 'source':
                is_cacheable = externalsource.isSourceCacheable(
                    self.aq_inner, node)
                if not is_cacheable:
                    return 0
        return 1

    security.declareProtected(SilvaPermissions.ReadSilvaContent,
                              'to_xml')
    def to_xml(self, context):
        """Render object to XML.
        """
        f = context.f

        if context.last_version == 1:
            version_id = self.get_next_version()
            if version_id is None:
                version_id = self.get_public_version()
        else:
            version_id = self.get_public_version()

        if version_id is None:
            return
        f.write('<silva_document id="%s">' % self.id)
        version = getattr(self, version_id)
        version.to_xml(context)
        f.write('</silva_document>')

    security.declareProtected(SilvaPermissions.ChangeSilvaContent,
                              'editor_storage')
    def editor_storage(self, string=None, editor='kupu', encoding=None):
        """provide xml/xhtml/html (GET requests) and (heuristic)
           back-transforming to xml/xhtml/html (POST requests)
        """
        transformer = EditorTransformer(editor=editor)

        # we need to know what browser we are dealing with in order to know
        # what html to produce, unfortunately Mozilla uses different tags in
        # some cases (b instead of strong, i instead of em)
        browser = 'Mozilla'
        if self.REQUEST['HTTP_USER_AGENT'].find('MSIE') > -1:
            browser = 'IE'

        if string is None:
            ctx = Context(f=StringIO(),
                            last_version=1,
                            url=self.absolute_url(),
                            browser=browser,
                            model=self)
            self.to_xml(ctx)
            htmlnode = transformer.to_target(sourceobj=ctx.f.getvalue(), context=ctx)
            if encoding is not None:
                ret = htmlnode.asBytes(encoding=encoding)
                ret = ret.replace('\xa0', '&nbsp;')
            else:
                ret = unicode(htmlnode.asBytes('utf8'),'utf8')
                ret = ret.replace(u'\xa0', u'&nbsp;')
            #print 'returning:', repr(ret)
            return ret
        else:
            #print 'incoming', repr(string)
            version = self.get_editable()
            if version is None:
                raise _("No version to store to!")

            # get any metadata elements saved to the metadata
            # XXX currently uses a completely different machinery to parse
            # the XML, perhaps we want to move that to the transformations
            # at some point (but would require too much hacking for now)
            errors = self._set_metadata_from_html(string, version)
            if errors:
                raise BindingError, errors

            ctx = Context(url=self.absolute_url(),
                            browser=browser,
                            model=self,
                            request=self.REQUEST)
            silvanode = transformer.to_source(targetobj=string, context=ctx)[0]
            title = silvanode.find('title')[0].extract_text()
            docnode = silvanode.find('doc')[0]
            content = docnode.asBytes(encoding="UTF8")
            version.content.manage_edit(content) # needs utf8-encoded string
            version.set_title(title) # needs unicode
            version.sec_update_last_author_info()

            # Clear widget cache for this version.
            version.clearEditorCache()
            #print 'storing:', repr(content)

    def _set_metadata_from_html(self, html, version):
        """Rip the metadata out of the HTML (meta tags), set it on version"""
        # XXX obviously it would make sense if the transformations could 
        # tackle this instead of doing it seperately, however, they are messy
        # enough as they are now (and not capable of such a thing yet)
        
        # XXX SAX took about 2 seconds (!) to get the meta values,
        # so I decided to try re instead... this made sense, parsing now takes
        # less than one hundreth of a second (usually even less than one 
        # thousandth!)
        metamapping = {}
        import re
        reg = re.compile(r'\<meta[^\>]+\/\>')
        while 1:
            match = reg.search(html)
            if not match:
                break
            tag = match.group(0)
            html = html.replace(tag, '')
            reg_props = re.compile(r'([a-zA-Z_]+)="([^"]+)"')
            found = {}
            while 1:
                match = reg_props.search(tag)
                if not match:
                    break
                tag = tag.replace(match.group(0), '')
                if match.group(1) in ['name', 'content', 'scheme']:
                    found[match.group(1)] = self._deentitize(match.group(2))
                if (found.has_key('name') and found.has_key('content') and 
                        found.has_key('scheme')):
                    if not metamapping.has_key(found['scheme']):
                        metamapping[found['scheme']] = {}
                    metamapping[found['scheme']][found['name']] = found['content']

        errors = {}
        binding = self.service_metadata.getMetadata(self.get_editable())
        for namespace, values in metamapping.items():
            # %!#$%# metadata system expects UTF-8 :(
            for key, value in values.items():
                del values[key]
                if value is None:
                    continue
                values[key.encode('UTF-8')] = value.encode('UTF-8')
            set = self.service_metadata.getMetadataSetFor(namespace)
            ret = binding.setValues(set.id, values, reindex=1)
            if ret:
                errors.update(ret)
        return errors
    
    def _deentitize(self, xml):
        return xml.replace('&lt;', '<').replace('&gt;', '>').\
                replace('&quot;', '"').replace('&amp;', '&')

    def revert_to_previous(self):
        """ Create a new version of public version, throw away the
        current one.

        Overrides Versioning.revert_to_previous() to be able to clear
        the widget cache too.
        """
        Document.inheritedAttribute('revert_to_previous')(self)
        version = self.get_editable()
        version.clearEditorCache()

    def manage_beforeDelete(self, item, container):
        Document.inheritedAttribute('manage_beforeDelete')(self, item, container)
        # Does the widget cache needs be cleared for all versions - I think so...
        for version in self._get_indexable_versions():
            version_object = getattr(self, str(version), None)
            if version_object:
                self.service_editor.clearCache(version_object.content)

    def transform_and_store(self, content_type, content):
        ret = content
        try:
            if content_type in ['text/html', 'application/xhtml+xml']:
                html = content
                self.editor_storage(html, 'kupu')

                # invalidate Document cache
                version = self.get_editable()
                version.clearEditorCache()

                # XXX This can be removed, right?
                ret = self.editor_storage(editor='kupu', encoding="UTF-8")
            else:
                # plain Silva document XML
                self.get_editable().content.manage_edit(content)
        except:
            # a tad ugly, but for debug purposes it would be nice to see
            # what's actually gone wrong
            exc, e, tb = sys.exc_info()
            # XXX should store to zLog instead
            print '%s: %s' % (exc, e)
            print '\n%s' % '\n'.join(traceback.format_tb(tb))
            print
            # reraise the exception so the enduser at least sees something's
            # gone wrong
            raise
        return ret

    # WebDAV API
    security.declareProtected(SilvaPermissions.ChangeSilvaContent,
                                'PUT')
    def PUT(self, REQUEST=None, RESPONSE=None):
        """PUT support"""
        # XXX we may want to make this more modular/pluggable at some point
        # to allow more content-types (+ transformations)
        if REQUEST is None:
            REQUEST = self.REQUEST
        if RESPONSE is None:
            RESPONSE = REQUEST.RESPONSE
        content = REQUEST['BODYFILE'].read()
        content_type = self._get_content_type_from_request(REQUEST, content)
        ret = self.transform_and_store(content_type, content)

    def _get_content_type_from_request(self, request, content):
        """tries to figure out the content type of a PUT body from a request

            the request may not have a content-type header available, if so
            we should use the contents itself to determine the content type
        """
        ct = request.get_header('content-type')
        if ct is not None:
            return ct
        content = re.sub('<?.*?>', '', content).strip()
        if content.startswith('<html') or content.startswith('<!DOCTYPE html'):
            return 'text/html'
        return 'application/x-silva-document-xml'

    security.declareProtected(SilvaPermissions.ChangeSilvaContent,
                                'manage_FTPget')
    def manage_FTPget(self):
        """return the raw XML-contents of this document"""
        editable = self.get_previewable()
        if editable is None:
            raise InternalError, 'no viewable version available'
        f = StringIO()
        editable.get_xml_content(f)
        return f.getvalue()

    security.declareProtected(SilvaPermissions.ChangeSilvaContent,
                                'manage_DAVget')
    manage_DAVget = manage_FTPget

InitializeClass(Document)

class DocumentVersion(CatalogedVersion):
    """Silva Document version.
    """
    meta_type = "Silva Document Version"
    __implements__ = IVersion

    security = ClassSecurityInfo()

    manage_options = (
        {'label':'Edit',       'action':'manage_main'},
        ) + CatalogedVersion.manage_options

    def __init__(self, id, title):
        DocumentVersion.inheritedAttribute('__init__')(self, id, title)
        self.content = ParsedXML('content', '<doc></doc>')

    # display edit screen as main management screen
    security.declareProtected('View management screens', 'manage_main')
    manage_main = PageTemplateFile('www/documentVersionEdit', globals())

    def to_xml(self, context):
        f = context.f
        f.write('<title>%s</title>' % translateCdata(self.get_title()))
        self.get_xml_content(f)
        export_metadata(self, context)

    def get_xml_content(self, f):
        self.content.documentElement.writeStream(f)

    def manage_afterClone(self, item):
        # if we're a copy, clear cache
        # XXX should be part of workflow system
        self.clearEditorCache()

    security.declareProtected(SilvaPermissions.AccessContentsInformation,
                              'fulltext')
    def fulltext(self):
        """Return the content of this object without any xml"""
        if self.version_status() == 'unapproved':
            return ''
        return ' '.join((self.get_title(),
            self._flattenxml(self.content_xml())))

    security.declareProtected(SilvaPermissions.AccessContentsInformation,
                              'content_xml')
    def content_xml(self):
        """Returns the documentElement of the content's XML
        """
        s = StringIO()
        self.content.documentElement.writeStream(s)
        value = s.getvalue()
        s.close()
        return value

    def _flattenxml(self, xmlinput):
        """Cuts out all the XML-tags, helper for fulltext (for content-objects)
        """
        # XXX this need to be fixed by using ZCTextIndex or the like
        return xmlinput

    def clearEditorCache(self):
        """ Clears editor cache for this version
        """
        editor_service = self.service_editor
        document_element = self.content.documentElement
        editor_service.clearCache(document_element)

InitializeClass(DocumentVersion)

manage_addDocumentForm = PageTemplateFile("www/documentAdd", globals(),
                                          __name__='manage_addDocumentForm')

def manage_addDocument(self, id, title, REQUEST=None):
    """Add a Document."""
    if not mangle.Id(self, id).isValid():
        return
    object = Document(id)
    self._setObject(id, object)
    object = getattr(self, id)
    object.manage_addProduct['SilvaDocument'].manage_addDocumentVersion(
        '0', title)
    object.create_version('0', None, None)
    add_and_edit(self, id, REQUEST)
    return ''

manage_addDocumentVersionForm = PageTemplateFile(
    "www/documentVersionAdd",
    globals(),
    __name__='manage_addDocumentVersionForm')

def manage_addDocumentVersion(self, id, title, REQUEST=None):
    """Add a Document version to the Silva-instance."""
    version = DocumentVersion(id, title)
    self._setObject(id, version)

    version = self._getOb(id)
    version.set_title(title)

    add_and_edit(self, id, REQUEST)
    return ''

def xml_import_handler(object, node):
    id = get_xml_id(node)
    title = get_xml_title(node)

    id = str(mangle.Id(object, id).unique())
    object.manage_addProduct['SilvaDocument'].manage_addDocument(id, title)

    newdoc = getattr(object, id)
    newdoc.sec_update_last_author_info()

    for child in node.childNodes:
        if child.nodeName == u'doc':
            version = getattr(newdoc, '0')
            childxml = writeStream(child).getvalue().encode('utf8')
            version.content.manage_edit(childxml) # expects utf8
        elif hasattr(newdoc, 'set_%s' % child.nodeName.encode('utf8')) \
                and child.nodeValue:
            getattr(newdoc, 'set_%s' % child.nodeName.encode('utf8'))(
                child.nodeValue.encode('utf8'))

    return newdoc

class SilvaDocumentPolicy(Persistent):

    __implements__ = IContainerPolicy

    def createDefaultDocument(self, container, title):
        container.manage_addProduct['SilvaDocument'].manage_addDocument(
            'index', title)
        container.index.sec_update_last_author_info()

def document_factory(self, id, title, body):
    """Add a Document."""
    if not mangle.Id(self, id).isValid():
        return
    obj = Document(id).__of__(self)
    version = DocumentVersion('0', title).__of__(self)
    obj._setObject('0', version)
    obj.create_version('0', None, None)
    version = obj.get_editable()
    version.content.manage_edit(body)

    return obj

def _should_create_document(id, ct, body):
    rightct = (ct in ['application/x-silva-document-xml'])
    rightext = id.endswith('.slv')
    return rightct or rightext

contentObjectFactoryRegistry.registerFactory(
    document_factory,
    _should_create_document)