source: products/quintagroup.transmogrifier/branches/plone-2.1/quintagroup.transmogrifier/quintagroup/transmogrifier/catalogsource.py @ 436

from zope.interface import classProvides, implements
from zope.app.annotation.interfaces import IAnnotations

from collective.transmogrifier.interfaces import ISection, ISectionBlueprint

from Products.CMFCore import utils

from quintagroup.transmogrifier.logger import VALIDATIONKEY

class CatalogSourceSection(object):
    classProvides(ISectionBlueprint)
    implements(ISection)

    def __init__(self, transmogrifier, name, options, previous):
        self.previous = previous
        self.context = transmogrifier.context

        # next is for communication with 'logger' section
        self.anno = IAnnotations(transmogrifier)
        self.storage = self.anno.setdefault(VALIDATIONKEY, [])

        self.pathkey = options.pop('path-key', '_path')
        self.entrieskey = options.pop('entries-key', '_entries')

        # remove 'blueprint' option - it cannot be a query
        options.pop('blueprint')

        self.query = {}
        for k, v in options.items():
            for p in v.split(';'):
                params = p.split('=', 1)
                if len(params) == 1:
                    self.query[k] = p.strip()
                else :
                    q = self.query.setdefault(k, {})
                    q[params[0].strip()] = params[1].strip()

        self.catalog = utils.getToolByName(self.context, 'portal_catalog')

    def __iter__(self):
        for item in self.previous:
            yield item

        exported = []

        results = list(self.catalog(**self.query))
        results.sort(key=lambda x: x.getPath())
        for brain in results:
            # discussion items are indexed and they must be replaced to
            # content objects to which they correspond
            # we need to skip them
            if brain.portal_type == 'Discussion Item':
                path =  '/'.join(brain.getPath().split('/')[:-2])
                cp, id_ = path.rsplit('/', 1)
                brain = self.catalog(path=cp, id=id_)[0]
            else:
                path = brain.getPath()

            # folderish objects are tried to export twice:
            # when their contained items are exported and when they are
            # returned in catalog search results
            if path in exported:
                continue
            exported.append(path)

            # export also all parents of current object
            containers = []
            container_path = path.rsplit('/', 1)[0]
            while container_path:
                if container_path in exported:
                    container_path = container_path.rsplit('/', 1)[0]
                    continue
                contained = self.getContained(container_path)
                if contained:
                    exported.append(container_path)
                    containers.append({
                        self.pathkey: '/'.join(container_path.split('/')[2:]),
                        self.entrieskey: contained,
                    })
                container_path = container_path.rsplit('/', 1)[0]
            containers.reverse()
            # order metter for us
            for i in containers:
                self.storage.append(i[self.pathkey])
                yield i

            item = {
                self.pathkey: '/'.join(path.split('/')[2:]),
            }
            if brain.is_folderish:
                contained = self.getContained(path)
                if contained:
                    item[self.entrieskey] = contained

            self.storage.append(item[self.pathkey])
            yield item

        # cleanup
        if VALIDATIONKEY in self.anno:
            del self.anno[VALIDATIONKEY]

    def getContained(self, path):
        """ Return list of (object_id, portal_type) for objects that are returned by catalog
            and contained in folder with given 'path'.
        """
        results = []
        seen = []
        raw_results = self.catalog(path=path, **self.query)
        for brain in raw_results:
            current = brain.getPath()
            relative = current[len(path):]
            relative = relative.strip('/')
            if not relative:
                # it's object with path that was given in catalog query
                continue
            elif '/' in relative:
                # object stored in subfolders, we need append to results their parent folder
                parent_path = '/'.join([path, relative.split('/', 1)[0]])
                if parent_path not in seen:                    
                    res = self.catalog(path=path) #, meta_type='Folder')
                    for i in res:
                        if i.getPath() == parent_path:
                            results.append(i)
                            seen.append(parent_path)
                            break
            elif current not in seen:
                # object is directly stored in folder, that has path given in query
                seen.append(current)
                results.append(brain)
        contained = [(i.getId, str(i.portal_type)) for i in results]
        return tuple(contained)