Logo Search packages:      
Sourcecode: zope-extendedpathindex version File versions  Download package


# Copyright (c) 2004 Zope Corporation and Plone Solutions

# ZPL 2.1 license

import logging

from Globals import DTMLFile
from BTrees.IIBTree import IISet, intersection, union, multiunion
from BTrees.OOBTree import OOBTree
from BTrees.OIBTree import OIBTree

from Products.PluginIndexes.common.util import parseIndexRequest
from Products.PluginIndexes.common import safe_callable
from Products.PluginIndexes.PathIndex.PathIndex import PathIndex

_marker = []
logger = logging.getLogger('ExtendedPathIndex')

00019 class ExtendedPathIndex(PathIndex):
    """ A path index stores all path components of the physical
    path of an object:

    Internal datastructure (regular pathindex):

    - a physical path of an object is split into its components

    - every component is kept as a  key of a OOBTree in self._indexes

    - the value is a mapping 'level of the path component' to
      'all docids with this path component on this level'

    In addition
    - there is a terminator (None) signifying the last component in the path


    meta_type = "ExtendedPathIndex"

    manage_options= (
        {'label': 'Settings',
         'action': 'manage_main',
         'help': ('ExtendedPathIndex','ExtendedPathIndex_Settings.stx')},

    query_options = ("query", "level", "operator", "depth", "navtree",

00049     def __init__(self, id, extra=None, caller=None):
        """ ExtendedPathIndex supports indexed_attrs """
        PathIndex.__init__(self, id, caller)

        def get(o, k, default):
            if isinstance(o, dict):
                return o.get(k, default)
                return getattr(o, k, default)

        attrs = get(extra, 'indexed_attrs', None)
        if attrs is None:
        if isinstance(attrs, str):
            attrs = attrs.split(',')
        attrs = filter(None, [a.strip() for a in attrs])

        if attrs:
            # We only index the first attribute so snip off the rest
            self.indexed_attrs = tuple(attrs[:1])

    def clear(self):
        self._index_parents = OOBTree()
        self._index_items = OIBTree()

00075     def insertEntry(self, comp, id, level, parent_path=None, object_path=None):
        """Insert an entry.

           parent_path is the path of the parent object

           path is the object path, it is assumed to be unique, i.e. there
           is a one to one mapping between physical paths and docids.  This
           will be large, and is only used for breadcrumbs.

           id is the docid

        PathIndex.insertEntry(self, comp, id, level)

        if parent_path is not None:
            if not self._index_parents.has_key(parent_path):
                self._index_parents[parent_path] = IISet()


        # We make the assumption that a full path corresponds one and only
        # one object.

        if object_path is not None:
            self._index_items[object_path] = id

00101     def index_object(self, docid, obj ,threshold=100):
        """ hook for (Z)Catalog """

        # PathIndex first checks for an attribute matching its id and
        # falls back to getPhysicalPath only when failing to get one.
        # The presence of 'indexed_attrs' overrides this behavior and
        # causes indexing of the custom attribute.

        attrs = getattr(self, 'indexed_attrs', None)
        if attrs:
            index = attrs[0]
            index = self.id

        f = getattr(obj, index, None)
        if f is not None:
            if safe_callable(f):
                    path = f()
                except AttributeError:
                    return 0
                path = f

            if not isinstance(path, (str, tuple)):
                raise TypeError('path value must be string or tuple '
                                'of strings: (%r, %s)' % (index, repr(path)))
                path = obj.getPhysicalPath()
            except AttributeError:
                return 0

        if isinstance(path, (list, tuple)):
            path = '/'+ '/'.join(path[1:])
        comps = filter(None, path.split('/'))
        parent_path = '/' + '/'.join(comps[:-1])

        # Make sure we reindex properly when path change
        if self._unindex.has_key(docid) and self._unindex.get(docid) != path:

        if not self._unindex.has_key(docid):

        for i in range(len(comps)):
            self.insertEntry(comps[i], docid, i)

        # Add terminator
        self.insertEntry(None, docid, len(comps)-1, parent_path, path)

        self._unindex[docid] = path
        return 1

00155     def unindex_object(self, docid):
        """ hook for (Z)Catalog """

        if not self._unindex.has_key(docid):
                       'Attempt to unindex nonexistent object with id '
                       '%s' % docid)

        # There is an assumption that paths start with /
        path = self._unindex[docid]
        if not path.startswith('/'):
            path = '/'+path
        comps =  path.split('/')
        parent_path = '/'.join(comps[:-1])

        def unindex(comp, level, docid=docid, parent_path=None,

                if not self._index[comp][level]:
                    del self._index[comp][level]

                if not self._index[comp]:
                    del self._index[comp]
                # Remove parent_path and object path elements
                if parent_path is not None:
                    if not self._index_parents[parent_path]:
                        del self._index_parents[parent_path]
                if object_path is not None:
                    del self._index_items[object_path]
            except KeyError:
                           'Attempt to unindex object with id '
                           '%s failed' % docid)

        for level in range(len(comps[1:])):
            comp = comps[level+1]
            unindex(comp, level)

        # Remove the terminator
        level = len(comps[1:])
        comp = None
        unindex(comp, level-1, parent_path=parent_path, object_path=path)

        del self._unindex[docid]

00205     def search(self, path, default_level=0, depth=-1, navtree=0,
        path is either a string representing a
        relative URL or a part of a relative URL or
        a tuple (path,level).

        level >= 0  starts searching at the given level
        level <  0  not implemented yet

        if isinstance(path, basestring):
            startlevel = default_level
            startlevel = int(path[1])
            path = path[0]

        absolute_path = isinstance(path, basestring) and path.startswith('/')

        comps = filter(None, path.split('/'))

        orig_comps = [''] + comps[:]
        # Optimization - avoid using the root set
        # as it is common for all objects anyway and add overhead
        # There is an assumption about catalog/index having
        # the same container as content
        if default_level == 0:
            indexpath = list(filter(None, self.getPhysicalPath()))
            while min(len(indexpath), len(comps)):
                if indexpath[0] == comps[0]:
                    del indexpath[0]
                    del comps[0]
                    startlevel += 1

        if len(comps) == 0:
            if depth == -1 and not navtree:
                return IISet(self._unindex.keys())

        # Make sure that we get depth = 1 if in navtree mode
        # unless specified otherwise

        orig_depth = depth
        if depth == -1:
            depth = 0 or navtree

        # Optimized navtree starting with absolute path
        if absolute_path and navtree and depth == 1 and default_level==0:
            set_list = []
            # Insert root element
            if navtree_start >= len(orig_comps):
                navtree_start = 0
            # create a set of parent paths to search
            for i in range(len(orig_comps), navtree_start, -1):
                parent_path = '/'.join(orig_comps[:i])
                parent_path = parent_path and parent_path or '/'
                except KeyError:
            return multiunion(set_list)
        # Optimized breadcrumbs
        elif absolute_path and navtree and depth == 0 and default_level==0:
            item_list = IISet()
            # Insert root element
            if navtree_start >= len(orig_comps):
                navtree_start = 0
            # create a set of parent paths to search
            for i in range(len(orig_comps), navtree_start, -1):
                parent_path = '/'.join(orig_comps[:i])
                parent_path = parent_path and parent_path or '/'
                except KeyError:
            return item_list
        # Specific object search
        elif absolute_path and orig_depth == 0 and default_level == 0:
                return IISet([self._index_items[path]])
            except KeyError:
                return IISet()
        # Single depth search
        elif absolute_path and orig_depth == 1 and default_level == 0:
            # only get objects contained in requested folder
                return self._index_parents[path]
            except KeyError:
                return IISet()
        # Sitemaps, relative paths, and depth queries
        elif startlevel >= 0:

            pathset = None # Same as pathindex
            navset  = None # For collecting siblings along the way
            depthset = None # For limiting depth

            if navtree and depth and \
                   self._index.has_key(None) and \
                navset = self._index[None][startlevel]

            for level in range(startlevel, startlevel+len(comps) + depth):
                if level-startlevel < len(comps):
                    comp = comps[level-startlevel]
                    if not self._index.has_key(comp) or not self._index[comp].has_key(level): 
                        # Navtree is inverse, keep going even for
                        # nonexisting paths
                        if navtree:
                            pathset = IISet()
                            return IISet()
                        pathset = intersection(pathset,
                    if navtree and depth and \
                           self._index.has_key(None) and \
                        navset  = union(navset, intersection(pathset,
                if level-startlevel >= len(comps) or navtree:
                    if self._index.has_key(None) and self._index[None].has_key(level):
                        depthset = union(depthset, intersection(pathset,

            if navtree:
                return union(depthset, navset) or IISet()
            elif depth:
                return depthset or IISet()
                return pathset or IISet()

            results = IISet()
            for level in range(0,self._depth + 1):
                ids = None
                error = 0
                for cn in range(0,len(comps)):
                    comp = comps[cn]
                        ids = intersection(ids,self._index[comp][level+cn])
                    except KeyError:
                        error = 1
                if error==0:
                    results = union(results,ids)
            return results

00352     def _apply_index(self, request, cid=''):
        """ hook for (Z)Catalog
            'request' --  mapping type (usually {"path": "..." }
             additionaly a parameter "path_level" might be passed
             to specify the level (see search())

            'cid' -- ???

        record = parseIndexRequest(request,self.id,self.query_options)
        if record.keys==None: return None

        level    = record.get("level", 0)
        operator = record.get('operator', self.useOperator).lower()
        depth    = getattr(record, 'depth', -1) # Set to 0 or navtree later
                                                # use getattr to get 0 value
        navtree  = record.get('navtree', 0)
        navtree_start  = record.get('navtree_start', 0)

        # depending on the operator we use intersection of union
        if operator == "or":  set_func = union
        else: set_func = intersection

        res = None
        for k in record.keys:
            rows = self.search(k,level, depth, navtree, navtree_start)
            res = set_func(res,rows)

        if res:
            return res, (self.id,)
            return IISet(), (self.id,)

00385     def getIndexSourceNames(self):
        """ return names of indexed attributes """

        # By default PathIndex advertises getPhysicalPath even
        # though the logic in index_object is different.

            return tuple(self.indexed_attrs)
        except AttributeError:
            return ('getPhysicalPath',)

    index_html = DTMLFile('dtml/index', globals())
    manage_workspace = DTMLFile('dtml/manageExtendedPathIndex', globals())

manage_addExtendedPathIndexForm = DTMLFile('dtml/addExtendedPathIndex', globals())

def manage_addExtendedPathIndex(self, id, extra=None, REQUEST=None, RESPONSE=None, URL3=None):
    """Add an extended path index"""
    return self.manage_addIndex(id, 'ExtendedPathIndex', extra=extra,

Generated by  Doxygen 1.6.0   Back to index