############################################################################## # # Copyright (c) 2002 Nexedi SARL. All Rights Reserved. # Copyright (c) 2001 Zope Corporation and Contributors. All Rights Reserved. # # This software is subject to the provisions of the Zope Public License, # Version 2.0 (ZPL). A copy of the ZPL should accompany this distribution. # THIS SOFTWARE IS PROVIDED "AS IS" AND ANY AND ALL EXPRESS OR IMPLIED # WARRANTIES ARE DISCLAIMED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED # WARRANTIES OF TITLE, MERCHANTABILITY, AGAINST INFRINGEMENT, AND FITNESS # FOR A PARTICULAR PURPOSE # ############################################################################## from Persistence import Persistent import Acquisition import ExtensionClass from string import lower, split, join from thread import get_ident from DateTime import DateTime from Products.PluginIndexes.common.randid import randid from Products.CMFCore.Expression import Expression from Acquisition import aq_parent, aq_inner, aq_base, aq_self from zLOG import LOG import time import sys UID_BUFFER_SIZE = 900 MAX_UID_BUFFER_SIZE = 20000 class Catalog(Persistent, Acquisition.Implicit, ExtensionClass.Base): """ An Object Catalog An Object Catalog maintains a table of object metadata, and a series of manageable indexes to quickly search for objects (references in the metadata) that satisfy a search where_expression. This class is not Zope specific, and can be used in any python program to build catalogs of objects. Note that it does require the objects to be Persistent, and thus must be used with ZODB3. uid -> the (local) universal ID of objects path -> the (local) path of objects bgrain defined in meyhods... TODO: - optmization: indexing objects should be deferred until timeout value or end of transaction """ _after_clear_reserved = 0 def __init__(self): self.schema = {} # mapping from attribute name to column self.names = {} # mapping from column to attribute name self.indexes = {} # empty mapping def clear(self): """ Clears the catalog by calling a list of methods """ methods = self.sql_clear_catalog for method_name in methods: method = getattr(self,method_name) try: method() except: pass self._after_clear_reserved = 1 def clearReserved(self): """ Clears reserved uids """ method_id = self.sql_catalog_clear_reserved method = getattr(self, method_id) method() self._after_clear_reserved = 1 def __getitem__(self, uid): """ Get an object by UID Note: brain is defined in Z SQL Method object """ method = getattr(self, self.sql_getitem_by_uid) search_result = method(uid = uid) if len(search_result) > 0: return search_result[0] else: return None def editSchema(self, names_list): """ Builds a schema from a list of strings Splits each string to build a list of attribute names Columns on the database should not change during this operations """ i = 0 schema = {} names = {} for cid in self.getColumnIds(): name_list = [] for name in names_list[i].split(): schema[name] = cid name_list += [name,] names[cid] = tuple(name_list) i += 1 self.schema = schema self.names = names def getColumnIds(self): """ Calls the show column method and returns dictionnary of Field Ids XXX This should be cached """ method_name = self.sql_catalog_schema keys = {} for table in self.getCatalogSearchTableIds(): try: method = getattr(self, method_name) search_result = method(table=table) for c in search_result: keys[c.Field] = 1 keys['%s.%s' % (table, c.Field)] = 1 # Is this inconsistent ? except: pass keys = keys.keys() keys.sort() return keys def getColumnMap(self): """ Calls the show column method and returns dictionnary of Field Ids XXX This should be cached """ method_name = self.sql_catalog_schema keys = {} for table in self.getCatalogSearchTableIds(): try: method = getattr(self, method_name) search_result = method(table=table) for c in search_result: key = c.Field if not keys.has_key(key): keys[c.Field] = [] keys[key].append(table) key = '%s.%s' % (table, c.Field) if not keys.has_key(key): keys[key] = [] keys[key].append(table) # Is this inconsistent ? except: pass return keys def getResultColumnIds(self): """ Calls the show column method and returns dictionnary of Field Ids """ method_name = self.sql_catalog_schema keys = {} for table in self.getCatalogSearchTableIds(): try: method = getattr(self, method_name) search_result = method(table=table) for c in search_result: keys['%s.%s' % (table, c.Field)] = 1 except: pass keys = keys.keys() keys.sort() return keys def getTableIds(self): """ Calls the show table method and returns dictionnary of Field Ids """ keys = [] method_name = self.sql_catalog_tables try: method = getattr(self, method_name) search_result = method() for c in search_result: keys.append(c[0]) except: pass return keys # the cataloging API def produceUid(self): """ Produces reserved uids in advance """ method_id = self.sql_catalog_produce_reserved method = getattr(self, method_id) thread_id = get_ident() uid_list = getattr(self, '_v_uid_buffer', []) if self._after_clear_reserved: # Reset uid list after clear reserved self._after_clear_reserved = 0 uid_list = [] if len(uid_list) < UID_BUFFER_SIZE: date = DateTime() new_uid_list = method(count = UID_BUFFER_SIZE, thread_id=thread_id, date=date) uid_list.extend( filter(lambda x: x != 0, map(lambda x: x.uid, new_uid_list ))) self._v_uid_buffer = uid_list def newUid(self): """ This is where uid generation takes place. We should consider a multi-threaded environment with multiple ZEO clients on a single ZEO server. The main risk is the following: - objects a/b/c/d/e/f are created (a is parent of b which is parent of ... of f) - one reindexing node N1 starts reindexing f - another reindexing node N2 starts reindexing e - there is a strong risk that N1 and N2 start reindexing at the same time and provide different uid values for a/b/c/d/e Similar problems may happen with relations and acquisition of uid values (ex. order_uid) with the risk of graph loops """ self.produceUid() uid_list = getattr(self, '_v_uid_buffer', []) if len(uid_list) > 0: return uid_list.pop() else: raise CatalogError("Could not retrieve new uid") def catalogObject(self, object, path, is_object_moved=0): """ Adds an object to the Catalog by calling all SQL methods and providing needed arguments. 'object' is the object to be cataloged 'uid' is the unique Catalog identifier for this object """ #LOG('Catalog object:',0,str(path)) # Prepare the dictionnary of values kw = {} # Check if already Catalogued if hasattr(object, 'uid'): # Try to use existing uid # WARNING COPY PASTE.... uid = object.uid else: # Look up in (previous) path uid = 0 if is_object_moved: index = uid # We trust the current uid else: index = self.getUidForPath(path) if index: if (uid != index): # Update uid attribute of object uid = int(index) #LOG("Write Uid",0, "uid %s index %s" % (uid, index)) object.uid = uid # We will check if there is an filter on this # method, if so we may not call this zsqlMethod # for this object for method_name in self.sql_update_object: if self.isMethodFiltered(method_name): if self.filter_dict.has_key(method_name): portal_type = object.getPortalType() if portal_type not in (self.filter_dict[method_name]['type']): #LOG('catalog_object',0,'XX1 this method is broken because not in types: %s' % method_name) continue else: expression = self.filter_dict[method_name]['expression_instance'] if expression is not None: econtext = self.getExpressionContext(object) result = expression(econtext) if not result: #LOG('catalog_object',0,'XX2 this method is broken because expression: %s' % method_name) continue #LOG('catalog_object',0,'this method is not broken: %s' % method_name) # Get the appropriate SQL Method # Lookup by path is required because of OFS Semantics method = getattr(self, method_name) if method.meta_type == "Z SQL Method": # Build the dictionnary of values arguments = method.arguments_src for arg in split(arguments): try: value = getattr(object, arg) if callable(value): value = value() kw[arg] = value except: #LOG("SQLCatalog Warning: Callable value could not be called",0,str((path, arg, method_name))) kw[arg] = None method = aq_base(method).__of__(object.__of__(self)) # Use method in the context of object # Generate UID kw['path'] = path kw['uid'] = int(index) kw['insert_catalog_line'] = 0 #LOG("SQLCatalog Warning: insert_catalog_line, case1 value",0,0) # LOG # LOG("Call SQL Method %s with args:" % method_name,0, str(kw)) # Alter row #LOG("Call SQL Method %s with args:" % method_name,0, str(kw)) method(**kw) else: # Get the appropriate SQL Method # Lookup by path is required because of OFS Semantics if uid: # Make sure no duplicates - ie. if an object with different path has same uid, we need a new uid # This can be very dangerous with relations stored in a category table (CMFCategory) # This is why we recommend completely reindexing subobjects after any change of id catalog_path = self.getPathForUid(uid) if catalog_path == "reserved": # Reserved line in catalog table insert_catalog_line = 0 #LOG("SQLCatalog Warning: insert_catalog_line, case2",0,insert_catalog_line) elif catalog_path is None: # No line in catalog table insert_catalog_line = 1 #LOG("SQLCatalog Warning: insert_catalog_line, case3",0,insert_catalog_line) else: #LOG('SQLCatalog WARNING',0,'assigning new uid to already catalogued object %s' % path) uid = 0 insert_catalog_line = 0 #LOG("SQLCatalog Warning: insert_catalog_line, case4",0,insert_catalog_line) if not uid: # Generate UID index = self.newUid() object.uid = index insert_catalog_line = 0 #LOG("SQLCatalog Warning: insert_catalog_line, case5",0,insert_catalog_line) else: index = uid for method_name in self.sql_catalog_object: # We will check if there is an filter on this # method, if so we may not call this zsqlMethod # for this object if self.isMethodFiltered(method_name): if self.filter_dict.has_key(method_name): portal_type = object.getPortalType() if portal_type not in (self.filter_dict[method_name]['type']): #LOG('catalog_object',0,'XX1 this method is broken because not in types: %s' % method_name) continue else: expression = self.filter_dict[method_name]['expression_instance'] if expression is not None: econtext = self.getExpressionContext(object) result = expression(econtext) if not result: #LOG('catalog_object',0,'XX2 this method is broken because expression: %s' % method_name) continue #LOG('catalog_object',0,'this method is not broken: %s' % method_name) method = getattr(self, method_name) if method.meta_type == "Z SQL Method": # Build the dictionnary of values arguments = method.arguments_src for arg in split(arguments): try: value = getattr(object, arg) if callable(value): value = value() kw[arg] = value except: #LOG("SQLCatalog Warning: Callable value could not be called",0,str((path, arg, method_name))) kw[arg] = None method = aq_base(method).__of__(object.__of__(self)) # Use method in the context of object # Generate UID kw['path'] = path kw['uid'] = index kw['insert_catalog_line'] = insert_catalog_line # LOG # Alter row # Create row try: zope_root = self.getPortalObject().aq_parent root_indexable = int(getattr(zope_root,'isIndexable',1)) if root_indexable: #LOG("Call SQL Method %s with args:" % method_name,0, str(kw)) method(**kw) except: LOG("SQLCatalog Warning: could not catalog object with method %s" % method_name,100, str(path)) raise #except: # # # This is a real LOG message # # # which is required in order to be able to import .zexp files # LOG("SQLCatalog Warning: could not catalog object with method %s" % method_name, # 100,str(path)) def uncatalogObject(self, path): """ Uncatalog and object from the Catalog. Note, the uid must be the same as when the object was catalogued, otherwise it will not get removed from the catalog This method should not raise an exception if the uid cannot be found in the catalog. XXX Add filter of methods """ #LOG('Uncatalog object:',0,str(path)) uid = self.getUidForPath(path) methods = self.sql_uncatalog_object for method_name in methods: method = getattr(self, method_name) try: #if 1: method(uid = uid) except: # This is a real LOG message # which is required in order to be able to import .zexp files LOG("SQLCatalog Warning: could not uncatalog object uid %s with method %s" % (uid, method_name),0,str(path)) def uniqueValuesFor(self, name): """ return unique values for FieldIndex name """ method = getattr(self, self.sql_unique_values) return method() def getPaths(self): """ Returns all object paths stored inside catalog """ method = getattr(self, self.sql_catalog_paths) return method() def getUidForPath(self, path): """ Looks up into catalog table to convert path into uid """ try: if path is None: return None # Get the appropriate SQL Method method = getattr(self, self.sql_getitem_by_path) search_result = method(path = path) # If not emptyn return first record if len(search_result) > 0: return search_result[0].uid else: return None except: # This is a real LOG message # which is required in order to be able to import .zexp files LOG("Warning: could not find uid from path",0,str(path)) return None def hasPath(self, path): """ Checks if path is catalogued """ return self.getUidForPath(path) is not None def getPathForUid(self, uid): """ Looks up into catalog table to convert uid into path """ try: if uid is None: return None # Get the appropriate SQL Method method = getattr(self, self.sql_getitem_by_uid) search_result = method(uid = uid) # If not empty return first record if len(search_result) > 0: return search_result[0].path else: return None except: # This is a real LOG message # which is required in order to be able to import .zexp files LOG("Warning: could not find path from uid",0,str(uid)) return None def getMetadataForUid(self, uid): """ Accesses a single record for a given uid """ if uid is None: return None # Get the appropriate SQL Method method = getattr(self, self.sql_getitem_by_uid) brain = method(uid = uid)[0] result = {} for k in brain.__record_schema__.keys(): result[k] = getattr(brain,k) return result def getIndexDataForUid(self, uid): """ Accesses a single record for a given uid """ return self.getMetadataForUid(uid) def getMetadataForPath(self, path): """ Accesses a single record for a given path """ try: if uid is None: return None # Get the appropriate SQL Method method = getattr(self, self.sql_getitem_by_path) brain = method(path = path)[0] result = {} for k in brain.__record_schema__.keys(): result[k] = getattr(brain,k) return result except: # This is a real LOG message # which is required in order to be able to import .zexp files LOG("Warning: could not find uid from path",0,str(path)) return None def getIndexDataForPath(self, path): """ Accesses a single record for a given path """ return self.getMetadataForPath(path) def buildSQLQuery(self, query_table='catalog', REQUEST=None, **kw): """ Builds a complex SQL query to simulate ZCalatog behaviour """ # Get search arguments: if REQUEST is None and (kw is None or kw == {}): # We try to get the REQUEST parameter # since we have nothing handy try: REQUEST=self.REQUEST except AttributeError: pass # If kw is not set, then use REQUEST instead if kw is None or kw == {}: kw = REQUEST acceptable_key_map = self.getColumnMap() acceptable_keys = acceptable_key_map.keys() full_text_search_keys = self.sql_catalog_full_text_search_keys keyword_search_keys = self.sql_catalog_keyword_search_keys # We take additional parameters from the REQUEST # and give priority to the REQUEST if REQUEST is not None: for key in acceptable_keys: if REQUEST.has_key(key): # Only copy a few keys from the REQUEST if key in self.sql_catalog_request_keys: kw[key] = REQUEST[key] # Let us start building the where_expression if kw: where_expression = [] from_table_dict = {'catalog': 1} # Always include catalog table for key in kw.keys(): # Do not use kw.items() because this consumes much more memory value = kw[key] if key not in ('where_expression', 'sort-on', 'sort_on', 'sort-order', 'sort_order'): # Make sure key belongs to schema if key in acceptable_keys: if key.find('.') < 0: # if the key is only used by one table, just append its name if len(acceptable_key_map[key]) == 1 : key = acceptable_key_map[key][0] + '.' + key # query_table specifies what table name should be used elif query_table: key = query_table + '.' + key elif key == 'uid': # uid is always ambiguous so we can only change it here key = 'catalog.uid' # Add table to table dict from_table_dict[acceptable_key_map[key][0]] = 1 # We use catalog by default # Default case: variable equality if type(value) is type(''): if value != '': # we consider empty string as Non Significant if value == '=': # But we consider the sign = as empty string value='' if '%' in value: where_expression += ["%s LIKE '%s'" % (key, value)] elif value[0] == '>': where_expression += ["%s > '%s'" % (key, value[1:])] elif value[0] == '<': where_expression += ["%s < '%s'" % (key, value[1:])] elif key in keyword_search_keys: # We must add % in the request to simulate the catalog where_expression += ["%s LIKE '%%%s%%'" % (key, value)] elif key in full_text_search_keys: # We must add % in the request to simulate the catalog where_expression += ["MATCH %s AGAINST ('%s')" % (key, value)] else: where_expression += ["%s = '%s'" % (key, value)] elif type(value) is type([]) or type(value) is type(()): # We have to create an OR from tuple or list query_item = [] for value_item in value: if value_item != '': # we consider empty string as Non Significant # also for lists if type(value_item) in (type(1), type(1.0)): query_item += ["%s = %s" % (key, value_item)] else: if '%' in value_item: query_item += ["%s LIKE '%s'" % (key, str(value_item))] elif key in keyword_search_keys: # We must add % in the request to simulate the catalog query_item += ["%s LIKE '%%%s%%'" % (key, str(value_item))] elif key in full_text_search_keys: # We must add % in the request to simulate the catalog query_item += ["MATCH %s AGAINST ('%s')" % (key, value)] else: query_item += ["%s = '%s'" % (key, str(value_item))] if len(query_item) > 0: where_expression += ['(%s)' % join(query_item, ' OR ')] else: where_expression += ["%s = %s" % (key, value)] elif key == 'where_expression': # Not implemented yet pass if kw.get('where_expression'): if len(where_expression) > 0: where_expression = "(%s) AND (%s)" % (kw['where_expression'], join(where_expression, ' AND ') ) else: where_expression = join(where_expression, ' AND ') # Compute "sort_index", which is a sort index, or none: if kw.has_key('sort-on'): sort_index=kw['sort-on'] elif hasattr(self, 'sort-on'): sort_index=getattr(self, 'sort-on') elif kw.has_key('sort_on'): sort_index=kw['sort_on'] else: sort_index=None # Compute the sort order if kw.has_key('sort-order'): so=kw['sort-order'] elif hasattr(self, 'sort-order'): so=getattr(self, 'sort-order') elif kw.has_key('sort_order'): so=kw['sort_order'] else: so=None # We must now turn so into a string if type(so) is not type('a'): so = 'ascending' # We must now turn sort_index into # a dict with keys as sort keys and values as sort order if type(sort_index) is type('a'): sort_index = [(sort_index, so)] elif type(sort_index) is not type(()) and type(sort_index) is not type([]): sort_index = None # If sort_index is a dictionnary # then parse it and change it sort_on = None if sort_index is not None: try: new_sort_index = [] for (k , v) in sort_index: if len(acceptable_key_map[key]) == 1 : k = acceptable_key_map[key][0] + '.' + k elif query_table: k = query_table + '.' + k if v == 'descending' or v == 'reverse': from_table_dict[acceptable_key_map[k][0]] = 1 # We need this table to sort on it new_sort_index += ['%s DESC' % k] else: from_table_dict[acceptable_key_map[k][0]] = 1 # We need this table to sort on it new_sort_index += ['%s' % k] sort_index = join(new_sort_index,',') sort_on = str(sort_index) except: pass # Use a dictionary at the moment. return { 'from_table_list' : from_table_dict.keys(), 'order_by_expression' : sort_on, 'where_expression' : where_expression } def queryResults(self, sql_method, REQUEST=None, used=None, **kw): """ Returns a list of brains from a set of constraints on variables """ query = self.buildSQLQuery(REQUEST=REQUEST, **kw) kw['where_expression'] = query['where_expression'] kw['sort_on'] = query['order_by_expression'] kw['from_table_list'] = query['from_table_list'] # Return the result #LOG('acceptable_keys',0,'acceptable_keys: %s' % str(acceptable_keys)) #LOG('acceptable_key_map',0,'acceptable_key_map: %s' % str(acceptable_key_map)) #LOG('queryResults',0,'kw: %s' % str(kw)) #LOG('queryResults',0,'from_table_list: %s' % str(from_table_dict.keys())) return sql_method(**kw) def searchResults(self, REQUEST=None, used=None, **kw): """ Builds a complex SQL where_expression to simulate ZCalatog behaviour """ """ Returns a list of brains from a set of constraints on variables """ # The used argument is deprecated and is ignored try: # Get the search method method = getattr(self, self.sql_search_results) # Return the result kw['used'] = used kw['REQUEST'] = REQUEST return self.queryResults(method, **kw) except: LOG("Warning: could not search catalog",0,'', error=sys.exc_info()) return [] __call__ = searchResults def countResults(self, REQUEST=None, used=None, **kw): """ Builds a complex SQL where_expression to simulate ZCalatog behaviour """ """ Returns the number of items which satisfy the where_expression """ try: # Get the search method #LOG("countResults: scr:",0,str(self.sql_count_results)) #LOG("countResults: used:",0,str(used)) #LOG("countResults: kw:",0,str(kw)) method = getattr(self, self.sql_count_results) # Return the result kw['used'] = used kw['REQUEST'] = REQUEST return self.queryResults(method, **kw) except: LOG("Warning: could not count catalog",0,str(self.sql_count_results), error=sys.exc_info()) return [[0]] class CatalogError(Exception): pass