Fix surname index for autocompletion

Fix check tool and start with importgrdb upgrade


svn: r15949
This commit is contained in:
Benny Malengier 2010-10-02 16:01:34 +00:00
parent 0ca48dd811
commit 371205ec70
5 changed files with 172 additions and 34 deletions

View File

@ -46,7 +46,7 @@ import logging
#
#-------------------------------------------------------------------------
from gen.lib import (MediaObject, Person, Family, Source, Event, Place,
Repository, Note, GenderStats, Researcher)
Repository, Note, GenderStats, Researcher, NameOriginType)
from gen.db.dbconst import *
from gen.utils.callback import Callback
from gen.db import (BsddbBaseCursor, DbReadBase)
@ -67,6 +67,41 @@ _SIGBASE = ('person', 'family', 'source', 'event',
DBERRS = (db.DBRunRecoveryError, db.DBAccessError,
db.DBPageNotFoundError, db.DBInvalidArgError)
#-------------------------------------------------------------------------
#
# Helper functions
#
#-------------------------------------------------------------------------
def find_surname(key, data):
"""
Creating a surname from raw data of a person, to use for sort and index
"""
return __index_surname(data[3][5])
def find_surname_name(key, data):
"""
Creating a surname from raw name, to use for sort and index
"""
return __index_surname(data[5])
def __index_surname(surn_list):
"""
All non pa/matronymic surnames are used in indexing.
pa/matronymic not as they change for every generation!
"""
if surn_list:
surn = " ".join([x[0] for x in surn_list if not (x[3][0] in [
NameOriginType.PATRONYMIC, NameOriginType.MATRONYMIC]) ])
else:
surn = ""
return str(surn)
#-------------------------------------------------------------------------
#
# class DbBookmarks
#
#-------------------------------------------------------------------------
class DbBookmarks(object):
def __init__(self, default=[]):
self.bookmarks = list(default) # want a copy (not an alias)
@ -1356,12 +1391,8 @@ class DbBsddbRead(DbReadBase, Callback):
return self.__has_handle(self.source_map, handle)
def __sortbyperson_key(self, person):
surnlist = self.person_map.get(str(person))[3][5]
if surnlist:
surn = " ".join([x[0] for x in surnlist])
else:
surn = ""
return locale.strxfrm(surn)
return locale.strxfrm(find_surname(str(person),
self.person_map.get(str(person))))
def __sortbyplace(self, first, second):
return locale.strcoll(self.place_map.get(str(first))[2],

View File

@ -25,17 +25,17 @@ from __future__ import with_statement
"""
methods to upgrade a database from version 13 to current version
"""
from bsddb import db
from gen.db import BSDDBTxn
from gen.lib.nameorigintype import NameOriginType
from gen.db.write import _mkname, SURNAMES
def gramps_upgrade_15(self):
"""Upgrade database from version 14 to 15. This upgrade adds:
* tagging
* surname list
"""
length = len(self.person_map)
length = len(self.person_map)+10
self.set_total(length)
# ---------------------------------
@ -94,7 +94,13 @@ def gramps_upgrade_15(self):
)
with BSDDBTxn(self.env, self.person_map) as txn:
txn.put(str(handle), new_person)
self.update()
self.update(length)
#surname is now different, remove secondary index with names
_db = db.DB(self.env)
try:
_db.remove(_mkname(self.full_name, SURNAMES), SURNAMES)
except db.DBNoSuchFileError:
pass
# Bump up database version. Separate transaction to save metadata.
with BSDDBTxn(self.env, self.metadata) as txn:

View File

@ -52,7 +52,7 @@ from gen.lib import (GenderStats, Person, Family, Event, Place, Source,
MediaObject, Repository, Note)
from gen.db import (DbBsddbRead, DbWriteBase, BSDDBTxn,
DbTxn, BsddbBaseCursor, DbVersionError,
DbUpgradeRequiredError,
DbUpgradeRequiredError, find_surname, find_surname_name,
DbUndoBSDDB as DbUndo)
from gen.db.dbconst import *
from gen.utils.callback import Callback
@ -120,10 +120,7 @@ KEY_TO_CLASS_MAP = {PERSON_KEY: Person.__name__,
#
# Helper functions
#
#-------------------------------------------------------------------------
def find_surname(key, data):
return str(data[3][5])
#-------------------------------------------------------------------------
def find_idmap(key, data):
return str(data[1])
@ -1315,7 +1312,8 @@ class DbBsddb(DbBsddbRead, DbWriteBase, UpdateCallback):
"""
Build surname list for use in autocompletion
"""
self.surname_list = sorted(map(unicode, set(self.surnames.keys())), key=locale.strxfrm)
self.surname_list = sorted(map(unicode, set(self.surnames.keys())),
key=locale.strxfrm)
def add_to_surname_list(self, person, batch_transaction):
"""
@ -1323,7 +1321,8 @@ class DbBsddb(DbBsddbRead, DbWriteBase, UpdateCallback):
"""
if batch_transaction:
return
name = unicode(person.get_primary_name().get_surname())
name = unicode(find_surname_name(person.handle,
person.get_primary_name().serialize()))
i = bisect.bisect(self.surname_list, name)
if 0 < i <= len(self.surname_list):
if self.surname_list[i-1] != name:
@ -1340,7 +1339,8 @@ class DbBsddb(DbBsddbRead, DbWriteBase, UpdateCallback):
If not then we need to remove the name from the list.
The function must be overridden in the derived class.
"""
name = str(person.get_primary_name().get_surname())
name = str(find_surname_name(person.handle,
person.get_primary_name().serialize()))
try:
cursor = self.surnames.cursor(txn=self.txn)
cursor.set(name)
@ -1400,7 +1400,10 @@ class DbBsddb(DbBsddbRead, DbWriteBase, UpdateCallback):
self.genderStats.count_person(person)
# Update surname list if necessary
if (old_person.primary_name.surname !=person.primary_name.surname):
if (find_surname_name(old_person.handle,
old_person.primary_name.serialize()) !=
find_surname_name(person.handle,
person.primary_name.serialize())):
self.remove_from_surname_list(old_person)
self.add_to_surname_list(person, transaction.batch)
else:

View File

@ -77,8 +77,9 @@ def find_surname(key, data):
"""
Return the surname from the data stream. Used for building a secondary
index.
This function is not needed, as we don't use the secondary index.
"""
return str(data[3][5])
return str("a")
def find_idmap(key, data):
"""
@ -138,6 +139,10 @@ class GrampsBSDDB(DbGrdb, UpdateCallback):
This is replaced for internal use by gen/db/dbdir.py
However, this class is still used for import of the 2.2.x
GRDB format. In 3.0+ this format is no longer used.
We only need to upgrade the old main tables.
That will be used to append data to the database this GrampsBSDDB is
imported to.
"""
def __init__(self, use_txn = True):
@ -1247,19 +1252,6 @@ class GrampsBSDDB(DbGrdb, UpdateCallback):
self.surname_list = list(set(self.surnames.keys()))
self.sort_surname_list()
def remove_from_surname_list(self, person):
"""
Check whether there are persons with the same surname left in
the database. If not then we need to remove the name from the list.
The function must be overridden in the derived class.
"""
name = str(person.get_primary_name().get_surname())
try:
if self.surnames.keys().count(name) == 1:
self.surname_list.remove(unicode(name))
except ValueError:
pass
def __get_obj_from_gramps_id(self, val, tbl, class_init, prim_tbl):
if tbl.has_key(str(val)):
#if str(val) in tbl:
@ -1563,6 +1555,8 @@ class GrampsBSDDB(DbGrdb, UpdateCallback):
self.gramps_upgrade_13()
if version < 14:
self.gramps_upgrade_14()
if version < 15:
self.gramps_upgrade_15()
LOG.debug("Upgrade time: %s %s", int(time.time()-t), "seconds")
def gramps_upgrade_10(self):
@ -2589,6 +2583,108 @@ class GrampsBSDDB(DbGrdb, UpdateCallback):
name_type, prefix, patronymic,
group_as, sort_as, display_as, call)
def gramps_upgrade_15(self):
"""Upgrade database from version 14 to 15. This upgrade adds:
* tagging
* surname list
"""
length = len(self.person_map)+10
self.set_total(length)
# ---------------------------------
# Modify Person
# ---------------------------------
for handle in self.person_map.keys():
person = self.person_map[handle]
(junk_handle, # 0
gramps_id, # 1
gender, # 2
primary_name, # 3
alternate_names, # 4
death_ref_index, # 5
birth_ref_index, # 6
event_ref_list, # 7
family_list, # 8
parent_family_list, # 9
media_list, # 10
address_list, # 11
attribute_list, # 12
urls, # 13
ord_list, # 14
psource_list, # 15
pnote_list, # 16
change, # 17
marker, # 18
pprivate, # 19
person_ref_list, # 20
) = person
new_primary_name = self.convert_name_15(primary_name)
new_alternate_names = [self.convert_name_15(altname) for altname in
alternate_names]
new_person = (junk_handle, # 0
gramps_id, # 1
gender, # 2
new_primary_name, # 3
new_alternate_names,# 4
death_ref_index, # 5
birth_ref_index, # 6
event_ref_list, # 7
family_list, # 8
parent_family_list, # 9
media_list, # 10
address_list, # 11
attribute_list, # 12
urls, # 13
ord_list, # 14
psource_list, # 15
pnote_list, # 16
change, # 17
marker, # 18
pprivate, # 19
person_ref_list, # 20
[] # 21, tags
)
the_txn = self.env.txn_begin()
self.person_map.put(str(handle), new_person, txn=the_txn)
the_txn.commit()
self.update(length)
#surname is now different, normally remove secondary index with names
#we skip this, as this database will not be used after the import
# Bump up database version. Separate transaction to save metadata.
the_txn = self.env.txn_begin()
self.metadata.put('version', 15, txn=the_txn)
the_txn.commit()
def convert_name_15(self, name):
(privacy, source_list, note_list, date,
first_name, surname, suffix, title,
name_type, prefix, patronymic,
group_as, sort_as, display_as, call) = name
connector = u""
origintype = (NameOriginType.NONE, u"")
patorigintype = (NameOriginType.PATRONYMIC, u"")
if patronymic.strip() == u"":
#no patronymic, create a single surname
surname_list = [(surname, prefix, True, origintype, connector)]
else:
#a patronymic, if no surname or equal as patronymic, a single surname
if (surname.strip() == u"") or (surname == patronymic and prefix == u""):
surname_list = [(patronymic, prefix, True, patorigintype, connector)]
else:
#two surnames, first patronymic, then surname which is primary
surname_list = [(patronymic, u"", False, patorigintype, u""),
(surname, prefix, True, origintype, connector)]
#return new value, add two empty strings for nick and family nick
return (privacy, source_list, note_list, date,
first_name, surname_list, suffix, title, name_type,
group_as, sort_as, display_as, call, u"", u"")
def set_auto_remove(self):
"""
BSDDB change log settings using new method with renamed attributes

View File

@ -878,6 +878,7 @@ class CheckIntegrity(object):
marker, # 18
pprivate, # 19
person_ref_list, # 20
tags, # 21
) = person
# Take apart person reference list:
new_person_ref_list = []
@ -935,6 +936,7 @@ class CheckIntegrity(object):
marker, # 18
pprivate, # 19
new_person_ref_list, # 20
tags, # 21
)
p = gen.lib.Person(new_person)
self.db.commit_person(p, self.trans)