Browse Source

Make the archiver more extensible

by extracting methods. Now we are able to extend the module to implement other
archiving stategies.
pull/409/head
Guewen Baconnier 10 years ago
parent
commit
870dee8c6a
  1. 2
      record_archiver/data/cron.xml
  2. 92
      record_archiver/models/record_lifespan.py

2
record_archiver/data/cron.xml

@ -11,7 +11,7 @@
<field name="numbercall">-1</field>
<field eval="False" name="doall"/>
<field eval="'record.lifespan'" name="model"/>
<field eval="'_scheduler_record_archiver'" name="function"/>
<field eval="'_scheduler_archive_records'" name="function"/>
<field eval="'()'" name="args"/>
</record>

92
record_archiver/models/record_lifespan.py

@ -21,7 +21,7 @@ import logging
from datetime import datetime
from dateutil.relativedelta import relativedelta
from openerp.osv import orm, fields, osv
from openerp.osv import orm, fields
from openerp.tools import DEFAULT_SERVER_DATE_FORMAT as DATE_FORMAT
from openerp.tools.translate import _
@ -61,50 +61,70 @@ class RecordLifespan(orm.Model):
"Months must be a value greater than 0"),
]
def _scheduler_record_archiver(self, cr, uid, context=None):
def _scheduler_archive_records(self, cr, uid, context=None):
lifespan_ids = self.search(cr, uid, [], context=context)
_logger.info('Records archiver starts archiving records')
for lifespan_id in lifespan_ids:
try:
self.archive_records(cr, uid, [lifespan_id], context=context)
except osv.except_osv as e:
except orm.except_orm as e:
_logger.error("Archiver error:\n%s", e[1])
_logger.info('Rusty Records now rest in peace')
return True
def archive_records(self, cr, uid, ids, context=None):
""" Search and deactivate old records for each configured lifespan
def _archive_domain(self, cr, uid, lifespan, expiration_date,
context=None):
""" Returns the domain used to find the records to archive.
Can be inherited to change the archived records for a model.
"""
model = self.pool[lifespan.model]
domain = [('write_date', '<', expiration_date),
('company_id', '=', lifespan.company_id.id)]
if 'state' in model._columns.keys():
domain += [('state', 'in', ('done', 'cancel'))]
return domain
def _archive_lifespan_records(self, cr, uid, lifespan, context=None):
""" Archive the records for a lifespan, so for a model.
Can be inherited to customize the archive strategy.
The default strategy is to change the field ``active`` to False
on the records having a ``write_date`` older than the lifespan.
Only done and canceled records will be deactivated.
Only done and cancelled records will be deactivated.
"""
lifespans = self.browse(cr, uid, ids, context=context)
today = datetime.today()
for lifespan in lifespans:
model = self.pool[lifespan.model]
if not model:
raise osv.except_osv(
_('Error'),
_('Model %s not found') % lifespan.model)
if 'active' not in model._columns.keys():
raise osv.except_osv(
_('Error'),
_('Model %s has no active field') % lifespan.model)
delta = relativedelta(months=lifespan.months)
expiration_date = (today - delta).strftime(DATE_FORMAT)
domain = [('write_date', '<', expiration_date),
('company_id', '=', lifespan.company_id.id)]
if 'state' in model._columns.keys():
domain += [('state', 'in', ('done', 'cancel'))]
rec_ids = model.search(cr, uid, domain, context=context)
if not rec_ids:
continue
# use a SQL query to bypass tracking always messages on write for
# object inheriting mail.thread
query = ("UPDATE %s SET active = FALSE WHERE id in %%s"
) % model._table
cr.execute(query, (tuple(rec_ids),))
_logger.info(
'Archived %s %s older than %s',
len(rec_ids), lifespan.model, expiration_date)
model = self.pool[lifespan.model]
if not model:
raise orm.except_orm(
_('Error'),
_('Model %s not found') % lifespan.model)
if 'active' not in model._columns.keys():
raise orm.except_orm(
_('Error'),
_('Model %s has no active field') % lifespan.model)
delta = relativedelta(months=lifespan.months)
expiration_date = (today - delta).strftime(DATE_FORMAT)
domain = self._archive_domain(cr, uid, lifespan, expiration_date,
context=context)
rec_ids = model.search(cr, uid, domain, context=context)
if not rec_ids:
return
# use a SQL query to bypass tracking always messages on write for
# object inheriting mail.thread
query = ("UPDATE %s SET active = FALSE WHERE id in %%s"
) % model._table
cr.execute(query, (tuple(rec_ids),))
_logger.info(
'Archived %s %s older than %s',
len(rec_ids), lifespan.model, expiration_date)
def archive_records(self, cr, uid, ids, context=None):
""" Call the archiver for several record lifespans """
for lifespan in self.browse(cr, uid, ids, context=context):
self._archive_lifespan_records(cr, uid, lifespan, context=context)
return True
Loading…
Cancel
Save