odoo/addons/import_base/import_framework.py

502 lines
21 KiB
Python

# -*- coding: utf-8 -*-
##############################################################################
#
# OpenERP, Open Source Management Solution
# Copyright (C) 2004-2010 Tiny SPRL (<http://tiny.be>).
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
##############################################################################
import pprint
import mapper
import pooler
import tools
from tools.translate import _
from threading import Thread
import datetime
import logging
import StringIO
import traceback
pp = pprint.PrettyPrinter(indent=4)
_logger = logging.getLogger(__name__)
class import_framework(Thread):
"""
This class should be extends,
get_data and get_mapping have to extends
get_state_map and initialize can be extended
for advanced purpose get_default_hook can also be extended
@see dummy import for a minimal exemple
"""
"""
for import_object, this domain will avoid to find an already existing object
"""
DO_NOT_FIND_DOMAIN = [('id', '=', 0)]
#TODO don't use context to pass credential parameters
def __init__(self, obj, cr, uid, instance_name, module_name, email_to_notify=False, context=None):
Thread.__init__(self)
self.external_id_field = 'id'
self.obj = obj
self.cr = cr
self.uid = uid
self.instance_name = instance_name
self.module_name = module_name
self.context = context or {}
self.email = email_to_notify
self.table_list = []
self.initialize()
"""
Abstract Method to be implemented in
the real instance
"""
def initialize(self):
"""
init before import
usually for the login
"""
pass
def init_run(self):
"""
call after intialize run in the thread, not in the main process
TO use for long initialization operation
"""
pass
def get_data(self, table):
"""
@return: a list of dictionaries
each dictionnaries contains the list of pair external_field_name : value
"""
return [{}]
def get_link(self, from_table, ids, to_table):
"""
@return: a dictionaries that contains the association between the id (from_table)
and the list (to table) of id linked
"""
return {}
def get_external_id(self, data):
"""
@return the external id
the default implementation return self.external_id_field (that has 'id') by default
if the name of id field is different, you can overwrite this method or change the value
of self.external_id_field
"""
return data[self.external_id_field]
def get_mapping(self):
"""
@return: { TABLE_NAME : {
'model' : 'openerp.model.name',
#if true import the table if not just resolve dependencies, use for meta package, by default => True
#Not required
'import' : True or False,
#Not required
'dependencies' : [TABLE_1, TABLE_2],
#Not required
'hook' : self.function_name, #get the val dict of the object, return the same val dict or False
'map' : { @see mapper
'openerp_field_name' : 'external_field_name', or val('external_field_name')
'openerp_field_id/id' : ref(TABLE_1, 'external_id_field'), #make the mapping between the external id and the xml on the right
'openerp_field2_id/id_parent' : ref(TABLE_1,'external_id_field') #indicate a self dependencies on openerp_field2_id
'state' : map_val('state_equivalent_field', mapping), # use get_state_map to make the mapping between the value of the field and the value of the state
'text_field' : concat('field_1', 'field_2', .., delimiter=':'), #concat the value of the list of field in one
'description_field' : ppconcat('field_1', 'field_2', .., delimiter='\n\t'), #same as above but with a prettier formatting
'field' : call(callable, arg1, arg2, ..), #call the function with all the value, the function should send the value : self.callable
'field' : callable
'field' : call(method, val('external_field') interface of method is self, val where val is the value of the field
'field' : const(value) #always set this field to value
+ any custom mapper that you will define
}
},
}
"""
return {}
def default_hook(self, val):
"""
this hook will be apply on each table that don't have hook
here we define the identity hook
"""
return val
def _import_table(self, table):
data = self.get_data(table)
map = self.get_mapping()[table]['map']
hook = self.get_mapping()[table].get('hook', self.default_hook)
model = self.get_mapping()[table]['model']
final_data = []
for val in data:
res = hook(val)
if res:
final_data.append(res)
return self._save_data(model, dict(map), final_data, table)
def _save_data(self, model, mapping, datas, table):
"""
@param model: the model of the object to import
@param table : the external table where the data come from
@param mapping : definition of the mapping
@see: get_mapping
@param datas : list of dictionnaries
datas = [data_1, data_2, ..]
data_i is a map external field_name => value
and each data_i have a external id => in data_id['id']
"""
_logger.info(' Importing %s into %s' % (table, model))
if not datas:
return (0, 'No data found')
mapping['id'] = 'id_new'
res = []
self_dependencies = []
for k in mapping.keys():
if '_parent' in k:
self_dependencies.append((k[:-7], mapping.pop(k)))
for data in datas:
for k, field_name in self_dependencies:
data[k] = data.get(field_name) and self._generate_xml_id(data.get(field_name), table)
data['id_new'] = self._generate_xml_id(self.get_external_id(data), table)
fields, values = self._fields_mapp(data, mapping, table)
res.append(values)
model_obj = self.obj.pool.get(model)
if not model_obj:
raise ValueError(_("%s is not a valid model name.") % model)
_logger.debug(_(" fields imported : ") + str(fields))
(p, r, warning, s) = model_obj.import_data(self.cr, self.uid, fields, res, mode='update', current_module=self.module_name, noupdate=True, context=self.context)
for (field, field_name) in self_dependencies:
self._import_self_dependencies(model_obj, field, datas)
return (len(res), warning)
def _import_self_dependencies(self, obj, parent_field, datas):
"""
@param parent_field: the name of the field that generate a self_dependencies, we call the object referenced in this
field the parent of the object
@param datas: a list of dictionnaries
Dictionnaries need to contains
id_new : the xml_id of the object
field_new : the xml_id of the parent
"""
fields = ['id', parent_field]
for data in datas:
if data.get(parent_field):
values = [data['id_new'], data[parent_field]]
obj.import_data(self.cr, self.uid, fields, [values], mode='update', current_module=self.module_name, noupdate=True, context=self.context)
def _preprocess_mapping(self, mapping):
"""
Preprocess the mapping :
after the preprocces, everything is
callable in the val of the dictionary
use to allow syntaxical sugar like 'field': 'external_field'
instead of 'field' : value('external_field')
"""
map = dict(mapping)
for key, value in map.items():
if isinstance(value, basestring):
map[key] = mapper.value(value)
#set parent for instance of dbmapper
elif isinstance(value, mapper.dbmapper):
value.set_parent(self)
return map
def _fields_mapp(self,dict_sugar, openerp_dict, table):
"""
call all the mapper and transform data
to be compatible with import_data
"""
fields=[]
data_lst = []
mapping = self._preprocess_mapping(openerp_dict)
for key,val in mapping.items():
if key not in fields and dict_sugar:
fields.append(key)
value = val(dict(dict_sugar))
data_lst.append(value)
return fields, data_lst
def _generate_xml_id(self, name, table):
"""
@param name: name of the object, has to be unique in for a given table
@param table : table where the record we want generate come from
@return: a unique xml id for record, the xml_id will be the same given the same table and same name
To be used to avoid duplication of data that don't have ids
"""
sugar_instance = self.instance_name
name = name.replace('.', '_').replace(',', '_')
return sugar_instance + "_" + table + "_" + name
"""
Public interface of the framework
those function can be use in the callable function defined in the mapping
"""
def xml_id_exist(self, table, external_id):
"""
Check if the external id exist in the openerp database
in order to check if the id exist the table where it come from
should be provide
@return the xml_id generated if the external_id exist in the database or false
"""
if not external_id:
return False
xml_id = self._generate_xml_id(external_id, table)
id = self.obj.pool.get('ir.model.data').search(self.cr, self.uid, [('name', '=', xml_id), ('module', '=', self.module_name)])
return id and xml_id or False
def name_exist(self, table, name, model):
"""
Check if the object with the name exist in the openerp database
in order to check if the id exist the table where it come from
should be provide and the model of the object
"""
fields = ['name']
data = [name]
return self.import_object(fields, data, model, table, name, [('name', '=', name)])
def get_mapped_id(self, table, external_id, context=None):
"""
@return return the databse id linked with the external_id
"""
if not external_id:
return False
xml_id = self._generate_xml_id(external_id, table)
return self.obj.pool.get('ir.model.data').get_object_reference(self.cr, self.uid, self.module_name, xml_id)[1]
def import_object_mapping(self, mapping, data, model, table, name, domain_search=False):
"""
same as import_objects but instead of two list fields and data,
this method take a dictionnaries : external_field : value
and the mapping similar to the one define in 'map' key
@see import_object, get_mapping
"""
fields, datas = self._fields_mapp(data, mapping, table)
return self.import_object(fields, datas, model, table, name, domain_search)
def import_object(self, fields, data, model, table, name, domain_search=False):
"""
This method will import an object in the openerp, usefull for field that is only a char in sugar and is an object in openerp
use import_data that will take care to create/update or do nothing with the data
this method return the xml_id
To be use, when you want to create an object or link if already exist
use DO_NOT_LINK_DOMAIN to create always a new object
@param fields: list of fields needed to create the object without id
@param data: the list of the data, in the same order as the field
ex : fields = ['firstname', 'lastname'] ; data = ['John', 'Mc donalds']
@param model: the openerp's model of the create/update object
@param table: the table where data come from in sugarcrm, no need to fit the real name of openerp name, just need to be unique
@param unique_name: the name of the object that we want to create/update get the id
@param domain_search : the domain that should find the unique existing record
@return: the xml_id of the ressources
"""
domain_search = not domain_search and [('name', 'ilike', name)] or domain_search
obj = self.obj.pool.get(model)
if not obj: #if the model doesn't exist
return False
xml_id = self._generate_xml_id(name, table)
xml_ref = self.mapped_id_if_exist(model, domain_search, table, name)
fields.append('id')
data.append(xml_id)
obj.import_data(self.cr, self.uid, fields, [data], mode='update', current_module=self.module_name, noupdate=True, context=self.context)
return xml_ref or xml_id
def mapped_id_if_exist(self, model, domain, table, name):
"""
To be use when we want link with and existing object, if the object don't exist
just ignore.
@param domain : search domain to find existing record, should return a unique record
@param xml_id: xml_id give to the mapping
@param name: external_id or name of the object to create if there is no id
@param table: the name of the table of the object to map
@return : the xml_id if the record exist in the db, False otherwise
"""
obj = self.obj.pool.get(model)
ids = obj.search(self.cr, self.uid, domain, context=self.context)
if ids:
xml_id = self._generate_xml_id(name, table)
ir_model_data_obj = obj.pool.get('ir.model.data')
id = ir_model_data_obj._update(self.cr, self.uid, model,
self.module_name, {}, mode='update', xml_id=xml_id,
noupdate=True, res_id=ids[0], context=self.context)
return xml_id
return False
def set_table_list(self, table_list):
"""
Set the list of table to import, this method should be call before run
@param table_list: the list of external table to import
['Leads', 'Opportunity']
"""
self.table_list = table_list
def run(self):
"""
Import all data into openerp,
this is the Entry point to launch the process of import
"""
self.data_started = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
self.cr = pooler.get_db(self.cr.dbname).cursor()
error = False
result = []
try:
self.init_run()
imported = set() #to invoid importing 2 times the sames modules
for table in self.table_list:
to_import = self.get_mapping()[table].get('import', True)
if not table in imported:
res = self._resolve_dependencies(self.get_mapping()[table].get('dependencies', []), imported)
result.extend(res)
if to_import:
(position, warning) = self._import_table(table)
result.append((table, position, warning))
imported.add(table)
self.cr.commit()
except Exception, err:
sh = StringIO.StringIO()
traceback.print_exc(file=sh)
error = sh.getvalue()
print error
self.date_ended = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
self._send_notification_email(result, error)
self.cr.close()
def _resolve_dependencies(self, dep, imported):
"""
import dependencies recursively
and avoid to import twice the same table
"""
result = []
for dependency in dep:
if not dependency in imported:
to_import = self.get_mapping()[dependency].get('import', True)
res = self._resolve_dependencies(self.get_mapping()[dependency].get('dependencies', []), imported)
result.extend(res)
if to_import:
r = self._import_table(dependency)
(position, warning) = r
result.append((dependency, position, warning))
imported.add(dependency)
return result
def _send_notification_email(self, result, error):
if not self.email:
return False
email_obj = self.obj.pool.get('mail.message')
email_id = email_obj.create(self.cr, self.uid, {
'email_from' : 'import@module.openerp',
'email_to' : self.email,
'body_text' : self.get_email_body(result, error),
'subject' : self.get_email_subject(result, error),
'auto_delete' : True})
email_obj.send(self.cr, self.uid, [email_id])
if error:
_logger.error(_("Import failed due to an unexpected error"))
else:
_logger.info(_("Import finished, notification email sended"))
def get_email_subject(self, result, error=False):
"""
This method define the subject of the email send by openerp at the end of import
@param result: a list of tuple
(table_name, number_of_record_created/updated, warning) for each table
@return the subject of the mail
"""
if error:
return _("Data Import failed at %s due to an unexpected error") % self.date_ended
return _("Import of your data finished at %s") % self.date_ended
def get_email_body(self, result, error=False):
"""
This method define the body of the email send by openerp at the end of import. The body is separated in two part
the header (@see get_body_header), and the generate body with the list of table and number of record imported.
If you want to keep this it's better to overwrite get_body_header
@param result: a list of tuple
(table_name, number_of_record_created/updated, warning) for each table
@return the subject of the mail
"""
body = _("started at %s and finished at %s \n") % (self.data_started, self.date_ended)
if error:
body += _("but failed, in consequence no data were imported to keep database consistency \n error : \n") + error
for (table, nb, warning) in result:
if not warning:
warning = _("with no warning")
else:
warning = _("with warning : %s") % warning
body += _("%s has been successfully imported from %s %s, %s \n") % (nb, self.instance_name, table, warning)
return self.get_body_header(result) + "\n\n" + body
def get_body_header(self, result):
"""
@return the first sentences written in the mail's body
"""
return _("The import of data \n instance name : %s \n") % self.instance_name
#TODO documentation test
def run_test(self):
back_get_data = self.get_data
back_get_link = self.get_link
back_init = self.initialize
self.get_data = self.get_data_test
self.get_link = self.get_link_test
self.initialize = self.intialize_test
self.run()
self.get_data = back_get_data
self.get_link = back_get_link
self.initialize = back_init
def get_data_test(self, table):
return [{}]
def get_link_test(self, from_table, ids, to_table):
return {}
def intialize_test(self):
pass
# vim:expandtab:smartindent:tabstop=4:softtabstop=4:shiftwidth=4: