Merge lp:~numerigraphe-team/ocb-addons/7.0-inventory-lines-sorted into lp:ocb-addons

[Lionel Sausin - Initiatives/Numérigraphe]

This branch makes the inventory lines sorted on location name, product name and serial number.
It greatly eases the work of inventory workers when counting big inventories.
This solution is backported from trunk-wms (future v8), with an additionnal improvement to sort on serial numbers' names instead of their IDs.

[Holger Brunn (Therp) (hbrunn)]

In this context, I would have written
self
instead of
self.pool.get('stock.inventory.line')
but given that it's the same in upstream, I'm fine with it as it is.

review: Approve (code review)

[Holger Brunn (Therp) (hbrunn)]

and thankfully, upstream is not me and did it right because self is not stock.inventory.line in this context. So please ignore the comment above...

[Raphaël Valyi - http://www.akretion.com (rvalyi)]

rather new feature while this will likely be rewritten in v8. I would say if you need it merge it, but not sure that should be in what OCB is defining itself.

review: Abstain

[Raphaël Valyi - http://www.akretion.com (rvalyi)]

Oh, this is actually a backport of trunk-wms. Then I'm fine with it!

review: Approve

[Loïc Bellier - Numérigraphe (lb-b)]

review: Approve (code review)

[Guewen Baconnier @ Camptocamp (gbaconnier-c2c)]

Couldn't a composed index on the columns "inventory_id, location_name, product_code, product_name, prodlot_name" be useful?

review: Needs Information

[Loïc Bellier - Numérigraphe (lb-b)]

Hi gbaconnier,

Yes it's useful to always order the lines on the same order.
Without this index, search for productions lot for a same product, or a same product on different locations can be painful on thousands of inventory lines.
This index is useful for having the same order on printing and screen without re-order.

[Holger Brunn (Therp) (hbrunn)]

While the index is a good thing to have, I merge this already as we can add the index afterwards

[Lionel Sausin - Initiatives/Numérigraphe (ls-initiatives)]

Thanks for merging. I've begun work on adding the index here: lp:~numerigraphe-team/ocb-addons/7.0-line-order-index
I still need to make sure it's worth pushing it. I'll soon be able to test on a DB with lots of rows: I'll keep you posted.

[Lionel Sausin - Initiatives/Numérigraphe (ls-initiatives)]

I've made a quick regarding the index, it seems not very useful - once the rows are filtered on inventory_id, PG prefers to simply do a table scan + sort rather than an index scan.
The only point I see in this index would be to let the DBA schedule a CLUSTER on it to speed up the table scan, but that's a bit overkill for most cases I suppose.

=== modified file 'stock/stock.py'
--- stock/stock.py	2014-03-04 13:08:40 +0000
+++ stock/stock.py	2014-03-11 18:17:12 +0000
@@ -2954,6 +2954,17 @@
     _name = "stock.inventory.line"
     _description = "Inventory Line"
     _rec_name = "inventory_id"
+    _order = "inventory_id, location_name, product_code, product_name, prodlot_name"
+
+    def _get_product_name_change(self, cr, uid, ids, context=None):
+        return self.pool.get('stock.inventory.line').search(cr, uid, [('product_id', 'in', ids)], context=context)
+
+    def _get_location_change(self, cr, uid, ids, context=None):
+        return self.pool.get('stock.inventory.line').search(cr, uid, [('location_id', 'in', ids)], context=context)
+        
+    def _get_prodlot_change(self, cr, uid, ids, context=None):
+        return self.pool.get('stock.inventory.line').search(cr, uid, [('prod_lot_id', 'in', ids)], context=context)
+
     _columns = {
         'inventory_id': fields.many2one('stock.inventory', 'Inventory', ondelete='cascade', select=True),
         'location_id': fields.many2one('stock.location', 'Location', required=True),
@@ -2963,6 +2974,18 @@
         'company_id': fields.related('inventory_id','company_id',type='many2one',relation='res.company',string='Company',store=True, select=True, readonly=True),
         'prod_lot_id': fields.many2one('stock.production.lot', 'Serial Number', domain="[('product_id','=',product_id)]"),
         'state': fields.related('inventory_id','state',type='char',string='Status',readonly=True),
+        'product_name': fields.related('product_id', 'name', type='char', string='Product name', store={
+                                                                                            'product.product': (_get_product_name_change, ['name', 'default_code'], 20),
+                                                                                            'stock.inventory.line': (lambda self, cr, uid, ids, c={}: ids, ['product_id'], 20),}),
+        'product_code': fields.related('product_id', 'default_code', type='char', string='Product code', store={
+                                                                                            'product.product': (_get_product_name_change, ['name', 'default_code'], 20),
+                                                                                            'stock.inventory.line': (lambda self, cr, uid, ids, c={}: ids, ['product_id'], 20),}),
+        'location_name': fields.related('location_id', 'complete_name', type='char', string='Location name', store={
+                                                                                            'stock.location': (_get_location_change, ['name', 'location_id', 'active'], 20),
+                                                                                            'stock.inventory.line': (lambda self, cr, uid, ids, c={}: ids, ['location_id'], 20),}),
+        'prodlot_name': fields.related('prod_lot_id', 'name', type='char', string='Serial Number name', store={
+                                                                                            'stock.production.lot': (_get_prodlot_change, ['name'], 20),
+                                                                                            'stock.inventory.line': (lambda self, cr, uid, ids, c={}: ids, ['prod_lot_id'], 20),}),
     }
 
     def _default_stock_location(self, cr, uid, context=None):