Odoo13 新变化：存货核算

Odoo13将于2019年10月发布，本次发布也包含了大量的改进，例如，对存货核算的重构。
去掉了 产品历史价格product.price.history ，增加了 stock valuation layer 对象，重写了 存货计价的逻辑
具体提交记录如下：

[REF] stock_account: standard cost method with valuation layers

Valuation of standard move: We introduce _create_in_svl and
_create_out_svl and call these method in _action_done only on the
standard move, the other valuation methods still use the old code. We
hijack _account_entry_move to create journal entries anyway for the
valuation layer with a minimal diff. Helpers for the valuation layers
creation are added on the product because we will need to create similar
layer without stock move later on when we'll implement changes of cost
method. We also make a first implementation of the edition in the past
to make some tests pass, it will be completed with other cost method
implementation

Change of standard price: Before this commit, the wizard to change the
standard price was only displayed in perpetual, because it was used to
create accountign entries. We also use this wizard in manual now,
because we want to create a stock valuation layer. Fortunetaly, the
accounting fields of this wizard are not required if the product has a
manual valuation. And the new method is private.

We adapt the existing tests as well as we can and include the start of a
new battery of tests: test_stockvaluationlayer.py

task-1875873

[REF] stock_account: average cost method with valuation layers

One of the goal of this refactoring is also to run the vacuum when the
cost method is AVCO. This will be used allow already out landed cost
on AVCO and also to have a consistent value when going into negative.

We update _prepare_in_svl_vals to set the remaining_qty field that
will be used when fifo is implemented. On the stock move, we create the
layer in with the real price unit of the move and adapt
product_price_update_before_done. This method will now compute the AVCO
according to the quantity valued in stock (quantity_svl) instead of all
the quantities in stock (qty_available), which was wrong (qty_available
could contain consigned quants).

We also make sure to recompute the avco when editing a done move.

We don't implement yet the negative handling as we'll need fifo for
that.

task-1875873

[REF] stock_account: fifo "normal" cost method with valuation layers

We don't implement fifo negative handling in this commit.
We make sure we run fifo in average mode to get a proper remaining_qty
field.
_run_fifo is now a method on product (since it could be user outside
of a stock move context). It is really similar to its ancestor on the
stock move, with a few float_is_zero and compare more and a missing
remaing_value.

We move the valuation sanity check into a proper helper,
_sanity_check_for_valuation.

We complete the _create_correction_svl method with accounting entries.

task-1875873

[REF] stock_account: fifo "negative" cost method with valuation layers

Re-implement _run_fifo_vacuum on the product this time with
float_is_zero and without remaining_value.

Trigger the vacuum at each receipt. Before it was done in the scheduler
each night, meaning the valuation on the move was not really
understandable nor consistent (the estimated value could be fixed by
some other move than the next receipt).

task-1875873

同时，代发货模式，到岸成本，以及制造成本等，也进行了重构

[REF] stock_account: dropshipping valuation

We standardize the dropshipping by handling it as it was an receipt
directly followed by an out.

We also follow the same logic whatever the cost method, ie we don't run
fifo or update the AVCO.

task-1875873

[REF] mrp_account: adapt to valuation layers

A finished move is considered as an incoming move and will pass by
_create_in_svl. We thus compute the price_unit that will be set on the
valuation layer.

We also set a stat button on the MO displaying the valuation
layers, similar to the one on the picking.

task-1875873

[REF] stock_dropshipping: adapt to stock valuation layers

The tests of this module showed multiple issues with the anglos axon
accounting entries creation so we fix them by getting the value on the
valuation layer and not directly on the stock moves.

About the changes in
test_crossdock.py

To change a standard price, we need to open the wizard
test_lifo_price.py
To change a standard price, we need to open the wizard + read the
valuation on the layers.
test_stockvaluation.py
Now that we create a layer in and out for a dropshipped move, the
accounting entries are similar to an in followed by an out, thus we need
to increment the number of generated AML. Not that the balance of the
accounts stay the same.

There's a slight change in fifo_perpetual_anglosaxon_ordered though.
Let's say for a start that this config doesn't make sense. Ordered means
we don't read the value on the stock move to create the accounting
entries but only the standard price. We had to change 8 by 10 in this
test because, before this task, the standard price was updated when
running fifo, even in dropshipped. As the PO line was valued at 8, the
anglos axon in stock out was at 8 because the standard price was changed
to 8 after the delivery.

task-1875873

[REF] stock_landed_costs: adapt to valuation layers

We need to increment the value of another layer. We don't want to write
on this layer in place as we used to do for multiple reasons:

• valuation at date: before the application of an lc, the value should
  not change
• if the layer was consumed, we would need to adapt the consumers and
  if one day we would have needed to understand what happened in the
  fifo stack it would be hell

So we allow to create a valuation layer linked to another valuation
layer through a simple M2O O2M. At the time it is only used on the
landed costs to add value and no quantity so the code expects it would
be the case and it is not enforced. We adapt _run_fifo to use the full
value.

We also add a stat button on the landed cost to access the valuation
layers, like what is done for the picking and manufacturing orders.

task-1875873