Purpose of this commit is to simplify fields used in leave and allocation
used to store the duration of the requests. We want notably to get rid
of "number of days temp" that, contrary to its name, is not temporary
at all as it stores the real number of days. We also want to clarify
the hours or days encoding that is different on leave and allocation models.
Specifications
* number_of_days on leave and allocation is the master field holding the
leave or allocation duration in days;
* two fields are used for interface that are suffixed by _display in order
to indicate they are UX only: number_of_hours_display and number_of_days_
display. They are interface fields used for the user and correct number
of days is computed and stored in number_of_days;
* more specifically for leave model
* display fields are readonly as it is computed based on dates given by
the user, directly (from / to) or indirectly (half-day leaves for example);
* duration field is computed through an onchange. It could be directly
set by hand by managers if some confusion in the number of days appear;
* more specifically for allocation model
* display fields trigger an inverse as they are the real duration given
by HR officers or managers;
* duration field is computed through the inverse of display fields;
Various reports, data and use of leave / allocation models are updated to
use the correct field.
This commit is linked to task ID 1876795 and PR #26571.
Purpose of this commit is to simplify the configuration of leave type and
use the job of HR officers / manager by making options clearer and easier
to use.
Specifications for leave type :
* display name as header;
* remove accrual option from leave type. It is now configured directly
on allocation requests;
* add a "X days allocated" stmart button. If you set a general allocation
of 20 days, the stat button will display 20 days. Even if you go update
the allocation by employee after that. Purpose is to have the value of
main allocation done in group, not individual cases of some employees;
* add a "Group leaves" stmart button. It leads to a new leave requests
in group mode where mode is visible;
* have a unique allocation type field replacing the setting allowing to make
leave or allocation as well as limit and negative fields;
* limit leave duration encoding to two options, either day-based (allowing to
take full days or half a day), either hours-based (allowing to take full
days, half a day or some specific hours in a given day);
More technically
* limit and negative fields are removed. It was confusing as limit set to
True means you can actually exceed the limit). It can be computed based
on allocation_type ('no' meaning you can bypass the limit);
* request_unit is simplified. Day and half are merged into day;
* accrual and balance limit are moved to allocation model;
Specifications for allocation requests
* set accrual specific on allocation itself and not coming from the leave
type configuration anymore. Indeed we now consider accrual to be done
case by case independently from the leave type;
* remove negative and balance_limit. Indeed we now consider people have to
check by hand what is done by accrual;
Leave and allocation are slightly impacted by other changes, such as domain
for default leave types. Changes have been made as small as possible. Future
commits will still improve leave and allocation models and views, going
further in the cleaning process. Some tests are udpated to match the new
configuration of leave type.
This commit is linked to task ID 1876795 and PR #26571.
Demo data prevent from reinitializing hr_holidays module. Indeed confirming
already confirmed leaves or allocations does not work. Demo data are
updated so that we can initialize hr_holidays several times.
This commit is linked to task ID 1876795 and PR #26423.
We need to support the accrual allocation for leaves.
This is a special type of allocations that adds some leaves to the
allocation every X amount of time worked.
This kind of allocations is especially used in the USA, but can also be
used in other contexts (eg: In Belgium we could allocate 20 days per
year for the legal leaves)
These allocations have support for :
* A max balance limit
* Going in negative (if set on the leave_type and the manager
approves the leaves)
* Removing the unpaid leaves from the quantity of days worked
* Setting an end date after which it does not allocate anymore leaves
Task #46114Closes#22292
We would like to be able to take leaves using different units, currently
we only support taking leaves by hours because we use datetimes.
We need to support three units:
* days
* half days
* hours
These units will be defined on the leave type.
Task #40995Closes#21760
This commit improves leave request and allocation management through a
better integration of activities and addition of automated activities.
Several things are done in this commit :
* leave and allocation models do not inherit from mail.activity.mixin.
This commit adds the inherit so that HR users and managers can now
schedule and manage activities on leave and allocation requests. This
will help them in their daily job;
* automatic activities generation is added for leave and allocation.
Activities are generated for approval and second approval. They are also
automatically set as done when validating or unlinked when refusing or
resetting to avoid bloating users with unnecessary activities;
* a menu to configure activity types is added. Indeed HR managers should be
able to see and configure activity types related to their job;
* filters are added to be able to use the systray and to filter the kanban
view based on activities
Having automated activities allow to replace some messages and tracking
that were implemented to warn people of leave and allocation to approve
or validate. This commit therefore
* simplifies the tracking as only approved or refused state now trigger
a subtype;
* removes to approve and to validate subtypes on leave and allocation
model as well as their parent subtypes on the department. This allows
to simplify chatter a lot in hr_holidays app;
It is simpler to have data split by main model or application like we
already do for models and views. It allows to easily have an overview
of data a module holds.
As we will work on mail related data like adding activities or tweaking
subtypes and templates, having them all in a single file and not lost
between other data helps finding and working with it.
This commit only moves code. No functional change should occur.
In order to help limit leaves recording errors. We need to add some
configuration on leave type object. For example we need to avoid people
taking legal leaves 2016 on 2017.
* Add validity dates : start and end dates on the leave type. An empty
end date means forever
* Double validation : change the checkbox into a radio button with now
three choices :
- Human Resource Responsible
- Manager
- Double Validation
* Add a sequence field, using handle widget
* Add a field, to specify if the employee can select the leave type in
his allocations, requests or in both of them
Purpose
=======
Follow the coding guidelines (https://www.odoo.com/documentation/11.0/reference/guidelines.html)
Specification
=============
* each model has now its own .py file
* view files have better names
* reordered the methods for readability
* reimplemented email subtypes
Purpose
=======
Have a clear distinction between leave and allocation requests instead of
using the same model to mix 2 different concepts.
Splitting the model will allow different business code to be run on each
model as allocations and leaves are not exactly the same thing; this
will add some code but simplify future improvements
Specification
=============
1/ Completely separate the models
Model hr.holidays has been split into leave.request and
leave.allocation.
2/ Make reporting working again by using an SQL view aggregating data
from leaves and allocations
A new report has been added in order to aggregate
datas from both allocations and requests. The views have been modified
accordingly.
If an user (or a developer) wants to install this module twice, she
could not before this patch. Because those demo data call a function
that crashes if the state of the demo data has been modified... and
modifies it!
We avoid this issue by resetting the states of the demo data
explicitely (before a deeper refactoring)
addendum to commit da6b9d1be443999028f021ff81b4cf128a4eaaa2
- better employee name and image for default employee in data (xml-id: hr.employee_root)
- in demo data hr.employee_fp renamed to hr.employee_root to not have 2 employees linked to the root user
- renamed all occurences of employee_fp to employee_root
- added encoding to xml files
- replace search_read() with search() for _search_absent_employee()
- fixed setting first date of current month as default value for wizards
- reduce calls to <dict>.get()
Christophe Simonis