The open_action_with_context was not using _for_xml_id, producing an
error when reading the action content.
In open_action, the action_name was taken from the context and needs
sanity checks. Ensure only actions from the account module can be
read and only if the user has access to the target model.
This is a limitation of the previous behaviour but, at the moment, all
known calls are made refering to an action from the account module.
Limit the scope of this method while the 14.0 is still early to avoid
having a door open to ready any action, and difficult to close later.
Remove old action fetching from the context in create_move that is no
longer used.
closesodoo/odoo#61602
X-original-commit: a39e94f7e4dc74e850650ac90bbc5f85af130bc8
Related: odoo/enterprise#14695
Signed-off-by: Martin Trigaux (mat) <mat@odoo.com>
Co-authored-by: Xavier Morel <xmo@odoo.com>
Use _for_xml_id to replace all the self.env.ref().read()[0]
This has the advantage of having a single point of control and to add
the fields filtering and model verification.
Add sudo for other operations on ir.actions.*
Purpose
=======
This commit intends to improve the animations on the leaderboard during survey
sessions. We now show both the ranking of each attendee before and after taking
into account the current question scores.
It allows attendees to see their position in the leaderboard shifting up / down
depending on how well they answered, and adds a bit more interactivity during
the session.
Specs
=====
The code changes to make this happen implied having multiple layers of
animations running one after the other.
The use of async/await in the JS allowed for a much smoother and less indented
code, simplifying the reading.
We now have 2 different "bars" to handle:
The bar that shows the previous leaderboard position.
The bar that shows the current question score.
This means we handle both jQuery items and modify their size & transition
properties depending on our animation.
Unfortunately, all these animations required a lot of testing to find fitting
timings, which means using quite a few setTimeout with "magic" delays.
These were just the most good looking ones and it feels like naming them with
a constant would not really help.
We also slightly improved the graphic font size / styling to make it easier to
read for potentially far away attendees.
Task 2235815
closesodoo/odoo#50991
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
As there is no session object, it is not possible to generate a randomised set
of questions that will be displayed for all the clients.
When getting the most voted answers while checking which question is the next
one to display, or if this question is the last one, predefined questions are
regenarated each time a fake user input is created (the most voted answers.)
So at the end, while checking if a question is the last one, we compare
different objects.
This is not an issue for non-randomised survey, as everything is constant.
So, until a solution is found to keep information related to a session in
particular, randomised questions mode will be ignored in live session mode.
Task ID: 2230980
X-original-commit: 7f0981c4a61429802a5fe24dd794221b1275dfe3
Unanswered questions or sections are now hidden in the survey review mode.
Only the question answered by the user will be visible.
If the survey was done during a session, the users will see their own answers,
even if the theme they selected was not the most voted one (in case of
conditional questions).
To get the questions to display :
- questions that have been answered
- + the ones that should have been answered = active questions
In session, active questions are based on most voted answers)
This commit also fixes the randomized questions handling in most voted answers,
as copied predefined questions where <NewIds> and those was not usable in
some templates.
Task ID: 2230980
X-original-commit: 8ca79727d5a93b91746c1e63d59d08550417906c
Before this commit, is_last_page_or_question was determined only in function
of next active questions : If there are no active question after the current
one, current question is the last one.
But, in conditional questions context, if the current question is conditionning
the display of latter questions, when we display the current question, we
cannot already know if we are on the last question, as user did not answered
yet.
After this commit, the current question is the last one only if
- there are no active questions that follows
- AND the current question is not conditionning the display of other questions
For page_per_section, we check the same rule for all section's questions.
Task ID: 2232973
X-original-commit: 12d962e579d2480fbf821fb381adebf47668dbcc
As section with description are now considered as potential questions to
display, the method to get the next question to display needs to check if
the next section should be displayed or not, depending on conditional
configuration and user inputs.
For page_per_question and page_per_section layout, the following rule now
applies:
If the section is empty (has no sub-questions):
Show it if section has description, else skip it. (in both layout)
If the section has sub-questions:
If a least one sub-question is active:
In page per section: show it.
In page per question: show if section has description, else skip it.
If no active sub-question: Hide it.
Apply also empty section rules to non conditional survey :
If a section is empty (no description and no active questions inside), the section
should not be displayed.
Task ID: 2230980
X-original-commit: f0ce3623e3cfc2451e3f7359ec09399e6e474c10
Before this commit, when computing the result graph data in the survey finished
page, a user answer for a question with no correct answer was counted twice,
once as correct, once as incorrect, as :
correct_answer == question_correct_answer == none
After this commit, we count an answer as correct only if question has a correct
answer.
Task ID: 2176630
closesodoo/odoo#54755
X-original-commit: 74a00111908556bce10ea3a5e748f5a5c81f6a40
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
This commit fixes all issues detected by the new pylint
gettext-variable test.
It converts some calls to the new syntax
_("Foo %s", bar)
to progressively migrate the code to the new syntax.
A few calls were not technically incorrect but still detected by the
linter.
_("Foo" +
"Bar")
has been converted to
_("Foo"
"Bar")
as it has the same effect and make sure the argument is of type
asteroid.Const instead of BinOp).
closesodoo/odoo#53683
Related: odoo/enterprise#11467
Signed-off-by: Raphael Collet (rco) <rco@openerp.com>
PURPOSE
Try to move from onchange / default_get to stored editable computed fields.
Behavior should be the same (computed or set by user), with support of
create / write / onchange field update without additional code.
SPECIFICATIONS
Update classic fields updated in some cases by onchange and/or default methods
by fields with store=True, readonly=False. It means their value comes either
from manual user input, either from trigger based computation.
Remove onchange and default_get when possible, leading to an unique computation
method and clearing fields definition.
Also clean some fields definition inconsistencies, notably required fields
that should instead be correctly computed or default that have no real meaning.
SURVEY USE CASES
- On survey.question, the question_type field should be emptied if the question
is a page (is_page = True)
- On survey.survey, if scoring is configured, the minimum scoring success
should be between 0 and 100 (added SQL constraint).
- On survey.survey, the certification field cannot be True if we're in no
scoring mode (scoring_type = 'no_scoring')
- On survey.survey, the attempts number cannot be limited (is_attempts_limited
is False) if the access mode is public and the login is not required.
- On survey.survey, we cannot give a badge (certification_give_badge is False)
if it's not a certification or if the login is not required.
- On survey.invite, the "subject" and "body" field are based on the template
used (template_id field).
MISC
- As "validation_email" works with "validation_required", we removed the
onchange on "validation_email".
- To avoid extra complexity with an onchange on "is_time_limited" and
"time_limit", we simply default the time_limit field to 10.
LINKS
Upgrade PR odoo/upgrade#847
Task 2187377
closesodoo/odoo#46291
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
- Create a new survey 'Test Survey';
- Add a section 'S1';
- Add a multiple lines text box question 'Q1';
- Add a section 'S2';
- Add a section 'S3';
- Don't move any of the section or question to avoid changing the
sequences, for the moment all the sequences are equal to 10;
- Change the survey layout to 'One page per section';
- Save the survey;
- Test the survey;
- Fill the 'Q1';
- Go to the last page and Submit the survey;
- Review your answers.
Before this commit, the question is empty, this occurs because the
question page_id is not set, in _compute_page_id, the sequence of the
question should be bigger than the question of the page (in this case
the section).
Now, the _compute_page_id was change to take into account the case when
the question has the same sequence as the page (section).
opw-2222045
closesodoo/odoo#49366
X-original-commit: 99f35075e3d1e418e096a29b7440152e746a1cd5
Signed-off-by: Jorge Pinna Puissant (jpp) <jpp@odoo.com>
Co-authored-by: Nicolas Lempereur <nle@odoo.com>
This commit fixes timer related issues on survey flows, for both regulars
surveys and survey sessions.
The timer was unreliable in its completion by a whole second, which had not a
big impact at the user level but could make some tests fail (randomly).
Side changes from this commit:
- Reduced delays in the survey session tests to speed up testing
- Removed some timer flickering on survey views
LINKS
Task 2230526
Runbot error ID 535
closesodoo/odoo#49178
X-original-commit: 15c72aec71d42624685fb0fc9df25279c7855310
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
On redirection, werkzeug converts locations from relative to absolute
URLs. Apparently the specific algorithm when given a relative *path*
changed between 0.14 and 0.15, or possibly these paths suddenly became
processed whereas before they were interpreted as absolute?
So in Werkzeug 0.14, `redirect('foo/bar')` was treated as
`redirect('/foo/bar')` but in 0.15 it's first urljoin-ed to the
current path before the redirection is effected.
This is an issue with slides, because survey's get_start_url returns
the url `survey/start/<access_token>` without a leading slash.
The result is that where with 0.14 the certification tour would do this:
GET /slides_survey/slide/get_certification_url?slide_id=39&fullscreen=1
GET /survey/start/<id>?answer_token=<token>
with 0.15 or above it does this:
GET /slides_survey/slide/get_certification_url?slide_id=39&fullscreen=1
GET /slides_survey/slide/survey/start/<id>?answer_token=<token>
Which weirdly enough doesn't work at all.
closesodoo/odoo#48982
X-original-commit: 82a23883748172c05d2562f244ffbc1558ce01fc
Signed-off-by: Xavier Morel (xmo) <xmo@odoo.com>
Purpose of this commit is to synchronize active status of certifications to
their badges. When archiving or activating a certification, its badge is also
archived or activated.
Task ID 2170708
Community PR odoo/odoo#46563
Purpose of this commit is to call toggle_active or archive / unarchive methods
instead of manually writing on active field. Indeed this allows to trigger
business code related to archive / unarchive which is normally located
in toggle_archive (called by action_archive and action_unarchive).
Task ID 2170708
Community PR odoo/odoo#46563
Purpose
=======
Speed up survey completion.
Specifications
==============
This commit apply an auto commit when user select a choice on a simple choice
question or on a matrix question with simple choice.
The auto commit happens following those conditions:
- Only on simple choice and matrix with simple choice questions
- For matrix, only when last subquestion is answered
- Only when Back mode is active
- If this is not the last question
Link
====
PR #46768
Task 2208574
This commit slightly changes the way users navigate through the survey form:
- The back button was moved to the bottom-right of the screen.
- The "Continue" button stays at the center but there is also a "next" button
at the bottom-right of the screen (same behavior -> submits the form)
- Arrow keys (left + right) navigation is now supported
Technical note: As the survey navigation is now outside of the survey content,
that was re-rendered on every submit, we have to manually refresh the
navigation with a second template render on the controller side.
LINKS
PR #46768
Task 2208574
PURPOSE
Improve the various sessions flow to make it easier for attendees to reach the
session and to allow better interactions between the host and his audience.
This comes with small usability improvements as well as increased
"beautifulness" by using custom layouts & animations.
SPECS
This commit changes the survey sessions fade in/out delay for both the host
and the attendees.
When going from one question to another, the total delay is now 1 second
(instead of 2 previously), with 500 ms for fade out and 500 ms for fade in.
The previous delay gave an impression of "slowness".
The delays and "server lag reduction" on the attendees side were adapted
accordingly.
LINKS
PR #46768
Task 2208574
PURPOSE
Improve the various sessions flow to make it easier for attendees to reach the
session and to allow better interactions between the host and his audience.
This comes with small usability improvements as well as increased
"beautifulness" by using custom layouts & animations.
SPECS
- For live session only:
Currently, the "short link" feature is good but it still needs to be improved
with the following points:
- Make the default "code" shorter (4) and digit only.
- Allow to customize the "code"
- Check for potential collisions
Added at the same time:
- input element is focused when you land on the '/s' page.
LINKS
PR #46768
Task 2208574
== Fix results scored question display ==
This commit fixes the display for scored choice question.
The text "Right answer" was displayed twice.
== Add default values to avoid constraints inconsistencies ==
This commit adds a 0 default values to some of our validation fields on survey
questions.
Without a default value, the ORM would store "NULL" in the fields and then
allow users to define a "min" without a "max", that would result in an error
while trying to validate the answers to that question.
Furthermore, when creating a question from the frontend, the web framework will
always send 0 for all field values anyway.
The code change adds consistency and avoids potential errors on that issue.
LINKS
Task 2208574
closesodoo/odoo#48371
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
With this commit, Selection fields with `required=True` which are
extended via `selection_add` are given proper ondelete policies to
ensure the cleanup of records containing these extended options during
uninstall of the extending module.
This commit also cleans up leftover uninstall hooks that were being used
to handle the same set of problems prior to the ondelete mechanism being
implemented for Selection fields.
closesodoo/odoo#46325
Related: odoo/enterprise#9117
Signed-off-by: Raphael Collet (rco) <rco@openerp.com>
This commit applies conditional questions management on the
"_get_next_page_or_question" method in "page_per_section" questions layout.
We choose the first section that has at least one active question when
searching the next section to display.
Indeed, a whole section could contain only disabled questions based on the
previously selected answers of the attendee.
(The logic is obviously reverted when "go_back" is True).
PR #48442
Task ID 2208574
This commit fixes 2 small problems when the user goes back on the survey form:
- The back button should not be shown on "page_per_question" layout if we are
on the first page of the survey.
Pages can be displayed on "page_per_question" layout if they have a
description since #44271
- Can't go back on "page_per_section" layout if you are on the second page of
the survey, introduced by 0675d68afb
LINKS
PR #48442
Task 2208574
We now have a way to display "free text sections" during the survey flow.
(Since 5b74f54396 )
However, the behavior during survey sessions was not optimal:
- Attendees could click on "submit" and it would raise an error
- The section description was not displayed on the host screen
- The section description was never the "current question" of the session of it
was the first item of the question_and_page_ids
This commit fixes all those points to have a nice session flow.
While the host is explaining the section description to the attendees, they get
a waiting message on their screen.
In addition, we fixed the display to avoid showing pages that lack a
description, and we reworked the '_get_pages_or_questions' method to have more
meaningful and readable cases for every layouts.
LINKS
PR #48442
Task 2208574
Under a survey's results (i.e. summary of all user submissions) the
question type 'numerical_box' has a "Most Common" tab of its results
displaying the top 5 survey user answers. This commit adds the same tab
to question type 'date'. It is expected this tab will be most useful for
'date' questions that are scored.
Implementation was left generic to allow for other question types to
easily adopt similar display. For now, question types: 'text_box' and
'char_box' results do not have a "Most Common" tab since this currently
adds no benefit (once scoring is implemented for 'char_box', then there
may be a benefit for it). 'datetime' is also expected to not benefit
from this additional tab and requires a more complicated implementation
to handle timezone display consistently between tabs and potential
"Right answer" (see related task id: 2169184).
Upgrade PR: odoo/upgrade#835
Task ID: 2061937
Previously only the question_types 'simple_choice' and 'multiple_choice' were
scoreable and could have "correct" answer(s).
In this commit we change this behavior so that question_types 'numerical_box',
'date', and 'datetime' can also have a "correct" answer and an associated
answer_score.
Task ID: 2061937
Community PR #43014
Upgrade PR odoo/upgrade#835
The purpose of this task is to allow the user to create pages which describe
the next survey questions. He can now edit the description in a form when he
clicks on the info icon when he creates or edits a page/section.
Page will be displayed on the frontend in all cases. This allows notably
to have pages with some text / comment for presenters when doing live
action surveys.
Task 2182886
Pr #44271
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
Purpose
=======
This commit makes the live session mode work with conditional question.
The point is to determine which question to display next. As multiple users
are 'voting' for answers of the same questions, the way to determine the
next question to display is to keep only the most voted answer.
The most voted answer will be the only one to trigger the linked conditional
question.
E.g.: considering
- following survey configuration:
- Q1: A1, A2
- Q2: triggered by Q1-A1
- Q3: triggered by Q1-A2
- following user inputs:
- U1: A1 selected
- U2: A2 selected
- U3: A2 selected
The next question displayed to ALL users will be Q3.
Note: If U4 had selected A1, the next question would have been Q2 as
when there is an equality, we select the first one.
Link
====
Task ID: 23509
PR #44950
Purpose
=======
This commit adds a new survey mode : conditional questions. A conditional
question is a question that will be displayed to the user only if the user
selects a specific answer from another previous question i.e. question B is
displayed if user answers 'Choice 1' at question A.
Specifications
==============
There are 2 different ways to handle conditional questions.
* For one_page and page_per_section survey layout: conditional management is
done mainly at frontend side. The idea is to be able to load all the
questions of the section, hide the conditional questions and display them
dynamically on answer check.
E.g.: If user checks answer 'Choice 1' from question A, question B is
displayed immediately.
If question B is on section 2 and survey layout is set to section per page,
question B will be displayed immediately on next page template load.
Indeed, on each page load, server sends all the selected answers and the
template displays the conditional question depending on the already
elected answers;
* For page_per_question survey layout: conditional management is done mainly
at backend side. The idea is to find on each page the next question to
display depending on:
* navigation direction (next - previous)
* already selected answers and conditional question dependance
* survey questions sequence
E.g.: If questions survey are Question A, B and C and question B must be
displayed if answer 'Choice 1' from question A is selected.
At Question A, if user select 'Choice 1' and goes next: display B
At Question B, no matter the answer, if user goes next: display C
At Question B, no matter the answer, if user goes prev: display A
At Question A, if user select 'Choice 2' and goes next: display C
At Question C, no matter the answer, if user goes next: End survey
At Question C, if user has selected 'Choice 1' at question A and goes prev -> B
At Question C, if user has selected 'Choice 2' at question A and goes prev -> A
In both modes answers on questions that should not have been displayed to the
user depending on the previous selected answers are cleared.
E.g. : User checks 'choice 1' from question A, then complete question B.
Then user unchecks 'choice 1' -> answer on question B is cleared.
Ignore conditional configuration :
* during survey session, to avoid having different users answering different
questions on the same session. That would be confusing for users to see
questions on presenter screen that they didn't have the opportunity to
answer;
* if survey questions selection is random, because it will not display
conditional questions if they are not kept by randomization. To avoid
confusion, conditional configuration is simply ignored, not unset, so the
user can still reactivate it without having to reconfigure everything.
Remove questions that should not have been displayed to the user (depending of
the made choices) in scoring computation and print layout.
--> Inactives questions are removed from predefined questions when user finish
the survey.
E.g.: If the user selects first path on 2, he will never see second path.
If the user answers correctly to all question of first path, he should have
100%, instead of 50%, even if questions of path 2 are mandatory and scored.
Link
====
Task ID: 23509
PR #44950
This commit improves the answers count in the context of survey sessions as
well as displayed answers to narrow them down to the "current session".
We simply use a custom domain to consider all user_inputs and user_input_lines
that have a create_date after the start of the session.
It will help users to be able to re-use the same survey multiple times and run
different sessions without having to duplicate the survey every time.
LINKS
PR #45111
Task 2191924
PURPOSE
Improve layouts and usability for the host screens when using survey sessions.
We focus on the "host" screens, as the attendee screen is based on regular
survey flows/design.
SPECS
In short: copy mentimeter "wow effect".
Main points:
- Remove the "question" container (keep only "Results/Ranking")
- Remove the tabs concept, views are displayed in full screen
- Purpose is to have a "clean view"
- The display must be "full screen" with only relevant info
- Enable a easy keyboard navigation with left/right keyboard arrows
and spacebar
- Switching from results to ranking to next question is always linear
(next/previous instead of tabs)
- Improve charts
- Remove the pie chart : make bar charts instead for simple / multiple
choices
- Show the chart results "as people answer to it"
Only if question is not scored, otherwise show when all answers are
received / when user goes to "next"
- For free text/date/datetime: show small bubbles with attendees answers
- Add resize, fade in/out, ... animations for "wow effect"
LINKS
PR #45111
Upgrade PR odoo/upgrade#774
Task 2191924
The computed field 'session_show_ranking' was only computed based on the fact
that the survey had a scoring mechanism.
It's not very accurate since we also need a question that saves as the attendee
nickname to be able to display a relevant ranking chart.
This commit fixes the compute method by adding the necessary condition.
Task ID 2188482
closesodoo/odoo#45852
X-original-commit: 86ca11517d3872b2bf3cf89eda513a3bc8c84292
Signed-off-by: awa-odoo <awa-odoo@users.noreply.github.com>
PURPOSE
Add a "live session" mode for survey that allows the host to interact with its
audience.
He controls the pace of the survey and the audience answers questions one at a
time.
Results for each question can be displayed by the host to adapt his speech in
real time.
A gamification component even allows to give points to the attendees, based on
the quickness of their answer, and create ranking to keep everyone's attention.
SPECIFICATIONS
This whole feature is deeply integrated with survey.
It adds a layer of "survey session" on top if it that gathers attendees and
answers during the lifespan of the session.
The attendee identity is filled in by a special question that is marked as
"Save as user nickname".
If the host has not configured that question or if the user doesn't answer,
they are marked as "Anonymous" in the rankings.
The answer score computation is also altered during the session to take the
speed of the answer into account.
To ease the access to the survey to attendees, we have introduced two new
routes in the survey main controller:
- '/s/123456'
That uses the 6 first characters of the token to give quick access to the
survey and allow typing the URL manually somewhat conveniently.
- '/s' that renders a view that asks the attendee to enter the survey 'code',
which then redirects to the previous route.
This commit also fixes a few bugs:
- A bugged use case with the "enter key" listener.
When the survey is done, we display a "result template" that shows the user
his score and allows him to retry if possible.
When this template is displayed, we don't want to listen to the "enter" key
that allows to submit the survey form.
- We now correctly remove the timer and initialize the result
widget when the user presses 'enter' on the last page / question.
SURVEY SESSION FLOW
HOST point of view:
- On the Survey Form, the host configures:
- "Reward quick answers" that gives more points to attendees if they answer
quickly
- Question Time Limit, on each question, that defines the time limit for
that specific question, allowing to have different (or no) timer for each
question.
- From the Survey Form, the host can start a "live session"
Only one session per survey can be running at a time.
- When he starts the session, the session_state is marked as 'ready', meaning
all attendees landing on the survey page will be part of the session.
- The host will land on a page showing the current number of attendees as well
as the link for the attendees to join that session.
- The host starts the session, activating the first question of this session.
We keep an active reference to the "current_question_id" and use the bus to
trigger an event that will refresh the survey page for all attendees, showing
them the question and allowing them to answer.
- The host gets a "question management screen", from which he can:
- See the current question text & suggested answers
- See the current question timer
- See the number of answers received for the question
- Display the answers of the question (same view as the survey "results"
page, but only for the question)
- Display the ranking of attendees (if "competitive mode" is enabled)
- The host controls the pace of the survey by moving to the next question until
it's the last one of the survey, then ends the current session.
When the session is closed, we mark the answers of the attendees as "done".
- The host lands on a screen where he sees the survey results and the ranking
of all attendees.
ATTENDEE point of view:
- He reaches the survey when a session and open, and gets a screen asking him
to wait until the host decides to start the session.
- When the sessions starts, he's automatically redirected to the first question
(see above).
- If the session is already started when he reaches the link, he lands on the
current question (and NOT on the first one)
- If the host has configured a time limit, the attendee sees the countdown
while he's answering.
- When the attendee answers the question, the screen tells him the answer is
registered and he has to wait for the host to go to the next question, which
will happen automatically.
- If the timer reaches 0 and he has not submitted his answer, it's too late and
he waits for the next one.
- It continues like that until the end of the session.
- At the end of the session, the attendee gets a screen with his global results
(same as a regular scored survey).
PR #43568
Task 1972640
Co-authored-by: David Beguin <dbe@odoo.com>
This commit fixes a potential division by 0 error by correctly applying the
default division by 1.
Triggered by task ID 1972640 (live sessions on survey)
PR #44205
As access to survey description field is limited to survey users and managers
we can let them put whatever they want in the description field. Otherwise
they cannot use the full power of web editor, notably embedded videos.
PR #44205
Because 0 evaluates as False in "if" statements, generalized answer_type
constraint check logic would not allow user to submit a survey with 0
as a numerical value. This commit fixes this issue by explicitly allowing
this case.
Triggered by task ID 2061937 (improve scoring in surveys)
PR #44205
- Purpose:
Allow the addition of media on a question and/or answers in order to describe
or illustrate them.
- Specifications:
During the creation of a question, you have now the possibility to set a
description of a question by filling an html field under the tab "Description".
This field can be used to enter text or to illustrate the question with an
image or video.
There is also an option to add pictures on answers for simple and multiple
choice questions. Those are displayed as clickable images.
Task ID 2083823
PR #42228
In this commit, all the questions type have been redesigned to look more fancy.
(impacted question types : radio and checkboxes, text boxes (textarea and text
inputs, date and datetime inputs and matrix)
For radio and checkboxes, both design have been aligned to work the same way
(except checkboxes can still have more than one selected option).
Selection by key have been applied on those two question types if the number of
option is under 26 (to use all the alphabeat character for selection). This
selection by key is only available on page per question layout.
Tests have been adapted accordingly to the redesign (typically for choice and
matrix inputs)
* Progress Bar
This commit adds a progress bar to the survey to inform the user
where he is in the survey filling process.
There are two progress modes:
- Number : will display the number of the current page on the total number
of pages (or questions if layout mode is question_per_page)
- Percentage : will display the percentage of page or question already done
So this leads to, on last page:
- in Number mode :
the progress bar div is 100% filled 3 / 3 pages [===]
- in percentage mode :
the progress bar is aligned to percentage 66% completed [== ]
* Print Widget
A widget is added for survey print mode in order to resize all textarea to fit
their content, instead of showing a scroll bar. This can be usefull if user
wants to print the results. He will get the entire content of the 'textarea
answers' instead of only the two first lines.
* Misc
This commit also :
- adds background image to survey.
- redo quizz correction and add some data to illustrate non scored
questions rendering in print template
- review breadcrumb style
Note : readonly data option on survey form widget is not set anywhere yet but
the usage is done in prevention of the future work on presenter view for survey
session mode.
Task ID: '2150291'
PR #43237
PURPOSE
To have a better UX, the transition between questions in a survey should not
be done by loading another page, but by loading next questions in AJAX.
This will allow to fade out / fade in the next question and improve the user
experience, as well as prepare the work for the complete survey redesign.
SPECIFICATIONS
General
=======
This commit refactors the way the questions are loaded.
To go to next question(s), instead of redirecting to a new page containing
the next question(s), the submit process is loading the next question(s)
using ajax. The template of the question is prepared at server side and is
given in html string to the client that display the next question in a smooth
transition (fade in / fade out).
This commit includes also the start and the end screen of the survey inside
the survey form in order to perform smooth transition between start screen
and first question(s) and between last question(s) and end screen.
Only one master template (survey form) is now used to display every part of
the survey : start + questions + end screens.
The answer_done error is now handled by the _prepare_survey_data itself.
If the user wants to display an already_done survey, the returned template
is the end screen.
This commit handles page transition for the survey results. When submitting
survey, the survey result widget is initialized and attached to the survey
result section.
A new route has been added : /survey/begin/.
If the user arrives on start screen, the state is 'New'. Once the user clicks
on Start button, an rpc call is made to survey/begin that will call the
survey_prepare_data to render the first question page and set the state to
in_progress and the start_datetime of the answer.
Validation
==========
Add the validation of all types of questions directly in the frontend.
This allows to reduce the latency before the survey is validated and was
necessary to go along with our new fade out / fade in mechanism.
The client does not need to wait the validation answer from the server to
display validation error messages.
The current questions are only faded out if the frontend validation has passed.
The server still makes this validation before submitting answers and loading
next question(s) to avoid direct rpc calls that would mess up the answers.
Error messages are now displayed with a slide down transition, for a wonderful
wow effect !
Timer
=====
As the survey form works with ajax to display start screen, question(s) and
end page, the timer has to be started only after the start screen.
This commit display and start the timer only after start survey button is
clicked.
Breadcrumb
==========
As the survey form now works in ajax to display in same page start screen,
question(s) and end page screen, the breadcrumb has to be controlled not only
on page loading but also manually between transition from a screen to another.
To ease the breadcrumb management, breadcrumb is now a widget that will reload
a js template, mainly depending of the current page id.
The js form initialize or update this widget giving this new page id.
On breadcrumb item click, the widget triggers a onclick caught by the form
that calls the submit with the given target.
This commit handles page transition for the breadcrumb. When going forward or
backward (next or previous page), the breadcrumb activates, deactivates and
links elements are reinitialized depending of the target page.
LINKS
Task ID: 2152223
PR #41453
Co-authored-by: Aurélien Warnon <awa@odoo.com>
This commit prepares the complete redesign of survey.
Simple choice question type will only use the radio button display mode.
So the dropdown display mode is deleted.
Task ID: 2152223
PR #41453
The matrix data structure awaited by the server is the following :
{
question_id: {
'row_id_1': ['col_id1', ..., 'col_idn'],
'row_id_n': ['col_id1', ..., 'col_idn'],
'comment': 'This is a comment',
}
}
Before this commit, the comment was encapsulated into a dict, into an array,
with an incorrect key :
-> 'questionId_comment': [{'comment': 'This is a comment'}]
This was leading to 2 errors :
- A traceback when nothing is selected in the matrix options
- Comment is never saved.
After this commit, the comment is correctly given to and extracted by the
server. So the comment is saved and the traceback is avoided.
Task ID: 2152223
closesodoo/odoo#43086
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
Co-authored-by: Aurélien Warnon <awa@odoo.com>
PURPOSE
Allow some questions to be linked to customer data. We would like to have an
email field filled with user info / updating user info automatically in the survey.
SPECIFICATION
For now, we only want to be able to collect the user email address for a specific
question of the survey. For that purpose, add a ``save_as_email`` field on the
question model.
If this option is checked, the answer to that specific question will override the
"email" stored on the survey.user_input allowing live udpate of user input
based on given information. If multiple questions are marked as such, the
last (not empty) answer will prevail.
If user input already has an email (invitation-based surveys for example)
prefill the answer with the email so that people don't have to retype it.
GOING FURTHER (aka, incoming)
With live mode: maybe some additional fields will need to be saved
nickname
avatar
LINKS
Task 1919290
PR #41429
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
Co-authored-by: Aurélien Warnon <awa@odoo.com>
Co-authored-by: Thibault Delavallée <tde@odoo.com>
From
"Their was an error during the validation of the survey."
To
"There was an error during the validation of the survey."
Some side code linting is also performed because my linter is screaming on
me.
Task ID 1919290
PR #41429
Co-Authored-By: Aurélien Warnon <awa@odoo.com>
Co-Authored-By: Thibault Delavallée <tde@odoo.com>
PURPOSE
As new features are about to land in survey, notably live interactions [1]
and new survey building [2] performing a pre cleaning is necessary. In this
PR we clean survey models by: removing unnecessary fields, cleaning some code
and finally renaming models.
SPECIFICATIONS
Purpose of this commit is to globally rewrite the way statistics are
pre computed and used in results page of surveys. Indeed current code is
very complex, having a lot of calls to the same methods while taking only
part of its results. It also have a mix of model and controller methods,
data is computed several times a bit differently, ...
In this commit we rewrite that part. Including
* a simpler code flow;
* a simpler data structure;
* use a better naming for methods instead of "prepare_results" called
randomly within the code and re-using part of its returned content;
* use record sets / records instead of giving sub-elements of a record
like question.title, label.value, ...
Data structure
* simply give a question_and_page_ids to the template that will iterate
on pages / questions;
* have a graph_data, table_data dictionaries available for graph and
table display with simple entries;
* graph_data: holding all necessary but minimalist data for graph display
if question allows it;
* table_data: holding all necessary but minimalist data for table / summary
display;
* hold record sets of done / failed answers and answer lines;
In this commit we also simplify and clean pagination in survey graph results
* remove page_range helper tool that can be inlined in the templates
directly, doing a simple range;
* fix and improve pagination code;
LINKS
[0] Related to Task ID 2061901 (survey models cleaning and preparation)
[1] Task ID 1972640 (live interactions)
[2] Task ID 2119587 (new frontend for building surveys)
PR #40765
PURPOSE
As new features are about to land in survey, notably live interactions [1]
and new survey building [2] performing a pre cleaning is necessary. In this
PR we clean survey models by: removing unnecessary fields, cleaning some code
and finally renaming models.
SPECIFICATIONS
In survey.user_input.line, clean code about constraints. Purpose is to make
it easier to read and understand.
Code creating user input lines is moved from survey.user_input.line model to
survey.user_input. Indeed it is now based on a user_input record, and create
answer line related to a question, based on values coming from controllers
(and therefore frontend). It makes more sense to delegate that code to the
user_input model itself that creates its lines. Code is also rewritten to
be cleaner and easier to read.
Code computing statistics on user input records is moved from survey model
to user_input model, where it belongs. Missing translation marks are added
because translations matter.
LINKS
[0] Related to Task ID 2061901 (survey models cleaning and preparation)
[1] Task ID 1972640 (live interactions)
[2] Task ID 2119587 (new frontend for building surveys)
PR #40765
PURPOSE
As new features are about to land in survey, notably live interactions [1]
and new survey building [2] performing a pre cleaning is necessary. In this
PR we clean survey models by: removing unnecessary fields, cleaning some code
and finally renaming models.
SPECIFICATIONS
To better understand question type and their input type, we update some of
``survey.question`` ``question_type`` keys :
Type----Old type-----New type
text----free_text----text_box
char----textbox------char_box
Untouched question types: ``numerical_box``, ``date``, ``datetime``,
``simple_choice``, ``multiple_choice``, ``matrix``. Those are already
understandable.
Then ``survey.user_input.line`` ``answer_type`` keys are also updated to
match their question type counterparts
QuestType--------Old type-----New type
text_box---------free_text----text_box
char_box---------text---------char_box
numerical_box----number-------numerical_box
Then ``survey.user_input.line`` fields used to store the value are updated to
propagate the new naming
AnswerLineType----Old field----------New field
text_box----------value_free_text----value_text_box
char_box----------value_text---------value_char_box
numerical_box-----value_number-------value_numerical_box
Untouched answer types and field storing value: ``date``, ``datetime`` still
refer to same question type and use value_date / value_datetime fields.
``simple_choice``, ``multiple_choice`` and ``matrix`` still use ``suggestion``
answer type and ``suggested_answer_id`` (+ ``matrix_row_id``) to store link
to ``survey.question.answer`` records.
LINKS
[0] Related to Task ID 2061901 (survey models cleaning and preparation)
[1] Task ID 1972640 (live interactions)
[2] Task ID 2119587 (new frontend for building surveys)
PR #40765
