In discourse/discourse#32618 we added the dependency on locales to the basic topic serializer gated behind SiteSetting.experimental_content_localization. This is so that fancy_titles can be rendered in the language of the user.
Error: `ActiveModel::MissingAttributeError (missing attribute 'locale' for Topic)`
This commit adds that dependency to extra_data_pluck_fields which lists fields required from the serializer.
This change updates the report generator script to handle groups first, that way when the option users_from_group: true is used we can determine query access based on the user rather than the whole group.
This change adds an option to the Schedule a PM with Data Explorer results script, which when selected will send individual PMs to each group member selected in the Send to User, Group or Email field.
When this field is not checked, the functionality for groups will work as it did before (ie. send a single PM to the group).
When setting up an automation to create a DM or a post with a report on a recurring basis, and using a Data Explorer query that has parameters, we encounter an error.
The `.params_to_hash` method currently expects an array of arrays for the parameters, but in reality they seem to be arrays of hashes.
This change makes `.params_to_hash` work with arrays of hashes. It also preserves the ability to work with nested arrays in case this is used somewhere else.
The code for listing all of the defined queries is mixed together with the code for editing a single query. Notably, this results in large amounts of unnecessary data being loaded for the list view, which causes substantial rendering slowdowns.
To address this issue, we now only load the necessary data for the list view, and load the full data when it's actually needed (any endpoint that returns a single query). The primary changes that achieve this are:
- Create a new `QueryDetailsSerializer` serialiser, which includes all of the query info, and change the existing `QuerySerializer` serialiser to only include the necessary attributes of each query for generating a list of them all.
- Split the monolith `/plugins/explorer` route into `/plugins/explorer` for showing just the list of queries, and `/plugins/explorer/queries/:query_id`, for showing/editing/running a specific query.
When reading an existing state from the params defaults or from URL params, the input elements weren't representing the right state for booleans and "3-state/null booleans".
internal /t/-/118495
When checking `attach_csv` in the `Schedule a PM with Data Explorer results`
automation script, `csv` has to be added to the list of authorized extensions in the site settings. This will now raise an error if it's not the case.
Followup da1c99ec2d
We already had the functionality to convert results like
this:
```
|blah_url|
|--------|
|3,https://test.com|
```
To a link clientside, so in the UI it looks like this:
```
<a href="https://test.com">3</a>
```
With the addition of the recurring report to post automation,
and the existing report to PM automation, we also need to be
able to do this server-side, so reports don't come out malformed
if they have these type of count + URL columns.
This script is similar to the existing one that schedules
report results to to be sent to a PM on a regular basis,
but instead takes a topic ID and posts to that. This way
people can have report results sent to a public topic regularly
too and not have to deal with PM recipients and so on.
This commit uses GroupChooser as the input for param input of type
group_id. Meanwhile, it improves invalid group validation and semantic
error prompts.
The old float validation had several bugs. It will recognize strings
like "a1.2" and "3.4b" as valid doubles, but will not recognize integers
like "1234" as doubles. Also, since an empty string is not falsy in Ruby,
it will recognize "Inf" as -Infinity.
This commit fixes these issues
Validation of `user_id` parameter will throw a 500 error because
`User.find_by_username_or_email` does not throw
`ActiveRecord::RecordNotFound`, but silently returns `nil`.
This results in a `NoMethodError` in `object.id` on the next line
What's the problem?
===================
TL;DR: When the user enters an incorrect default value, its
corresponding param-input will disappear
When creating a parameter from SQL, we perform cast_to_ruby, which means
that if the default value given by the user is invalid, the entire
parameter will not be added to the param_list of the query.
This behavior is very confusing, and users will not understand why an
incorrect initial value will cause the param-input to disappear.
What's the fix?
================
The cast_to_ruby process is canceled in create_from_sql, so that
param-input with incorrect default value will still be displayed.
We have a simple validation process on the front end, which is enough to
prompt whether some default inputs are incorrect.
We used `with_deleted` incorrectly in the code. This method does not exist
on `User`, `Badge`, and `Group`. This will cause an error when entering a
numeric id in these three parameter input, forcing the user to enter the
name/username of these inputs.
See https://github.com/discourse/discourse-data-explorer/pull/307#issuecomment-2291017256
* FIX: Wrong type in category_id param input
We will dasherize category_id. The dasherize function accepts a string,
but we don't type-check it, so the default null may be passed in. This
will cause a type error and crash the front-end.
The main change here is that we now send a single PM to a group rather than individual PMs to each group member.
This change also adds email recipients correctly as target_emails since support was added within Discourse Automation.
* Added can_see_bookmarkable? methods to BaseBookmarkable, implement
that in QueryGroupBookmarkable
* Update spec to check that Notification for reminder has
bookmarkable_id and bookmarkable_type
c.f. https://github.com/discourse/discourse/pull/25905