Currently, there are two ways (kind of) for accessing `params` inside a
service:
- when there is no contract or it hasn’t been reached yet, `params` is
just the hash that was provided to the service. To access a key, you
have to use the bracket notation `params[:my_key]`.
- when there is a contract and it has been executed successfully,
`params` now references the contract and the attributes are accessible
using methods (`params.my_key`).
This patch unifies how `params` exposes its attributes. Now, even if
there is no contract at all in a service, `params` will expose its
attributes through methods, that way things are more consistent.
This patch also makes sure there is always a `params` object available
even when no `params` key is provided to the service (this allows a
contract to fail because its attributes are blank instead of having the
service raising an error because it doesn’t find `params` in its context).
This optimizes a hot path when users are being removed from one or more groups since we use the `on(:user_removed_from_group)` event to call the `Chat::AutoLeaveChannels` with a `user_id` parameter. In such case, we don't need to clear the membership of all users, just that one user.
DEBUG: Also added a "-- event = <event>" comment on top of the SQL queries used by the "AutoLeaveChannels" so they show up in the logs and hopefully facilitate any performance issues that might arise in the future.
Blocks allow BOTS to augment the capacities of a chat message. At the moment only one block is available: `actions`, accepting only one type of element: `button`.
<img width="708" alt="Screenshot 2024-11-15 at 19 14 02" src="https://github.com/user-attachments/assets/63f32a29-05b1-4f32-9edd-8d8e1007d705">
# Usage
```ruby
Chat::CreateMessage.call(
params: {
message: "Welcome!",
chat_channel_id: 2,
blocks: [
{
type: "actions",
elements: [
{ value: "foo", type: "button", text: { text: "How can I install themes?", type: "plain_text" } }
]
}
]
},
guardian: Discourse.system_user.guardian
)
```
# Documentation
## Blocks
### Actions
Holds interactive elements: button.
#### Fields
| Field | Type | Description | Required? |
|--------|--------|--------|--------|
| type | string | For an actions block, type is always `actions` | Yes |
| elements | array | An array of interactive elements, maximum 10 elements | Yes |
| block_id | string | An unique identifier for the block, will be generated if not specified. It has to be unique per message | No |
#### Example
```json
{
"type": "actions",
"block_id": "actions_1",
"elements": [...]
}
```
## Elements
### Button
#### Fields
| Field | Type | Description | Required? |
|--------|--------|--------|--------|
| type | string | For a button, type is always `button` | Yes |
| text | object | A text object holding the type and text. Max 75 characters | Yes |
| value | string | The value returned after the interaction has been validated. Maximum length is 2000 characters | No |
| style | string | Can be `primary` , `success` or `danger` | No |
| action_id | string | An unique identifier for the action, will be generated if not specified. It has to be unique per message | No |
#### Example
```json
{
"type": "actions",
"block_id": "actions_1",
"elements": [
{
"type": "button",
"text": {
"type": "plain_text",
"text": "Ok"
},
"value": "ok",
"action_id": "button_1"
}
]
}
```
## Interactions
When a user interactions with a button the following flow will happen:
- We send an interaction request to the server
- Server checks if the user can make this interaction
- If the user can make this interaction, the server will:
* `DiscourseEvent.trigger(:chat_message_interaction, interaction)`
* return a JSON document
```json
{
"interaction": {
"user": {
"id": 1,
"username": "j.jaffeux"
},
"channel": {
"id": 1,
"title": "Staff"
},
"message": {
"id": 1,
"text": "test",
"user_id": -1
},
"action": {
"text": {
"text": "How to install themes?",
"type": "plain_text"
},
"type": "button",
"value": "click_me_123",
"action_id": "bf4f30b9-de99-4959-b3f5-632a6a1add04"
}
}
}
```
* Fire a `appEvents.trigger("chat:message_interaction", interaction)`
Chat channels that are linked to a category can be set to automatically join users.
This is handled by subscribing to the following events
- group_destroyed
- user_seen
- user_confirmed_email
- user_added_to_group
- user_removed_from_group
- category_updated
- site_setting_changed (for `chat_allowed_groups`)
As well as a
- hourly background job (`AutoJoinUsers`)
- `CreateCategoryChannel` service
- `UpdateChannel` service
There was however two issues with the current implementation
1. We were triggering a lot of background jobs, mostly because it was decided to batch to auto join/leave into groups of 1000 users, adding a lot of stress to the system
2. We had one "class" (a service or a background job) per "event" and all of them had slightly different ways to select users to join/leave, making it hard to keep everything in sync
This PR "simply" adds two new servicesL `AutoJoinChannels` and `AutoLeaveChannels` that takes care, in an efficient way, of all the cases when users might automatically join a leave a chat channel.
Every other changes come from the fact that we're now always calling either one of those services, depending on the event that happened.
In the making of these classes, a few bugs were encountered and fixed, notably
- A user is only ever able to access chat channels if and only if they're part of a group listed in the `chat_allowed_group` site setting
- A category that has no associated "category groups" is only accessible to staff members (and not "Everyone")
- A silenced user should not be able to automatically join channels
- We should not attempt to automatically join users to deleted chat channels
- There is no need to automatically join users to chat channels that have already more than `max_chat_auto_joined_users` users
Internal - t/135259 & t/70607
* DEV: add specs for auto join/leave channels services
* DEV: less hacky specs
* DEV: no instance variables in specs
There's no UI for it at the moment but when creating a channel or updating it, it's now possible to pass `icon_upload_id` as param. This will be available on the channel as `icon_upload_url`.
* DEV: join/leave presence chat-reply when streaming
This commit ensures that starting/stopping a chat message with the streaming option will automatically make the creator of the message as present in the chat-reply channel.
* implements start/stop reply
* not needed
When adding threads to DM channels in #29170 we intentionally didn't add them to the My Threads section. However this makes it easy to miss notifications as we don't get the new thread badge on the sidebar and footer tabs (drawer/mobile). However they were also missing from the chat header and sidebar too, which is fixed with this PR.
When a new thread or a reply to an existing thread is created within a DM channel (either 1:1 or group), we now show the standard badges like we do for public channels.
We now also show the green dot in the sidebar for My Threads and public channels when they contain an unread watched thread.
We decided to make contracts immutable once their validations have run.
Indeed, it doesn’t make a lot of sense to modify a contract value
outside the contract itself.
If processing is needed, then it should happen inside the contract
itself.
This patch replaces the parameters provided to a service through
`params` by the contract object.
That way, it allows better consistency when accessing input params. For
example, if you have a service without a contract, to access a
parameter, you need to use `params[:my_parameter]`. But with a contract,
you do this through `contract.my_parameter`. Now, with this patch,
you’ll be able to access it through `params.my_parameter` or
`params[:my_parameter]`.
Some methods have been added to the contract object to better mimic a
Hash. That way, when accessing/using `params`, you don’t have to think
too much about it:
- `params.my_key` is also accessible through `params[:my_key]`.
- `params.my_key = value` can also be done through `params[:my_key] =
value`.
- `#slice` and `#merge` are available.
- `#to_hash` has been implemented, so the contract object will be
automatically cast as a hash by Ruby depending on the context. For
example, with an AR model, you can do this: `user.update(**params)`.
Currently in services, we don’t make a distinction between input
parameters, options and dependencies.
This can lead to user input modifying the service behavior, whereas it
was not the developer intention.
This patch addresses the issue by changing how data is provided to
services:
- `params` is now used to hold all data coming from outside (typically
user input from a controller) and a contract will take its values from
`params`.
- `options` is a new key to provide options to a service. This typically
allows changing a service behavior at runtime. It is, of course,
totally optional.
- `dependencies` is actually anything else provided to the service (like
`guardian`) and available directly from the context object.
The `service_params` helper in controllers has been updated to reflect
those changes, so most of the existing services didn’t need specific
changes.
The options block has the same DSL as contracts, as it’s also based on
`ActiveModel`. There aren’t any validations, though. Here’s an example:
```ruby
options do
attribute :allow_changing_hidden, :boolean, default: false
end
```
And here’s an example of how to call a service with the new keys:
```ruby
MyService.call(params: { key1: value1, … }, options: { my_option: true }, guardian:, …)
```
This patch improves the custom `array` type available in contracts.
It’s now able to split strings on `|` on top of `,`, and to be more
consistent, it also tries to cast the resulting items to integers.
On the chat channel settings page, we want to show a single Send push notifications setting instead of the current Desktop notifications and Mobile push notifications settings.
For existing users, use the Mobile push notifications setting value for the new Send push notifications setting.
While using `OpenStruct` is nice, it’s generally not a very good idea as
it usually leads to performance problems.
The `OpenStruct` source code even says basically to avoid it.
Since the context object is crucial in our services, this patch replaces
`OpenStruct` with a custom implementation instead.
Currently in services, the `contract` step is only used to define where
the contract will be called in the execution flow. Then, a `Contract`
class has to be defined with validations in it.
This patch allows the `contract` step to take a block containing
validations, attributes, etc. directly. No need to then open a
`Contract` class later in the service.
It also has a nice side effect, as it’s now easy to define multiples
contracts inside the same service. Before, we had the `class_name:`
option, but it wasn’t really useful as you had to redefine a complete
new contract class.
Now, when using a name for the contract other than `default`, a new
contract will be created automatically using the provided name.
Example:
```ruby
contract(:user) do
attribute :user_id, :integer
validates :user_id, presence: true
end
```
This will create a `UserContract` class and use it, also putting the
resulting contract in `context[:user_contract]`.
With the current implementation, a service step can be written as:
```ruby
def my_step(a_default_value: 2)
…
end
```
That’s a pattern we want to avoid as default values (if needed) should
be probably defined in a contract.
This patch makes a service raise an exception if a default value is
encountered.
This will help to enforce a consistent pattern for creating service
actions.
This patch also namespaces actions and policies, making everything
related to a service available directly in
`app/services/<concept-name>`, making things more consistent at that
level too.
This change introduces a new thread notification level allowing users to get notified when someone replies to the thread.
Users who watch a thread will get a green notification on the chat icon and a user notification (blue). User notifications are consolidated based on thread id to prevent cluttering the original users notification area.
---------
Co-authored-by: Régis Hanol <regis@hanol.fr>
Now that `ActiveRecord` relations are properly handled in a `model`
step, putting this step back will allow the service to stops its
execution if there are no users to be removed without calling the pretty
big SQL query contained in the `CalculateMembershipsForRemoval` action.
When we use CTRL/CMD + K to search, on the server we run 4 queries:
- 1 for users
- 1 for groups
- 1 for category channels
- 1 for direct message channels
The server returns up to 10 results per type and the client concatenate all the results (in the order I described) and then takes the first 10 or so.
Let's say you've had 1:1s with john1, john2, and john3. Searching for “john” would return all the johns as “users”. It doesn’t make sense to also return them as 1:1 dms.
That’s what this fixes. When the search returns some users, we skip all 1:1s because we assume they will show up as the results of the “users” query.
We’ll always return matching direct messages with more than 2 users (aka. “group chats”). All 3 other queries (users, groups, and category channels) are unaffected.
Internal ref - t/136079
This is a follow-up to 671f40ce07 and
ed11ee9d05.
While the optimisations in the previous commits were sound, it did not
resolve the memory bloat we were seeing. It turns out that we call
`.blank?` on the model's result if the model has not been marked
optional. The problem with this is that if the model returns an
ActiveRecord relation, calling `.blank?` on the relation basically loads
everything into memory.
Therefore, this commit removes `users` as a model in the since it really isn't
a model but just a relation.
This is a follow up to ed11ee9d05.
In `Chat::AutoRemove::HandleCategoryUpdated`, we are currently loading
the related users record in batches and then handing it off to
`Chat::Action::CalculateMembershipsForRemoval.call`. However, we are
still seeing memory spike as a result of this.
This commit eliminates the allocation of `User` ActiveRecord objects until
absolutely necessary. `Chat::Action::CalculateMembershipsForRemoval.call` has been
updated to accept an ActiveRecord relation instead which allows us to
avoid the ActiveRecord allocations.
This commit seeks to reduce the memory footprint of `Chat::AutoRemove::HandleCategoryUpdated.call`
by optimizing the
`Chat::AutoRemove::HandleCategoryUpdated#remove_users_without_channel_permission` method which was
loading all the ActiveRecord users objects into memory at once. This
change updates the method call to load the ActiveRecord user objects in
batches instead.
This change allows us to distinguish between regular user generated chat messages and those created via the Chat SDK.
A new created_by_sdk boolean column is added to the Chat Messages table. When this value is true, we will not include the message in the user summary email that is sent to users.
This change allows the correct number of members to be added when creating a group direct message, based on the site setting chat_max_direct_message_users.
Previously we counted the current user within the max user limit and therefore the count was off by 1.
Before this fix we could only list messages of a thread if it was part of a `threading_enabled` channel or if the thread was set to `force`.
Due to our design of also using a thread id when this is just a chain of replies so we can switch from threading enabled to disabled at any time, we will allow `Chat:: ListChannelThreadMessages` to list the messages of any thread, the only important requirements are:
- having a thread id
- being able to access this thread
To allow this, this commit simply removes the check on `threading_enabled` or `force`.
When users click a link that points to an existing group chat, we should reopen that chat instead of creating a new group chat so users can more easily continue ongoing conversations.
Prior to this commit, only system users had this pass.
Another significant change of the PR, is to make membership of a channel the angular stone of the permission check to create/update/stop streaming a message. The idea being, if you are a member of a channel already we don't need to check if you can join it AGAIN.
We also have `Chat::AutoRemove::HandleCategoryUpdated` which will deal with permissions change so it's simpler and less prone to error to consider the membership as the only source of truth.
There's no point checking if a user can join a channel if they are already part of it. This case was frequent when using `enforce_membership: true` for custom bots for example.
This commit introduces several enhancements to the ChatSDK module, aiming to improve the functionality and usability of chat thread interactions. Here's what has been changed and added:
1. **New Method: `first_messages`:**
- Added a method to retrieve the first set of messages from a specified chat thread.
- This method is particularly useful for fetching initial messages when entering a chat thread.
- Parameters include `thread_id`, `guardian`, and an optional `page_size` which defaults to 10.
- Usage example added to demonstrate fetching the first 15 messages from a thread.
2. **New Method: `last_messages`:**
- Added a method to retrieve the last set of messages from a specified chat thread.
- This method supports reverse pagination, where the user may want to see the most recent messages first.
- Similar to `first_messages`, it accepts `thread_id`, `guardian`, and an optional `page_size` parameter, defaulting to 10.
- Usage example provided to illustrate fetching the last 20 messages from a thread.
The TextCleaner step has been moved from chat message’s validation to create_message/update_message services. It allows us to easily tweak part of its behavior depending on the needs.
For example we will now disable strip_whitespaces by default when streaming messages as we want to keep newlines and spaces at the end of the message.
This change encourages users to title their threads to make it easier for other users to join in on conversations that matter to them.
The creator of the chat thread will receive a toast notification prompting them to add a thread title when on mobile and the thread has at least 5 sent replies.
That could cause flakeyness in specs depending in which timing message bus would arrive and it's not necessary as it should be updated with `handleThreadOriginalMessageUpdate`.
This change moves the chat message excerpt into a new database column (string) on the chat_messages table.
As part of this change, we will now set the excerpt within the `Chat::CreateMessage` service, and update it within the `Chat::UpdateMessage` service.
This commit will now allow us to track read position in a thread and returns to this position when you open the thread.
Note this commit is also extracting the following components to make it possible:
- `<ChatMessagesScroller />`
- `<ChatMessagesContainer />`
The `UpdateUserThreadLastRead` has been updated to allow this.
Various refactorings have also been done to the code and specs to improve the support of last read.
Why this change?
When the site setting for chat_max_direct_message_users is set to 1, it is expected that users can have a 1:1 chat with other users. However, since the current user is counting as 1 user it makes starting a new chat impossible.
This change hands this validation off to DirectMessageChannel::MaxUsersExcessPolicy which handles the count correctly by filtering out the current user.
This enables the following in Discourse AI
```
plugin.register_modifier(:chat_allowed_bot_user_ids) do |user_ids, guardian|
if guardian.user
mentionables = AiPersona.mentionables(user: guardian.user)
allowed_bot_ids = mentionables.map { |mentionable| mentionable[:user_id] }
user_ids.concat(allowed_bot_ids)
end
user_ids
end
```
some bots that are id < 0 need to be discoverable in search otherwise people can not talk to them.
---------
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
We were incorrectly using `return` in a block which was causing exceptions at runtime. These exceptions were not causing much issues as they are in defer block.
While working on writing a test for this specific case, I noticed that our `upsert_custom_fields` function was using rails `update_all` which is not updating the `updated_at` timestamp. This commit also fixes it and adds a test for it.
Prior to this change we would pre-load all the user channels which making initial page load slower. This change will make them be loaded right after initial load. In the past this was not possible as the channels would have to be loaded on each page transition. However since about a year, we made the channels to be cached on the frontend and no other request will be needed.
I have decided for now to not show a loading state in the sidebar as I think it would be noise, but we can reconsider this later.
Note given we don't have the channels loaded at first certain things where harder to accomplish. The biggest UX change of this commit is that we removed all the complex logic of computing the best channel to display when you load /chat. We will now store the id of the last channel you visited and will use this id to decide which channel to show.
