* chore: Move Pricing validation to on create
The previous validation was getting triggered on all transactions,
this would prevent login to existing users and session data need to be updated.
Opting for a less intrusive approach.
* chore: minore change
- This PR introduces a modification to the channel fetching logic, ensuring that channels with older message_template_last_updated timestamps are prioritized during synchronization.
- Reorganizing installation config settings to move more configurations into UI from environment variables
- Changes to installation config to support premium plans in the enterprise edition
- Fixes the broken premium indicator in account/show and accounts/edit page
- Added a trait called microsoft_email for the Channel::Email factory.
- Rewrote the logic to make it simple to understand
- Rewrote the specs for readability
- We previously had this spec inside the enterprise folder which wouldn't be picked up by our build process, so moving to the correct folder instead.
Co-authored-by: Shivam Mishra <scm.mymail@gmail.com>
* feat: set cache keys for 3 days only
* feat: invalidate should set the latest timestamp
* refactor: cache_keys concern
* remove invalidate_cache method
* refactor reset to set to new value instead of delete
* ensure only one event is dispatched
* feat: set expiry to 24 hours
* chore: make expiry 48 hours
* feat: include destroy event
* feat: set expiry to 72 days
* fix: typo
* test: cache update after `touch`
* test: update cache keys
* refactor: remove after_touch, it's already handled in commit
The system did not detect the delivery errors earlier, resulting in some false positives. The user was not informed when an email failed to be delivered. While we do handle failure status in other channels, we were not able to capture the actual delivery status for the email channel.
This pull request makes the following changes:
- Updated the class EmailReplyWorker to use the deliver_now method instead of deliver_later. This change is made to raise any errors that may occur with the SMTP connection. The errors are then captured and sent to Sentry, and the email is marked as failed. Previously, we did not consider the case of retries in the email channel, so this feature is currently not functioning. So, I have disabled the retry option. We will address this in a follow-up ticket.
- Downgraded the net-smtp gem to version 0.3.4. This change is made to avoid an argument error when using XOAUTH2.
Fixes: https://linear.app/chatwoot/issue/CW-3032/argumenterror-wrong-authentication-type-xoauth2-argumenterror
Currently, we do not support reactions, ephemeral messages, or the request_welcome event for the WhatsApp channel. However, if this is the first event we receive in Chatwoot (i.e., there is no previous conversation or contact in Chatwoot), it will create a contact and a conversation without any messages. This confuses our customer, as it may appear that Chatwoot has missed some messages. There are multiple cases where this might be the first event we receive in Chatwoot. One quick example is when the user has sent an outbound campaign from another tool and their customers reacted to the message.
Another event like this is request_welcome event. WhatsApp has a concept for welcome messages. You can send an outbound message even though the user has not send a message. You can receive notifications through a webhook whenever a WhatsApp user initiates a chat with you for the first time. (Read the Welcome message section: https://developers.facebook.com/docs/whatsapp/cloud-api/phone-numbers/conversational-components/ ). Although this can help the business send a pro-active message to the user, we don't have it scoped in our feature set. For now, I'm ignoring this event.
Fixes https://linear.app/chatwoot/issue/CW-3018/whatsapp-handle-request-welcome-case-properly
Fixes https://linear.app/chatwoot/issue/CW-3017/whatsapp-handle-reactions-properly
Fixes the follow cases
- The ensure block released the lock even on LockAcquisitionError
- Custom timeout was not allowed
This also refactored the with_lock method, now the key has to be constructed in the parent function itself
Co-authored-by: Sojan Jose <sojan@pepalo.com>
This allows a user to add/update a custom regex and a cue while defining custom attributes(Only applicable for type- text).
While adding/editing custom attributes, the values are validated against the attribute definition regex, and if it is incorrect, a cue message or default error message is shown and restricts invalid values from being saved.
Fixes: #6866
Message search would frequently timeout. The reason was that the query would join the conversation too, the new query searches the message table directly
Co-authored-by: Sojan <sojan@pepalo.com>
Consider a scenario where two jobs are concurrently accessing a job with a lock mechanism. If the second job accesses lock_manager.locked? before the first job called lock_manager.lock(lock_key), it would return a false positive. At this point, both jobs can be executed freely.
To address this issue, the following change ensures that only the current thread that has set the lock can execute. Otherwise, a LockAcquisitionError will be thrown.
Co-authored-by: Sojan Jose <sojan@pepalo.com>
Co-authored-by: Shivam Mishra <scm.mymail@gmail.com>
- this ensures that the markdown formatted messages from the Chatwoot dashboard will render consistently in telegram UI for the supported types like bold, italics, links etc
While debugging a sentry error for "ActiveRecord::InvalidForeignKey ActiveStorage::Representations::RedirectController", it was noticed that we enqueue a Avatar::AvatarFromUrlJob for each setUser call, which is unnecessary. Hence making this call only if the contact doesn't have an existing avatar.
If one needs to have this avatar updated, they can go to the contacts tab and delete the current avatar, Chatwoot will pick up the new avatar in subsequent API call.
This PR adds support for Markdown in the telegram API for send message. The dashboard uses commonmark syntax for markdown but telegram is using MarkdownV2.
More info - https://core.telegram.org/bots/api#markdownv2-style
Adds support for bold - *bold*, _italic text_, __underline__, ~strikethrough~
Co-authored-by: Sojan Jose <sojan@pepalo.com>
In rare cases, the API call to Telegram for the file path fails. We were logging the error in sentry, switching to logs instead.
Co-authored-by: Sojan <sojan@pepalo.com>
In the previous release, we enabled "HTTP Basic Authentication" to secure all attachments requiring HTTP authentication. This is particularly important for media files that may contain sensitive data, as recommended by Twilio. However, some users experienced issues because they did not enable this option despite our alerts prompting them to do so. If the authenticated attachment download call fails, add another call to download the attachment without authentication.
The sentinel configuration set in this specification seems to be affecting other specifications. So, let's ensure that the memoised config variable gets unset after execution of the spec.
- This PR adds a feature to auto-trigger handoff events when an Agent bot toggles a conversation status from Pending to Open
Co-authored-by: Sojan <sojan@pepalo.com>
This PR logs additional information in content_attributes of a message in case it is unsupported. This info can be used by the client to render a fresh UI
- Ability to configure Chatwoot custom branding from UI when using paid plans
Co-authored-by: Nithin David Thomas <1277421+nithindavid@users.noreply.github.com>
- The cache stores the labels as label1, label2. Without removing the whitespace, the output of cached_label_list_array would be [label1, [whitespace]label2], which doesn't match with the existing labels. This list is returned with proper label names.
- Revamp the super admin design
- Introduce a new settings page for support and billing settings
- Move the access tokens into users, agent bots and platform app show pages
Co-authored-by: Sojan <sojan@pepalo.com>
Co-authored-by: Pranav Raj S <pranav@chatwoot.com>
- Mark all messages as read by providing the conversation ID and timestamp.
- For Instagram, ensure all previous messages that weren't marked as failed are now marked as read. This is because the read events are only triggered for the most recent message and not for any previous ones.
- We have observed some failures for data import jobs in the cloud due to race conditions with job executions and active storage file uploading. This PR adds delays and retries to accommodate that.
This PR refactors the schema we introduced in #7518 based on the feedback from production tests. Here is the change log
- Decouple Inbox association to a new table inbox_response_sources -> this lets us share the same response source between multiple inboxes
- Add a status field to responses. This ensures that, by default, responses are created in pending status. You can do quality assurance before making them active. In future, this status can be leveraged by the bot to auto-generate response questions from conversations which require a handoff
- Add response_source association to responses and remove hard dependency from response_documents. This lets users write free-form question answers based on conversations, which doesn't necessarily need a response source.
- Add webhook events for inbox creation/updation.
- Right now, the feature is added under a feature_flag. It is not available by default on all installations.
DISTINCT query with custom attributes return an error. To avoid the error, this PR refactors the query to include tags only when it is required.
Fixes#7931Fixes#7836
Co-authored-by: Sojan Jose <sojan@pepalo.com>
Within the article we are now rendering every link for youtube and vimeo as embeds. This isn't a good solution, as users might need to have plain links as well. This fix will render only links within two blank lines as embeds.
Co-authored-by: Sojan Jose <sojan@pepalo.com>
We've had some messages come in from a few different phone numbers that had null bytes in them. I don't know how this happens. They don't seem to be malicious.
They currently cause the Postgres gem to raise an error when Chatwoot attempts to save the message body to the database:
ArgumentError (string contains null byte)
Related Rails GitHub issue: rails/rails#26891
- Ensure existing contact information is updated on data import
- Refactor the existing job to make it more readable
- Fixes issues with import files in the wrong encoding
fixes: #7307
This update will mean that errors will roll back the current transaction and the error will be sent back to the frontend and the user will know that the Inbox did not finish setting up successfully.
- adds an audit log when an agent is added or removed from a team
- adds an audit log when an agent is added or removed from an inbox
Co-authored-by: Sojan Jose <sojan@pepalo.com>
Updating the `unattended` tab to include conversations where the customer responded and is awaiting an agent's response.
Previously it showed only the conversations where the first response was pending.
Co-authored-by: Pranav Raj S <pranav@chatwoot.com>
When using client APIs to create conversations and auto-assignment is turned on, welcome messages were getting triggered. This PR disable the behaviour and ensure template hooks are triggered only if there are incoming messages present.
Fixes: https://linear.app/chatwoot/issue/CW-2187
This commit introduces the ability to associate response sources to an inbox, allowing external webpages to be parsed by Chatwoot. The parsed data is converted into embeddings for use with GPT models when managing customer queries.
The implementation relies on the `pgvector` extension for PostgreSQL. Database migrations related to this feature are handled separately by `Features::ResponseBotService`. A future update will integrate these migrations into the default rails migrations, once compatibility with Postgres extensions across all self-hosted installation options is confirmed.
Additionally, a new GitHub action has been added to the CI pipeline to ensure the execution of specs related to this feature.
Co-authored-by: Shivam Mishra <scm.mymail@gmail.com>
Co-authored-by: Nithin David Thomas <1277421+nithindavid@users.noreply.github.com>
Co-authored-by: Pranav Raj S <pranav@chatwoot.com>
Currently, if Enable email collect box is enabled, and the user starting the chat already has an account with the same email, it will override the current account with {name} from the email name@email.com , this fixes it.
Co-authored-by: Sojan <sojan@pepalo.com>
We have display_name and the actual name for an agent. display_name is used in all public facing parts. The https://github.com/chatwoot/chatwoot/pull/7232/files PR missed this and used actual name instead.
This PR fixes it.
* logging the messages id for message validation exception
* Update the processed_message_content validation over length
* codeclimate
* specs failing for contacts
- Adds support for superscript when rendering article markdown
- Chatwoot Markdown Render to render markdown everywhere
Co-authored-by: Sojan <sojan@pepalo.com>
* fix: calculation for resolution count
* test: resolution count bug fix
- ensure enqueued jobs are run
- fix the dates check, conversations resolved today should show up in today
* feat: ensure conversations are resolved
* test: do not count extra events if the conversation is not resolved currently
* fix: typo
Update the cache headers for cache_keys to max-age=10, private, stale-while-revalidate=300
1. The cache will be fresh for 10 seconds (max-age=10). During this time, the browser will use the cached version without checking with the server.
2. After the initial 10 seconds, the browser can continue to serve the stale cache for up to 5 minutes (stale-while-revalidate=300). During this period, it will also try to revalidate and update the cache in the background.
3. After 310 seconds in total (10 seconds fresh, 300 seconds stale), if the browser has not been able to revalidate the cache, it will attempt to fetch the fresh resource directly from the server for subsequent requests, causing potential latency equivalent to a network request.
This means that the data will be directly revalidated only every 5 mins. Other times, it will stay fresh for 10 seconds and revalidate in the background. In most cases, we won't have to rely on a cache validation check because there is a WebSocket event for revalidation, so we know if something changes.
Right now the stale-while-revalidate is 5 minutes, we can then move it to 15 minutes.
> The stale-while-revalidate header is not supported in Safari, for Safari the cache keys will only stay in memory for 10 seconds before being marked stale
Allows super admins to reset the cache for an account. This will force the front end to fetch the data again on the next load.
fixes: https://linear.app/chatwoot/issue/CW-1817
Co-authored-by: Sojan <sojan@pepalo.com>
- update the templates updated at, even if the API request fails ( to prevent jobs from stacking up in case of API failures upstream )
- sequence the job in batches of 25 requests per minutes schedule ( in case API response time is high, also not to send too many requests in a single batch )
- move the sync job re-rerun to 3 hours ( since we are updating the updated at even in case of failures )(prev 15 minutes )
Fixes: https://linear.app/chatwoot/issue/CW-1590
This change adds the ability to include a `source_id` param when querying the `/api/v1/accounts/{account_id}/conversations/search` endpoint. It restricts to results to only conversations related to a contact_inbox with the provided parameter. My motivation for adding this feature was to allow an external API to communicate with a specific conversation with only an awareness of the conversation `source_id` from the client.
Co-authored-by: Sojan Jose <sojan@pepalo.com>
* feat: Enable features on successful subscription
* fix: Add specs
* disable features for default plan
* Remove mark as cloud customers
---------
Co-authored-by: Pranav Raj S <pranav@chatwoot.com>
- Add JsonSchemaValidator, which takes a declarative schema and validates it for a given property.
- Add specs for JsonSchemaValidator
- Enable the validator for template_params
* feat: update seed script to include prioritt
* feat: add sort_handler for conversations
* test: sort on priority order
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
* feat: include timezone offset in summary calculation
* fix: exlcude end in date range
* test: explicit end of day
* fix: test for report builder
* fix: reports.spec.js
---------
Co-authored-by: Tejaswini Chile <tejaswini@chatwoot.com>
When the CC field is generated in the UI, the email values are joined together
with ", " but when they are parsed, we currently split by just ",".
This causes an error on the backend and on the frontend.
It seems reasonable to update the code to allow whitespace in the input and to
split by `\s*,\s` and also to trim leading and trailing whitespace from the CC
list.
---------
Co-authored-by: Sojan <sojan@pepalo.com>
* fix: Invalidate cache if any of the related objects change
* Add specs:
* Update context
---------
Co-authored-by: Shivam Mishra <scm.mymail@gmail.com>
The current way of writing specs for automation in one single file automation_listener isn't effective. Hence we are breaking down the specs for each class into separate spec files.
fixes: CW-1447
- Strip search term before searching
- order messages by created_at desc
- order contacts by last_activity_at desc
- order conversations by created_at desc
- Search only resolved contacts
- Optimize resolved contacts query
ref: #6583
