In order to harmonize the doc options, we are adding
the "Move to my docs" and "Move into a doc" options
to the doc options menu.
This allows users to move documents directly from
the options menu, providing a more streamlined
experience.
We have added a new feature to the editor that allows
users to find and replace text within their documents.
This feature enhances the editing experience by
providing a convenient way to search for specific
words or phrases and replace them with new content.
We changed the responsive store to use throttling
instead of debouncing for the window resize event.
This change ensures that the store updates more
consistently during rapid window resizing,
providing a smoother user experience.
A major release was made on the ui-kit side,
ui-kit is replaced by ui-components, which is a new
package that contains the same components as ui-kit
but with some improvements.
To build the tokens, a new dev dependency is necessary,
which is @gouvfr-lasuite/ui-tokens.
These replacements bring better separation of concerns
and a better architecture for the future.
We need to adapt our codebase to use the new package
and the new dev dependency.
When MIT, we were not able to export anything.
But actually, the export HTML and Print are allowed
for MIT, so we should allow them.
We now allow partially export when MIT, but the
AGPL export is still not allowed (pdf / odt / docx).
We added a new feature that allows users to copy
a link to a specific block within the document.
We can copy the link to the block by clicking on
the "Copy link to block" button in the block's menu
in the document editor.
When the link is pasted in the browser, it will
automatically scroll to the block.
We want to have our own side menu, so we can add more
features to it and display the ui as we want.
The new side menu is called DocsSideMenu
and it will be used in the BlockNoteEditor.
Depend the size of the screen and how the elements
are displayed, we want to have different
shadows and transparency for the floating bar
and the left panel.
The new comment composer cllipped its own formatting toolbar.
That made styling text impossible so we let the toolbar extend past the box.
Signed-off-by: BOUKERFA Mohamed El Amine <boukerfa.ma@gmail.com>
Reset the comment filter to open whenever the side panel
closes, and close the active panel when navigating to
another document. This prevents panel state from
leaking across sessions and documents.
Signed-off-by: fch-aa <21101725+fch-aa@users.noreply.github.com>
Added:
- ♿️(frontend) restore skip to content link after header redesign
- 🌐(i18n) rename cn_CN to zh_CN, add eo_PL and zh_TW locales
- ✨(backend) conditional email notification in server to server api
- ✨(backend) profile api using django-silk
Changed:
- ♿️(frontend) use semantic `<dl>` structure in document info card
- ⚡️(frontend) replace onboarding assets with webm and webp
- 💄(frontend) use the same highlight color for cells and moves
- ⚡️(backend) optimize media_auth endpoint
- 🚸(frontend) print from document options menu
Fixed:
- 🐛(frontend) refresh pins after document deletion and restoration
- 🐛(frontend) redirect homepage to login when homepage feat
is disabled
- 🐛(backend) ignore CSPs for API docs in development
- 🐛(frontend) export images embedded with a relative url
- 🐛(y-provider) fix sentry init
- 🐛(backend) handle object storage metadata keys case-insensitively
- 🐛(keycloak) fix database env variables in the self-hosting example
- 🐛(helm) show the database error while jobs wait for it to be ready
The imported docs were not displayed anymore in the
list grid view directly after the import.
It was because the request was not matched with the
correct query key.
STEP was in uppercase in the assets names, on the
preprod the assets were not found because the
server is case sensitive.
We renamed the assets to lowercase to fix the issue,
as the previous assets.
A new language has been added to the Django system,
allowing for Polish translations and localization
support.
We need to initialize the Polish language files before
being able to download the translations from
Crowdin. This commit includes the initial setup for
the Polish language, including the necessary configuration
files and directory structure.
Lot of elements did not have a focus style, or the
focus style was not consistent. We now have a global focus
style that is applied to all elements that are
focusable. The style is still overridable for
specific elements, so buttons and other elements
still have their own focus style.
We will have probably to improve the focus style
in the future of some elements, but this is a
good start.
We updated the UI of the doc toolbox to fit
with the design system. The icons have been updated,
we moved it to the doc management folder in order
to use it to different places in the app.
We have removed the "Copy as Markdown" option from
the document options menu.
If user need to copy the document content as markdown,
they can just copy the content from the document
editor and paste it into a markdown editor.
Webp and webm are more efficient formats
than gif and png, so we replace the
onboarding assets with these formats to improve
performance and decrease file size.
Once the sql queries improved we have still a bottleneck on large
concurrent requests on this endpoint. We notive in the profiles generated
that lot of time was spent in creating a new s3 client instance on each
request. django_storage use a thread local cache for signed and unsigned
connection, but using uvicorn we have a new thread for each request, so
on each request a new s3 client is generated and it appears to be an
expensive operation. To fix this issue, we cache the client and share it
accross all the thread and requests.
On the media_auth endpoint the first bottleneck we have is with
postgresql. We are looking for too much data and no index is used on the
attachments colum. When the lookup filter on the attachement columns, a
full scan is made on all the document table looking for each element in
the array, this operation is really expensive. To fix this we created a
GIN index on the attachments column. Also the readable_per_se lookup was
selecting too much data combined with the filter_descendants function.
We remove the usage of the filter_descendants, we choose to first fetch
all the paths where the attachment is found, this operation is fast
thanks to the new index, split all the paths in candidate paths and then
filter readable_per_se queryset with these paths. All these
modifications make the endpoint faster.
We have performance issue with the media_auth endpoint. We need to
monitor and profile this endpoint to understand what is happening. We
decided to install and configure django-silk. By default django-silk is
not enabled and not accessible, once enabled migration must be run and
the /silk/ endpoint will be accessible only using a super user.
To have relecant profiles and monitor we must have a database comparable
to a production one because we are facing a scaling issue. Unless a
million documents probably nothing is happening, for this we added
several management commands to help us. A first one to anonymize a
database than can be reused without any risk to leak PII data, can be
called using python manage.py anonymize_database with some mandatory
options. An other command to generate a database profile, this command
should be run on a production database, this profile can then be used by
the generate_volumetry to populate a development database with same
volumetry than describe in the profile.
Object storage metadata keys are case insensitive per the S3
specification, but implementations don't agree on the case they give
back. When head_object returns a capitalized "Status", updating the
attachment status added a second, lowercase entry instead of replacing
it, and the copy request ended up carrying two x-amz-meta-status
headers. Ceph RadosGW loses one of them behind a proxy, which
invalidates the request signature.
The same assumption was made when reading the status back in media-auth
and media-check, where an attachment stored on such a backend stayed in
"processing" forever.
Metadata read from the storage is now normalized to lowercase keys
before being consumed or copied over.
Signed-off-by: risk-alt <aldu6974@gmail.com>
Favorite relationships persist across soft deletion, while the sidebar keeps
a stale cached list until reload. Invalidate the favorites query after deletion
and restoration so deleted documents disappear immediately and their pins
return when restored.
Signed-off-by: fch-aa <21101725+fch-aa@users.noreply.github.com>
The migrate and createsuperuser jobs polled `manage.py check` with its
output sent to /dev/null, then printed "Database not ready" whatever
the reason. Any failure the check reports, a missing setting or a bad
credential for instance, looked like a database that had not started
yet, and the job looped forever without a single clue in its logs.
The check output is now captured and printed on each failed attempt,
along with the attempt number, so the reason the job is still waiting
is readable with kubectl logs. The retry behaviour is unchanged.
Applied to both jobs in the chart values and in the dev and feature
helmfile environments. The chart README table is regenerated
accordingly.
Signed-off-by: risk-alt <aldu6974@gmail.com>
Blocknote highlights the selected table cells with #c8c8ff66 and draws
the drop cursor of a moved block with #ddeeff: two different colors,
both out of the palette.
Use --c--contextuals--background--semantic--brand--tertiary for both, and
for the drop indicator of a moved table row or column which was #adf.
The cell overlay is blended with multiply because it is drawn on top of
the cell content while the token is opaque.
Signed-off-by: risk-alt <aldu6974@gmail.com>
Images embedded with a relative url such as /assets/logo.png were routed
through the CORS proxy: the same origin check used a substring match on
window.location.hostname, which a relative url never contains. The
backend rejects those urls with a 400 as it only validates absolute
http(s) urls.
Resolve the url against the current origin instead, so relative and
absolute same origin urls are fetched directly. This also stops matching
unrelated external urls that merely contain the hostname, and replaces
the "base64" substring check by an explicit data: scheme check.
exportResolveFileUrl now checks response.ok as well: fetch() only
rejects on network errors, so the 400 payload was returned as a blob and
embedded in the export instead of falling back to the url.
Signed-off-by: risk-alt <aldu6974@gmail.com>
When tests were run in parallel, the language tests
could introduce flakiness due to a switch to the
default language not being completed before the
next test started.
ui-kit v0.28 breaks some UI components.
This commit updates the relevant files to ensure
compatibility with the new version of ui-kit.
We update the cunningham-tokens.css and
cunningham-tokens.ts files to reflect the changes
in the design tokens.
With Content Security Policies activated, swagger (and redoc)
auto-generated API documentation is no longer accessible even locally.
To restore this feature, we've excluded CSP for related URLs only in
the `Development` configuration.
Add a "Present" item to the block side menu that opens the presenter on
the slide containing that block. Map any block id (incl. nested or
divider ids) to its rendered content slide.
Closes#2470
Since pydantic-ai has ben upgraded deprecation warning were present in
the logs. This commint change how the instrumentation is enabled to be
compatible with hoe pydantic-ai manage it. Also prompt system are
ignored when they come from the front, like we are doing and dropped. To
maintain the system prompt we use the instruction parameter instead.
We upgraded pydantic-ai-slim to the latest version and add mistral has
extra dependency. Doing this, the direct dependance to mistral can be
removed and will be managed with pidantic-ai-slim package. The import of
mistralai has changed, we have to updagre them from mistralai to
mistralai.client
Anthony LC