* increases timeout
* refactor: reuse constants for redis db
* avoid closing connection on pong error
* fix: prevent electron from pausing timer executions
Our pings could be paused when our app is backgrounded.
* refactor: remove unnecessary refs
* fix: presence unreliable within same pod
It was broadcasting presence changes to other pods, but not notifying
clients connected to the same pod
- New cron job in cluster
- Handle presence and other concerns
- Toggle in app to disable email notifications
- Handles other edge cases such as cooldown period
- Simple html email with simple message
Closes#117
* setup infra for pusher service
* setup client sdk for pusher service
* fix: ping parse failure
* fix: send pong back to client
avoid disconnections every 2.5 minutes
* increase replicas
* feat: show presence and compose indicator
It wasn't progressing, likely because the metadata for a screen
recording was not read by the video html element. So we fallback to our
own duration metadata.
Closes#135
Closes#131
Fundamentally, we cannot afford to composite webcam and screen on the
client. We should create two streams and then combine them with ffmpeg
processing in the backend / orion.
* first pass implementation
* remove screenrecord shortcut hint
* refactor
* fix: prevent mirror review of screen recording
* feat: allow recording with webcam overlay
* fix: review screen recording without clipped content
* tidy keyboard hints consistent position
* refactor: re-use one component for screen picker
huddles and screen clips use same picker component now. We had to make
sure that tailwind works for both of them.
* fix: improve visibility of keyboard hints
During compose video or screen recording, the keyboard hints were
invisible if the content was super bright.
This created some distortion because while we were hiding the autoplay
window, this would not stop it's media. It's better to stop the media
completely
Resolves#110
Goes the extra mile and supports markdown for sending code and formatted
text, almost document-like. It feels more like GitHub's comments and
text areas.
Now we have network root and stream-view defining their own hints
instead of trying to re-use one component. And we also introduce an
overlay for additional commands for people.
- Reduces surface area
- Reduces options for people (less cognitive load)
- One way to use the product
- List view works better for keyboard shortcuts
- List view has basics such as recency
* refactor: use interfaces for function params
* cleanup message retention code
* support open / closed streams
- Tabs for viewing separately
- Context menu to close / open streams
- Update stream particle status field
* ui tweak
* show stream state in stream-view
* ui tweaks
* cleanup message retention from orion api
* add token endpoint for livekit
* fix: invalid type passed to hook
* fix: inject livekit env variables for orion
* fix: show controls indicator above stream # shortcut
* return livekit server url from api
* simple huddle implementation with streams
* set human name in livekit room context
* simplify deployment tooling
* join huddle with audio automatically
* feat: show when there is an active huddle
This introduces a webhook which listens to events from livekit and
updates our firestore stream particle. It keeps the client simple,
reacting to changes to firestore docs.
* use headphones icon for huddles
* fix: screenshare not working in electron
The default VideoConference component from livekit doesn't support
screenshare in electron. This attempts to compose our own layout with
livekit components ourselves and introduces our own flow for
screenshare.
* prototype with two windows
* use deps in separate window render process
* enhance autoplay into desktop overlay
* fix: autoplay window not applying tailwind styles
This was causing issues in which the last_child_created_at field for
stream particles was slightly after the actual particle's creation time,
and so comparing whether it's unseen merely via the playback markers
would be inaccurate.
This generates the transcript and shows the caption experience on the
client side for media particles. It also simplifies other side effects
that we must perform such as updating the `last_child_created_at` field
for stream and container particles.
Making the firestore subcollections the same name helps with creating
indices and queries across subcollections of different depths. This also
prevents the need for conditional logic for network root / top-level
particles.
* refactor: update api and client to reference humanIds
* fix: prevent deletion of network member
This may cause various side effects if there is data in other services
which reference this member
talksik