* Initial implementation; Happy flow
* Maybe revert this
* Functional prototype
* feat(facial-expressions): get stream when changing background effect and use presenter effect with camera
* add(facial-expressions): array that stores the expressions durin the meeting
* refactor(facial-expressions): capture imagebitmap from stream with imagecapture api
* add(speaker-stats): expression label
* fix(facial-expression): expression store
* revert: expression leabel on speaker stats
* add(facial-expressions): broadcast of expression when it changes
* feat: facial expression handling on prosody
* fix(facial-expressions): get the right track when opening and closing camera
* add(speaker-stats): facial expression column
* fix(facial-expressions): allow to start facial recognition only after joining conference
* fix(mod_speakerstats_component): storing last emotion in speaker stats component and sending it
* chore(facial-expressions): change detection from 2000ms to 1000ms
* add(facial-expressions): send expression to server when there is only one participant
* feat(facial-expressions): store expresions as a timeline
* feat(mod_speakerstats_component): store facial expresions as a timeline
* fix(facial-expressions): stop facial recognition only when muting video track
* fix(facial-expressions): presenter mode get right track to detect face
* add: polyfils for image capture for firefox and safari
* refactor(facial-expressions): store expressions by counting them in a map
* chore(facial-expressions): remove manually assigning the backend for tenserflowjs
* feat(facial-expressions): move face-api from main thread to web worker
* fix(facial-expressions): make feature work on firefox and safari
* feat(facial-expressions): camera time tracker
* feat(facial-expressions): camera time tracker in prosody
* add(facial-expressions): expressions time as TimeElapsed object in speaker stats
* fix(facial-expresions): lower the frequency of detection when tf uses cpu backend
* add(facial-expressions): duration to the expression and send it with durantion when it is done
* fix(facial-expressions): prosody speaker stats covert fro string to number and bool values set by xmpp
* refactor(facial-expressions): change expressions labels from text to emoji
* refactor(facial-expressions): remove camera time tracker
* add(facial-expressions): detection time interval
* chore(facial-expressions): add docs and minor refactor of the code
* refactor(facial-expressions): put timeout in worker and remove set interval in main thread
* feat(facial-expressions): disable feature in the config
* add(facial-expressions): tooltips of labels in speaker stats
* refactor(facial-expressions): send facial expressions function and remove some unused functions and console logs
* refactor(facial-expressions): rename action type when a change is done to the track by the virtual backgrounds to be used in facial expressions middleware
* chore(facial-expressions): order imports and format some code
* fix(facial-expressions): rebase issues with newer master
* fix(facial-expressions): package-lock.json
* fix(facial-expression): add commented default value of disableFacialRecognition flag and short description
* fix(facial-expressions): change disableFacialRecognition to enableFacialRecognition flag in config
* fix: resources load-test package-lock.json
* fix(facial-expressions): set and get facial expressions only if facial recognition enabled
* add: facial recognition resources folder in .eslintignore
* chore: package-lock update
* fix: package-lock.json
* fix(facial-expressions): gpu memory leak in the web worker
* fix(facial-expressions): set cpu time interval for detection to 6000ms
* chore(speaker-stats): fix indentation
* chore(facial-expressions): remove empty lines between comments and type declarations
* fix(facial-expressions): remove camera timetracker
* fix(facial-expressions): remove facialRecognitionAllowed flag
* fix(facial-expressions): remove sending interval time to worker
* refactor(facial-expression): middleware
* fix(facial-expression): end tensor scope after setting backend
* fix(facial-expressions): sending info back to worker only on facial expression message
* fix: lint errors
* refactor(facial-expressions): bundle web worker using webpack
* fix: deploy-facial-expressions command in makefile
* chore: fix load test package-lock.json and package.json
* chore: sync package-lock.json
Co-authored-by: Mihai-Andrei Uscat <mihai.uscat@8x8.com>
Skipping node-canvas dependency in resemblejs as it [fails in certain environments](https://github.com/rsmbl/Resemble.js#nodejs).
The reference states
> If you are using Resemble.js for in-browser analysis only, you can skip the node-canvas dependency.
The majority of the code is in the WASM file, the JS is just 9KB.
It's so little, in fact, that the performance hint for the main bundle didn't
have to be adjusted.
The majority of the code is in the WASM file and models, this is just a few KB.
It's so little, in fact, that the performance hint for the main bundle didn't
have to be adjusted.
* Update dateUtil.js
* version up moment
* exclude unnecessary languages in Moment.js from webpack
* add Occitan of Moment.js
* Fixed auto-formatting
* add require missing by mistake
app.bundle.js before: 3851549 after: 4506493.
The culprit for the bloat was Olm. It feature-detects the environment in order
to pick a suitable random byte generator, and alas Webpack includes the None
crypto pollyfill. This is due to the existence of the "node" block in our
Webpack configuration file.
The solution is to provide empty modules to make bundling work, as we did
already for the fs module, since we know they are not used at runtime.
Adapt to E2EE changes in lib-jitsi-meet. Notably:
---
e2ee: introduce per-participant randomly generated keys
This the second stage in our E2EE journey.
Instead of using a single pre-shared passphrase for deriving the key used for
E2EE, we now establish a secure E2EE communication channel amongst peers.
This channel is implemented using libolm, using XMPP groupchat or JVB channels
as the transport.
Once the secure E2EE channel has been established each participant will generate
a random 32 byte key and exchange it over this channel.
Keys are rotated (well, just re-created at the moment) when a participant joins
or leaves.
---
Up until now we relied on implicit loading of middlewares and reducers, through
having imports in each feature's index.js.
This leads to many complex import cycles which result in (sometimes) hard to fix
bugs in addition to (often) breaking mobile because a web-only feature gets
imported on mobile too, thanks to the implicit loading.
This PR changes that to make the process explicit. Both middlewares and reducers
are imported in a single place, the app entrypoint. They have been divided into
3 categories: any, web and native, which represent each of the platforms
respectively.
Ideally no feature should have an index.js exporting actions, action types and
components, but that's a larger ordeal, so this is just the first step in
getting there. In order to both set example and avoid large cycles the app
feature has been refactored to not have an idex.js itself.
Without changing the 'Host' header, a standalone instance we use for
jitsi dev work will generate:
config.websocket = 'wss://localhost:8080/xmpp-websocket'
This was fine with BOSH, but a Websocket will not connect.
With this change it will be(based on the default devServerProxyTarget):
config.websocket = 'wss://alpha.jitsi.net/xmpp-websocket'
The currently selected values are a bit above the actual sizes, so if a PR
increases the bundle size enough to trigger the failure, it should bump it.
It better have a good reason for it though!
Calinteodor
Avram Tudor
Gabriel Borlea
Saúl Ibarra Corretgé
Vlad Piersec
Werner Fleischer
Steffen Kolmer
hmuresan
tudordan7
るしふぁ
Hristo Terezov
Erik Demaine
Mihai Uscat
khajaamin
Felix C. Stegerman
Bettenbuk Zoltan
ludovicm67
paweldomas
Andrei Gavrilescu
Andrei Gavrilescu