* added transcode configs for nvenc,qsv and vaapi
* updated dev docker compose
* added software fallback
* working vaapi
* minor fixes and added tests
* updated api
* compile libvips
* move hwaccel settings to `hwaccel.yml`
* changed default dockerfile, moved `readdir` call
* removed unused import
* minor cleanup
* fix for arm build
* added documentation, minor fixes
* added intel driver
* updated docs
styling
* uppercase codec and api names
* formatting
* added tests
* updated docs
* removed semicolons
* added link to `hwaccel.yml`
* added newlines
* added `hwaccel` section to docker-compose.prod.yml
* ensure mesa drivers are installed
* switch to mimalloc for sharp
* moved build version and sha256 to json
* let libmfx set the render device
* possible fix for vp9 on qsv
* updated tests
* formatting
* review suggestions
* semicolon
* moved `LD_PRELOAD` to start script
* switched to jellyfin's ffmpeg package
* fixed dockerfile
* use cqp instead of icq for qsv vp9
* updated dockerfile
* added sha256sum for other platforms
* fixtures
* Remove unnecessary PG_DATA environement variable from docker-compose.yml
There is no need to set the PostgreSQL data directory to the default location, it just adds an additional unnecessary line to the docker-compose file.
In addition, the PG_DATA isn't even the correct environment variable name (it should be PGDATA, see: https://hub.docker.com/_/postgres/), so this environment variable was never doing anything to begin with.
* Update docker-compose.dev.yml
* Update docker-compose.prod.yml
* Update docker-compose.test.yml
Previously, if you'd shut down the `make-dev` command and restart
the docker daemon (say, on a system reboot), there would be 3 immich
containers already running.
* using pydantic BaseSetting
* ML API takes image file as input
* keeping image in memory
* reducing duplicate code
* using bytes instead of UploadFile & other small code improvements
* removed form-multipart, using HTTP body
* format code
---------
Co-authored-by: Alex Tran <alex.tran1502@gmail.com>
* updated dockerfile, added typing, packaging
apply env change
* added arm64 support
* added ml version pump, second try for arm64
* added linting config to pyproject.toml
* renamed ml input field
* fixed linter config
* fixed dev docker compose
A default entrypoint and command make it just a bit easier to use the images as
there is no longer a need for an explicit entrypoint. The exception is the
server image, which still requires the shell script to be specified.
* default NODE_ENV to production for server image
* update node image to use 3.17 alpine in server
* update web docker image to use alpine 3.17
* remove NODE_ENV from production docker-compose
* NODE_ENV is also needed default in machine-learning
* build: add typesense to docker
* feat(server): typesense search
* feat(web): search
* fix(web): show api error response message
* chore: search tests
* chore: regenerate open api
* fix: disable typesense on e2e
* fix: number properties for open api (dart)
* fix: e2e test
* fix: change lat/lng from floats to typesense geopoint
* dev: Add smartInfo relation to findAssetById to be able to query against it
---------
Co-authored-by: Alex Tran <alex.tran1502@gmail.com>
* fix(machine-learning): Add the command to execute at startup
Previously it wasn't set in the Docker container but it should be.
* fix(docker): remove machine-learning command arg
* fix(docker): machine-learning CMD argument
BREAKING CHANGES
* Users have to update the docker-compose file, machine-learning portion.
* Temporary dropping machine-learning support for Arm64 and Armv7
* chore: add typeorm commands to npm and set default database config values
* feat: move to server side authentication tokens
* fix: websocket should emit error and disconnect on error thrown by the server
* refactor: rename cookie-auth-strategy to user-auth-strategy
* feat: user tokens and API keys now use SHA256 hash for performance improvements
* test: album e2e test remove unneeded module import
* infra: truncate api key table as old keys will no longer work with new hash algorithm
* fix(server): e2e tests (#1435)
* fix: root module paths
* chore: linting
* chore: rename user-auth to strategy.ts and make validate return AuthUserDto
* fix: we should always send HttpOnly for our auth cookies
* chore: remove now unused crypto functions and jwt dependencies
* fix: return the extra fields for AuthUserDto in auth service validate
---------
Co-authored-by: Jason Rasmussen <jrasm91@gmail.com>
* feat(.well-known): add .well-known/immich to reference API endpoint
* feat(.well-known): make schema optional (defaults to https)
* adjust method comment to be a little less confusing
* fix casting issue with resovled url
* include when checking Well-known, update server hint
* add validation for login form's server url
* consolidate common process into resolveAndSetEndpoint
* fix missed prettier formatting
* revert translation changes
* update environment variable description, hopefully a bit clearer
* rename environment variable to IMMICH_API_URL_EXTERNAL
* comment out optional env variables
* fix(web): browser-side api client to include authorization token
* Revert "fix(web): browser-side api client to include authorization token"
This reverts commit 60e338938f.
* remove multi-domain related changes
* Add proxy changes
* Add web changes
* Add microservices changes
* Add examples
* Add header comment to nginx config
* Use URLs instead of host and port
Implemented a mechanism to extract the correct time zone from the GPS coordinate if presented in the file's EXIF, and to convert the timestamp to the correct UTC time so that the time will show correctly based on the mobile/web local time zone.
