Connect your remote data storage to Supervisely Platform without data duplication.
NOTE #1: most frequest usecase is when Enterprise Customer would like to connect huge existing data storage (tens of terabytes) and avoid data duplication. In other cases we recommend to use general import procedure to store data in Supervisely Data Storage.
NOTE #2: this release works only with data structured in Supervisely format. In future versions raw images, videos, and other formats will be added.
If you have ideas or suggestions, please post an idea in 💡Supervisely Ideas Exchange or chat with us in slack.
cd
to directory with the data you want to connect. For example: cd work/data
. We recommend to avoid using special characters in paths (spaces, etc…).docker run -p 8088:80 -v $(pwd):/mnt/data jetbrainsinfra/nginx-file-listing:0.2
nginx.conf
to disallow it.Now you can check that data is accesible in browser:
Step 1: Add app to your team from Ecosystem if it is not there.
Step 2: Go to Plugins & Apps
section in current team. And press Run
button in front of application.
Step 3: You will be redirected to Current Workspace
->Tasks
page. Wait until app is started and press Open
button.
Note: Running procedure is simialr for almost all apps that are started from context menu. Example steps with screenshots are here in how-to-run section.
Put address to the remote directory with project in Supervisely format (where meta.json
is). And press Preview remote
.
List of all files and directories will be shown. Every directory is dataset (because this release supports only supervisely format) with two folders img
and ann
. Select what datasets will be uploaded.
Define destination project. Workspace will be created if not exist. If project already exists then error appears. Just change project name.
Press Start upload
button.
Link to created project and progress bars are available in output section.
App shuts down automatically on finish. Or you can stop it manually from app settings page.