External storage
Last updated
Last updated
When dealing with a large volume of files, external storage can be used. This storage may take the form of a separate server.
It is IMPOSSIBLE to change the storage type, once it is chosen.
Our platform integrates with S3 and GCP storage to allow you to annotate large amounts of files.
If you are working with images, you need to follow these steps:
Go to the FILES tab and click 'Add files' button
Fill in the mandatory fields:
File: the name of the file. Fill in the name under which the file of stored in the storage.
Width and Height: fill in the size of your image in pixels.
Start frame, Frame step, and Frames: fill in this these field with "1" value if you are working with images.
You can upload the same values with the help of CSV.
The last step here is to create Files list - click 'Create' button.
If you are working with videos, you need to follow these steps:
Go to the FILES tab and click 'Add files' button
Fill in the mandatory fields:
File: the name of the file. Fill in the name under which the file of stored in the storage.
Width and Height: fill in the size of your image in pixels.
Start frame: the number of the frame you want to be the first one in the editor (If you want to start annotating your video from the 5th frame, you can enter "5" in this field and the platform will omit the first four frames).
Frame step: this field allows you to adjust FPS value. For example, if you enter "10" in this field, the platform will upload each 10th frame of the video. To keep original frame rate just enter "1" in this field.
Frames: the total amount of frames in the video.
Frame counts in source video: this field is filled in if video is split into segments and will be merged later.
You can upload the same values with the help of CSV.
The last step here is to create Files list - click 'Create' button.
Now, you need to create a path for the platform to find the files in the storage. To do this, you need to create a URL and use it in the Configuration tab.
The image URL consists of three parts. The first part is https://[instance_name].keylabs.ai:8200
Then, we need to create the second part - a prefix.
Prefixes are created by Keylabs technical team. You need to provide them with the following credentials:
For s3:
bucket name;
bucker region;
access key;
secret key.
For GCP:
You need to manage the access to your storage - provide us with access mode where we can download data.
Provide us with path to where the data is stored
It is unique for every project and links the platform and the storage. It contains the storage name (s3 or gcp):
and the the part that leads to the bucket where the data is stored:
The following variable is {frame:5} which will be replaced with the frame names i.e., the numeric value of the frame. If we put number 5 in the variable, the frame number must consist of 5 digits, e.g. 00001, 00002, if we put "3" in {frame:3} - the frame number must consist of 3 digits, e.g. 001, 012. This number can vary but must be consistent throughout the project.
The {frame:5} variable is required only when you work with video. When working with images the link will look like this: https://tutorials.keylabs.ai:8200/s3/test/{file}
The last part of the link is the file format. It depends on the format of the files you want to use.
To ensure your image is not smaller (see example) or larger than the editor, carefully fill in the width and height values.
To ensure your image is not smaller (see example) or larger than the editor, carefully fill in the width and height values.
Next part of the link is variable {file} . Platform will replace this variable with the file name
