Create a Dataset

With the Platform

Creating a Dataset is really simple, once on your "Datalake" page

Search for pictures

To perform operations, your need to select assets on the table, the right checkbox select all the visible assets, the left one select all filtered assets.
You can search pictures on your lake with our Data Query Language, basically you can search pictures in your lake with :
  • Tags
  • Width
  • Height
  • Source
  • Filename
  • Annotations
  • Dataset
For example, let's seach for pictures that are tagged penfun with a Picture width > 400 px and that have at least one annotation.
Once you have filtered your pictures, you can either select all filtered pictures or a subset.
You can now click on create dataset !
Then enter a Name and description for your Dataset, please note that description is optional :)

With Python SDK

1
pip install picsellia
Copied!
First make sure that you have Picsellia Python package installed
then you will need to initialize the Client with your API Token, available in you profile page.
1
from picsellia.client import Client
2
clt = Client(api_token="your token")
Copied!
you can now search for some assets on your lake with the datalake.fetch() method:
1
pictures = clt.datalake.picture.fetch(quantity=1, tags=['tag1'])
Copied!
You can use Client.datalake.pictures.status() to vizualize the fetched assets
then you can create your dataset
1
clt.datalake.dataset.create(name='dataset2',
2
description='this is a test',
3
pictures=pictures)
4
Copied!
You can find a complete reference to the SDK here.
Last modified 4mo ago