Converting Kili label format to GeoJSON

In this recipe, we will see how to convert Kili label format to GeoJSON format.

First, let's install and import the required libraries:

%pip install kili

import json
import urllib.request

from kili.client import Kili

kili = Kili(
    # api_endpoint="https://cloud.kili-technology.com/api/label/v2/graphql",
    # the line above can be uncommented and changed if you are working with an on-premise version of Kili
)

Data

For this tutorial, we will use a GeoTiff file created from Copernicus Sentinel-2A data.

urllib.request.urlretrieve(
    "https://github.com/mommermi/geotiff_sample/raw/master/sample.tif", "sample.tif"
)

Kili project creation

We will create a project with a single asset, the GeoTiff file we just downloaded.

json_interface = {
    "jobs": {
        "BBOX_DETECTION_JOB": {
            "content": {
                "categories": {"B_BOX_A": {"children": [], "color": "#472CED", "name": "BBox A"}},
                "input": "radio",
            },
            "instruction": "BBox job",
            "mlTask": "OBJECT_DETECTION",
            "required": 1,
            "tools": ["rectangle"],
            "isChild": False,
        },
        "POINT_DETECTION_JOB": {
            "content": {
                "categories": {"POINT_A": {"children": [], "color": "#D33BCE", "name": "Point A"}},
                "input": "radio",
            },
            "instruction": "Point job",
            "mlTask": "OBJECT_DETECTION",
            "required": 1,
            "tools": ["marker"],
            "isChild": False,
        },
        "POLYGON_DETECTION_JOB": {
            "content": {
                "categories": {
                    "POLYGON_A": {"children": [], "color": "#3BCADB", "name": "Polygon A"}
                },
                "input": "radio",
            },
            "instruction": "Polygon job",
            "mlTask": "OBJECT_DETECTION",
            "required": 1,
            "tools": ["polygon"],
            "isChild": False,
        },
        "LINE_DETECTION_JOB": {
            "content": {
                "categories": {"LINE_A": {"children": [], "color": "#5CE7B7", "name": "Line A"}},
                "input": "radio",
            },
            "instruction": "Line job",
            "mlTask": "OBJECT_DETECTION",
            "required": 1,
            "tools": ["polyline"],
            "isChild": False,
        },
        "SEGMENTATION_JOB": {
            "content": {
                "categories": {
                    "SEGMENTATION_A": {"children": [], "color": "#FB753C", "name": "Segmentation A"}
                },
                "input": "radio",
            },
            "instruction": "Segmentation job",
            "mlTask": "OBJECT_DETECTION",
            "required": 1,
            "tools": ["semantic"],
            "isChild": False,
        },
    }
}

project_id = kili.create_project(
    input_type="IMAGE", title="[Kili SDK Notebook]: Geojson tutorial", json_interface=json_interface
)["id"]

kili.append_many_to_dataset(project_id, content_array=["sample.tif"], external_id_array=["sample"])

At this point, we can visualize the asset in Kili: