Utility to visualize coordinates-detecting ML models results

🐳 Docker image

Checkout image on Docker Hub: https://hub.docker.com/r/deniskrumko/visual-coordinates-tool

docker pull deniskrumko/visual-coordinates-tool:latest

How to run in docker

To run services with configuration - create config folder with .toml configuration file and mount it to /usr/local/bin/config (config example is bellow in README):

docker run --rm -v $(pwd)/config/:/usr/local/bin/config -p 8080:8080 \
    deniskrumko/visual-coordinates-tool:latest --config config/test.toml

To run services without configuration:

docker run --rm -p 8080:8080 deniskrumko/visual-coordinates-tool:latest

⚒️ How to use

Define request parameters

Get response with coordinates

Request configuration

JSON request

If your service accepts JSON requests with image URL then specify template using which request can be performed. For example: {"inputs": "%s"}.

NOTE: %s is used to define where image URL will be passed.

Also you need to check JSON Request checkbox in near Get coordinates button.

Form data request

If your service accepts form-data requests, then put field name in "Request" section. For example: photo.

Also you need to uncheck JSON Request checkbox in near Get coordinates button.

Response configuration

For example, you service responds with JSON like this:

{
    "success": true,
    "data": {
        "coordinates": [
            {"x": 100, "y": 200},
            {"x": 200, "y": 300}
        ]
    }
}

Then you need to specify in Response section that field data.coordinates contains pairs of x-y coordinates.

NOTE: Currently visual-coordinates-tool supports only JSON responses and only pairs of {"x": ..., "y": ...} coordinates.

✅ Config example

[common]
# Name of the service to display in UI
name = "Cars Coordinates Testing"

# List of predefined service configurations to use
[services]

[services.my-service-1]
name = 'Service that uses JSON'
group = 'My services'
url = 'https://path-to-endpoint'
request_json_template = '{"inputs": "%s"}'
response_xy_field = 'coords'

[services.my-service-2]
name = 'Service that uses formdata'
group = 'My services'
url = 'http://path-to-endpoint'
request_formdata_field = 'photo'
response_xy_field = 'result.coords'
default = true

# List of image samples
[samples]
prefix = 'https://images-bucket-example.com/'
images = [
    'cars/1.jpg',
    'cars/2.jpg',
    'cars/3.jpg',
    '-',  # will act as divider
    'trucks/4.jpg',
    'trucks/5.jpg',
]

Service fields:

name - service name
group - group name
url - path to endpoint
request_json_template – predefined JSON request template
request_formdata_field – predefined form-data request field
response_xy_field – predefined response field with coordinates
default – use this service by default or not