Script 2 - Generating new designs with the model#

This example demonstrates how to generate new designs.

Before you begin#

or

  • Make sure you:
    1. instantiated a client with GeomAI objects,
    2. created a project and added training data to it,
    3. created a model configuration and built a model.

Import necessary libraries#

import ansys.simai.core
from ansys.simai.core.data.geomai.predictions import GeomAIPredictionConfiguration

Create the client#

Create a client to use the PySimAI library. This client will be the entrypoint of all “SimAI” and “GeomAI” objects.

simai = ansys.simai.core.SimAIClient(organization="my_organization")
client = simai.geomai

Retrieve your workspace#

Find the workspace your model has been saved to and set it as a current workspace:

Step 1. Set the current project:

client.set_current_project("new-bracket-project")

Step 2. Retrieve all the workspaces of the current project:

print(client.current_project.workspaces())

Step 3. Setting the current workspace:

client.set_current_workspace("new-bracket-project #1")
workspace = client.current_workspace
print(workspace)

Set up the parameters and run your prediction#

Configure the prediction:

configuration = GeomAIPredictionConfiguration(latent_params=[1, 1], resolution=(100, 100, 100))

Run the prediction:

prediction = client.predictions.run(configuration, workspace)

Check the status of your prediction:

if prediction.is_ready:
    print("✅ Prediction is ready!")
else:
    print("⏳ Prediction is not ready yet.")

print(prediction)

Download your prediction for further analysis and visualization#

To download your file, please indicate the ID of the prediction and the path to the desired storage location.

print(client.predictions.list(client.workspaces.list()[0]))

client.predictions.download("your_prediction_id", r"path/to/your/predictions/folder/output1.vtp")

The downloaded geometry can be now used as a SimAI training data or as an input for the SimAI prediction!

Gallery generated by Sphinx-Gallery