Creating a GeomAI Project and Uploading Training Data

This example demonstrates how to connect to the instance, create a new project, and upload geometry files as training data.

Before you begin

Make sure you have:

• Valid SimAI credentials and organization access.

• A folder containing geometry files (.vtp or .stl format).

• The ansys-simai-core library installed.

Import necessary libraries

import os

import ansys.simai.core as asc
from ansys.simai.core.errors import NotFoundError

Configure your settings

Update these variables with your specific settings:

ORGANIZATION = "my_organization"  # Replace with your organization name
DATASET_PATH = "path/to/your/data/folder"  # Folder with .vtp or .stl files
PROJECT_NAME = "new-bracket-project"  # Your project name

Create the client

Create a client to use the PySimAI library. This client will be the entrypoint for all Generative Design objects.

simai_client = asc.SimAIClient(organization=ORGANIZATION)
geomai_client = simai_client.geomai

List all available training data:

print("\nAvailable training data:")
available_tds = geomai_client.training_data.list()

Create or retrieve a project

Try to get an existing project by name, or create it if it does not exist:

try:
    project = geomai_client.projects.get(name=PROJECT_NAME)
    print(f"Using existing project: {PROJECT_NAME}")
except NotFoundError:
    project = geomai_client.projects.create(PROJECT_NAME)
    print(f"Created new project: {PROJECT_NAME}")

print(f"Current project: {project.name}")

Upload training data to the project

Loop through all geometry files in your dataset folder and upload them. The script handles duplicates by checking if the data already exist.

print("\nUploading training data files:")
successful_uploads = 0
failed_uploads = 0

for fname in os.listdir(DATASET_PATH):
    td_name = os.path.splitext(fname)[0]
    fpath = os.path.join(DATASET_PATH, fname)

    # Skip non-geometry files
    if not fname.lower().endswith((".vtp", ".stl")):
        print(f"Skipping non-geometry file: {fname}")
        continue

    # Check if training data already exist
    existing_tds = [td for td in available_tds if td.name == td_name]
    if existing_tds:
        print(f"Training data '{fname}' already exists in the datalake. Skipping upload.")
        td = existing_tds[0]
        try:
            td.add_to_project(project)
            print(f"✓ Added existing '{fname}' to project '{project.name}'")
            successful_uploads += 1
        except Exception as e:
            print(f"✗ Failed to add existing '{fname}' to project: {e}")
            failed_uploads += 1
        continue

    # Upload new training data
    try:
        training_data = geomai_client.training_data.create_from_file(file=fpath, project=project)
        print(f"✓ Uploaded '{fname}' -> ID: {training_data.id}")
        successful_uploads += 1
    except Exception as e:
        print(f"✗ Failed to upload '{fname}': {e}")
        failed_uploads += 1

print(f"\nUpload summary: {successful_uploads} successful, {failed_uploads} failed")

Check and wait for data processing

After uploading, the instance needs to process the geometries. This script displays the progress of the data processing.

project_data = project.data()


print("\nWaiting for data processing to complete...")
for data in project_data:
    print(f"Processing '{data.name}'...")
    data.wait()
    if data.is_ready:
        print(f"✓ Data '{data.name}' is ready")
    else:
        print(f"✗ Data '{data.name}' failed: {data.failure_reason}")

Display project status summary

Show a summary of the project’s data processing status:

project_data = project.data()

print("\nProject Summary")
print("=" * 50)
ready_data = [data for data in project_data if data.is_ready]
not_ready_data = [data for data in project_data if not data.is_ready]

print(f"Total data in project: {len(project_data)}")
print(f"Ready data: {len(ready_data)} of {len(project_data)}")
print(f"Not ready data: {len(not_ready_data)} of {len(project_data)}")

if not_ready_data:
    print(
        "\nFailed data processing details:\n"
        "Having an 'invalid geometry' means that the geometry is not compatible with Generative Design. "
        "Please check the geometry file for errors or issues (watertightness and manifold).\n"
    )
    for data in not_ready_data:
        print(f"- {data.name}: {data.failure_reason}")

Next steps

Once all data is ready, you can proceed to build a model. See the next example: Building a Generative Design Model.