GCP Cloud Functions

The following example shows the general structure of how to build a GCP cloud function. For concrete implementations of agents with specific abilities, see the examples section.

STEP 1: Create a Project

Create a new Encord Project:

mkdir my_project
cd my_project

Create and source a new virtual environment.

python -m venv venv
source venv/bin/activate

Create a requirements file.

requirements.txt

functions-framework
encord-agents

Install dependencies.

python -m pip install -r requirements.txt

STEP 2: Define the Agent

Create a main.py file using the following template:

main.py

from encord.objects.ontology_labels_impl import LabelRowV2

from encord_agents.core.data_model import FrameData
from encord_agents.gcp import editor_agent


@editor_agent()
def my_agent(frame_data: FrameData, label_row: LabelRowV2) -> None:
    ...
    # label_row.save()

Complete the my_agent function with the logic you want to execute when the agent is triggered.

You can inject multiple different dependencies into the function if necessary.

You can find multiple examples of what can be done with Custom Agents here.

STEP 3: Test the Agent

Trigger the agent by running it locally.

ENCORD_SSH_KEY_FILE=/path/to/your_private_key \
    functions-framework --target=my_agent --debug --source main.py

This means starting an API at localhost:8080/my_agent that expects a POST request with JSON data with the following format:

{
    "projectHash": "<project_hash>",
    "dataHash": "<data_hash>",
    "frame": <frame_number>
}

To test the agent endpoint, open the Label Editor in your browser on a frame where you want to run the agent. Then, copy the URL. Open a new terminal in the my_project directory and run:

source venv/bin/activate
encord-agents test local my_agent '<the_pasted_url>'

Notice the single quotes around <the_pasted_url>. They are important and should be there because you might copy a url with, e.g., an & character that have a special meaning if it is not within a string (or escaped).

Refresh the Label Editor in your browser to see the effect.

STEP 4: Deployment

To go from development to production, you must deploy your agent on the Google infrastructure. The following example shows how you can deploy the agent to the cloud. Notice how secrets (the access key that the agent should use) are set.

gcloud functions deploy my_agent \
    --entry-point my_agent \
    --runtime python311 \
    --trigger-http \
    --allow-unauthenticated \
    --gen2 \
    --region europe-west2 \
    --set-secrets="ENCORD_SSH_KEY=SERVICE_ACCOUNT_KEY:latest"

See the official Google run function deploy docs for more information. There are a couple of things that you need to pay attention to:

You must make sure to authenticate gcloud and select the appropriate project first
You should configure a secret with the ssh_key content. See Google Secrets docs

Get Started

Quick Start (By Role)

Data

Fundamentals

Annotate & Review

Label and Model Evaluation

Users and Workspaces

GCP Cloud Functions

STEP 1: Create a Project

STEP 2: Define the Agent

STEP 3: Test the Agent

STEP 4: Deployment

​STEP 1: Create a Project

​STEP 2: Define the Agent

​STEP 3: Test the Agent

​STEP 4: Deployment

STEP 1: Create a Project

STEP 2: Define the Agent

STEP 3: Test the Agent

STEP 4: Deployment