Skip to main content

Download datasets

After running a computation in a Data Clean Room, you can retrieve the results in different formats depending on your needs.

Setup​

First, import the necessary modules:

from decentriq_platform import create_client
from decentriq_platform.analytics import AnalyticsDcrBuilder, PythonComputeNodeDefinition

Then, create the Client instance with which you can communicate with the Decentriq platform:

user_email = "@@ YOUR EMAIL HERE @@"
api_token = "@@ YOUR TOKEN HERE @@"

client = create_client(user_email, api_token)
enclave_specs = dq.enclave_specifications.latest()

Retrieving computation results​

Once you have published a DCR and have access to a computation node, you can retrieve the results using various methods:

# Build and publish your DCR
builder = AnalyticsDcrBuilder(client=client)

script = """with open("/output/results.txt", "w") as output_file:
    output_file.write("Computation results")
"""

dcr_definition = (
    builder.with_name("My Analytics DCR")
    .with_owner(user_email)
    .with_description("Example DCR for downloading results")
    .add_node_definitions(
        [
            PythonComputeNodeDefinition(
                name="python_node",
                script=script,
            ),
        ]
    )
    .add_participant(user_email, analyst_of=["python_node"])
    .build()
)

dcr = client.publish_analytics_dcr(dcr_definition)

# Get the computation node
node = dcr.get_node("python_node")

# Retrieve results as a ZIP file (recommended for multiple output files)
result_zip = node.run_computation_and_get_results_as_zip()

# Access files in the ZIP
for filename in result_zip.namelist():
    content = result_zip.read(filename).decode()

Download methods​

The SDK provides different methods for retrieving computation results:

Use run_computation_and_get_results_as_zip() when your computation produces multiple output files:

result_zip = node.run_computation_and_get_results_as_zip()

# List all files in the results
print(result_zip.namelist())

# Read a specific file
content = result_zip.read("results.txt").decode()

2. As bytes​

Use run_computation_and_get_results_as_bytes() when you need the raw binary data:

result_bytes = node.run_computation_and_get_results_as_bytes()

# Save to file
with open("results.zip", "wb") as f:
    f.write(result_bytes)

Working with existing DCRs​

To download results from an existing DCR:

dcr_id = dcr.id
# Retrieve an existing DCR by its ID
existing_dcr = client.retrieve_analytics_dcr(
    dcr_id=dcr_id, 
    enclave_specs=enclave_specs.values()
)

# Get the computation node and retrieve results
computation_node = existing_dcr.get_node("python_node")
results = computation_node.run_computation_and_get_results_as_zip()