Exports allow you to download labeled data from projects or datasets. The Avala SDK provides methods to create exports, monitor their progress, and retrieve results.
Creating an Export
Export from a Project
Export all labeled data from a specific project:
from avala import Avala
client = Avala(api_key="your-api-key")
# Create export for a project
export = client.exports.create(project="proj_abc123")
print(f"Export created: {export.uid}")
print(f"Status: {export.status}")
Export from a Dataset
Export data from a dataset:
# Create export for a dataset
export = client.exports.create(dataset="ds_xyz789")
print(f"Export created: {export.uid}")
Export Both Project and Dataset
# Export with both project and dataset specified
export = client.exports.create(
project="proj_abc123",
dataset="ds_xyz789"
)
You can specify either project, dataset, or both when creating an export.
Checking Export Status
Exports are processed asynchronously. Check the status before downloading:
Create the export
export = client.exports.create(project="proj_abc123")
print(f"Export UID: {export.uid}")
Poll for completion
import time
while True:
export = client.exports.get(export.uid)
print(f"Status: {export.status}")
if export.status == "completed":
print("Export ready!")
break
elif export.status == "failed":
print(f"Export failed: {export.error}")
break
time.sleep(5) # Wait 5 seconds before checking again
Download the results
if export.status == "completed":
print(f"Download URL: {export.download_url}")
print(f"Expires at: {export.expires_at}")
Export Status Values
Exports go through several status stages:
pending: Export has been queuedprocessing: Export is being generatedcompleted: Export is ready for downloadfailed: Export encountered an error
Downloading Export Results
Once an export is completed, download the results:
import requests
# Get the export
export = client.exports.get("exp_123abc")
if export.status == "completed":
# Download the file
response = requests.get(export.download_url)
with open("export_data.zip", "wb") as f:
f.write(response.content)
print("Export downloaded successfully")
Download URLs are temporary and expire after a certain period (typically 24 hours). Check the expires_at field.
Listing Exports
View all exports created by your account:
# List all exports
exports = client.exports.list()
for export in exports:
print(f"{export.uid} - Status: {export.status}")
# Get exports with pagination
page = client.exports.list(limit=50)
for export in page:
print(f"Export {export.uid}: {export.status}")
# Get next page
if page.has_next:
next_page = client.exports.list(
cursor=page.next_cursor,
limit=50
)
Complete Export Workflow
import time
import requests
from avala import Avala
client = Avala(api_key="your-api-key")
# Create export
print("Creating export...")
export = client.exports.create(project="proj_abc123")
print(f"Export UID: {export.uid}")
# Wait for completion
print("Waiting for export to complete...")
while True:
export = client.exports.get(export.uid)
print(f"Status: {export.status}")
if export.status == "completed":
break
elif export.status == "failed":
print(f"Export failed: {export.error}")
exit(1)
time.sleep(5)
# Download results
print("Downloading export...")
response = requests.get(export.download_url)
with open("project_export.zip", "wb") as f:
f.write(response.content)
print("Export downloaded to project_export.zip")
import asyncio
import aiohttp
from avala import AsyncAvala
async def download_export():
client = AsyncAvala(api_key="your-api-key")
# Create export
print("Creating export...")
export = await client.exports.create(project="proj_abc123")
print(f"Export UID: {export.uid}")
# Wait for completion
print("Waiting for export to complete...")
while True:
export = await client.exports.get(export.uid)
print(f"Status: {export.status}")
if export.status == "completed":
break
elif export.status == "failed":
print(f"Export failed: {export.error}")
return
await asyncio.sleep(5)
# Download results
print("Downloading export...")
async with aiohttp.ClientSession() as session:
async with session.get(export.download_url) as response:
with open("project_export.zip", "wb") as f:
f.write(await response.read())
print("Export downloaded to project_export.zip")
asyncio.run(download_export())
Advanced: Automated Export Monitoring
Create a reusable function to monitor exports:
import time
from typing import Optional
def wait_for_export(
client,
export_uid: str,
poll_interval: int = 5,
timeout: Optional[int] = 300
):
"""Wait for an export to complete.
Args:
client: Avala client instance
export_uid: UID of the export to monitor
poll_interval: Seconds to wait between status checks
timeout: Maximum seconds to wait (None for no timeout)
Returns:
Export object when completed
Raises:
TimeoutError: If export doesn't complete within timeout
RuntimeError: If export fails
"""
start_time = time.time()
while True:
export = client.exports.get(export_uid)
if export.status == "completed":
return export
elif export.status == "failed":
raise RuntimeError(f"Export failed: {export.error}")
# Check timeout
if timeout and (time.time() - start_time) > timeout:
raise TimeoutError(f"Export did not complete within {timeout}s")
time.sleep(poll_interval)
# Usage
export = client.exports.create(project="proj_abc123")
print(f"Waiting for export {export.uid}...")
try:
completed_export = wait_for_export(client, export.uid, timeout=600)
print(f"Export ready: {completed_export.download_url}")
except TimeoutError as e:
print(f"Timeout: {e}")
except RuntimeError as e:
print(f"Error: {e}")
- Labeled annotations in JSON format
- Original data files (if applicable)
- Metadata and task information
- Project or dataset configuration
The export format depends on your project type and configuration. Check the downloaded archive for a README or schema file.
Best Practices
- Poll export status at reasonable intervals (5-10 seconds) to avoid rate limits
- Set timeouts when waiting for exports to prevent indefinite waiting
- Download exports promptly before the download URL expires
- Store export UIDs for tracking and auditing purposes
Large exports may take several minutes to process. Always implement timeout handling in production code.
Troubleshooting
If an export fails:
- Check the
error field on the export object for details
- Verify that the project or dataset exists and is accessible
- Ensure your API key has permission to export data
- Try creating a new export if the failure was temporary