List Evaluation Reports

Retrieve a list of evaluation reports with filtering and pagination

Overview

The list_evaluation_reports method allows you to retrieve a list of evaluation reports with optional filtering and pagination. This is useful for monitoring evaluation history and tracking performance over time.

Method Signature

Synchronous

def list_evaluation_reports(
    dataset_id: Optional[str] = None,
    status: Optional[str] = None,
    limit: Optional[int] = None,
    offset: Optional[int] = None
) -> Dict[str, Any]

Asynchronous

async def list_evaluation_reports(
    dataset_id: Optional[str] = None,
    status: Optional[str] = None,
    limit: Optional[int] = None,
    offset: Optional[int] = None
) -> Dict[str, Any]

Parameters

Parameter	Type	Required	Description
dataset_id	str	No	Filter by specific dataset ID
status	str	No	Filter by evaluation status (running, completed, failed)
limit	int	No	Maximum number of reports to return (default: 50)
offset	int	No	Number of reports to skip for pagination (default: 0)

Returns

Returns a dictionary containing the list of evaluation reports and pagination information.

Examples

Basic Usage

from keywordsai import KeywordsAI

client = KeywordsAI(api_key="your-api-key")

# List all evaluation reports
reports = client.datasets.list_evaluation_reports()

print(f"Found {len(reports['reports'])} evaluation reports")
for report in reports['reports']:
    print(f"ID: {report['evaluation_id']}, Status: {report['status']}")

Filter by Dataset

# List evaluations for a specific dataset
reports = client.datasets.list_evaluation_reports(
    dataset_id="dataset_123"
)

print(f"Found {len(reports['reports'])} evaluations for dataset_123")

Filter by Status

# List only completed evaluations
completed_reports = client.datasets.list_evaluation_reports(
    status="completed"
)

print(f"Found {len(completed_reports['reports'])} completed evaluations")

# List running evaluations
running_reports = client.datasets.list_evaluation_reports(
    status="running"
)

print(f"Found {len(running_reports['reports'])} running evaluations")

With Pagination

# Get first 10 reports
reports = client.datasets.list_evaluation_reports(
    limit=10,
    offset=0
)

print(f"Page 1: {len(reports['reports'])} reports")

# Get next 10 reports
next_reports = client.datasets.list_evaluation_reports(
    limit=10,
    offset=10
)

print(f"Page 2: {len(next_reports['reports'])} reports")

Asynchronous Usage

import asyncio
from keywordsai import AsyncKeywordsAI

async def list_reports_example():
    client = AsyncKeywordsAI(api_key="your-api-key")
    
    reports = await client.datasets.list_evaluation_reports(
        dataset_id="dataset_123",
        status="completed",
        limit=20
    )
    
    print(f"Retrieved {len(reports['reports'])} completed reports")
    return reports

asyncio.run(list_reports_example())

Combined Filtering

# Filter by both dataset and status
reports = client.datasets.list_evaluation_reports(
    dataset_id="dataset_123",
    status="completed",
    limit=5
)

print(f"Latest 5 completed evaluations for dataset_123:")
for report in reports['reports']:
    print(f"  {report['evaluation_id']}: {report['overall_score']}")

Error Handling

try:
    reports = client.datasets.list_evaluation_reports(
        dataset_id="dataset_123"
    )
    print(f"Successfully retrieved {len(reports['reports'])} reports")
except Exception as e:
    print(f"Error listing evaluation reports: {e}")

Response Structure

The response includes:

• reports: List of evaluation report summaries
• total_count: Total number of reports matching filters
• has_more: Whether more reports are available
• next_offset: Offset for the next page

Each report summary contains:

• evaluation_id: Unique identifier
• dataset_id: Associated dataset ID
• status: Current status
• overall_score: Aggregate score (if completed)
• created_at: Creation timestamp
• completed_at: Completion timestamp (if applicable)

Common Use Cases

• Monitoring evaluation history across datasets
• Tracking model performance trends
• Finding failed evaluations for debugging
• Generating performance dashboards
• Audit an