BReact SDK

A Python SDK for interacting with BReact's AI services, supporting both synchronous and asynchronous operations.

Installation

pip install breactsdk

Documentation

See the full documentation at https://github.com/BReact/BReact-sdk

Configuration

The SDK can be configured using environment variables or programmatically:

Environment Variables

Create a .env file in your project root:

BREACT_API_KEY=your_api_key
BREACT_BASE_URL=https://api-os.breact.ai  # Optional, defaults to this URL

Programmatic Configuration

from breactsdk.client import create_client

# Create client with custom configuration
client = create_client(
    api_key="your_api_key",  # If not provided, the SDK will use the one from the environment variable
    base_url="https://api-os.breact.ai"  # Optional
)

Usage Examples

Text Summarization

from breactsdk.client import create_client
import asyncio

# Initialize the client
client = create_client(
    api_key="your_api_key",
    base_url="https://api-os.breact.ai",
    async_client=True
)
    
async def summarize_text():
    try:
        async with client:
            # Generate text summary
            result = await client.summary.summarize(
                model_id="openai/gpt-4o",
                text='''One morning, when Gregor Samsa woke from troubled dreams, he found himself transformed in his bed into a horrible vermin. He lay on his armour-like back, and if he lifted his head a little he could see his brown belly, slightly domed and divided by arches into stiff sections. The bedding was hardly able to cover it and seemed ready to slide off any moment. His many legs, pitifully thin compared with the size of the rest of him, waved about helplessly as he looked. "What's happened to me?" he thought. It wasn't a dream.''',
                summary_type="brief",  # Options: "brief", "detailed", "bullet_points", "executive"
                max_words=200,  # Optional: control summary length
                output_format="json"  # Options: "paragraph", "bullets", "json"
            )
            
            print(result)
    except Exception as e:
        print(f"Error occurred: {str(e)}")
        if hasattr(e, 'response'):
            print(f"Status code: {e.response.status_code}")
            print(f"Error details: {e.response.text}")

asyncio.run(summarize_text())
 

Email Analysis and Response Generation

async with create_client(async_client=True) as client:
    # Analyze email thread
    analysis = await client.email_response.analyze_thread(
        email_thread=[{
            "sender": "client@example.com",
            "recipient": "support@company.com",
            "subject": "Urgent: Service Downtime",
            "content": "Email content here",
            "timestamp": "2024-01-20T09:00:00Z"
        }],
        analysis_type=["sentiment", "key_points", "action_items", "response_urgency"]
    )

    # Generate email response
    response = await client.email_response.generate_response(
        email_thread=[{
            "sender": "client@example.com",
            "recipient": "support@company.com",
            "subject": "Product Feature Inquiry",
            "content": "Email content here",
            "timestamp": "2024-01-20T10:30:00Z"
        }],
        tone="friendly",
        style_guide={
            "language": "en",
            "max_length": 150,
            "greeting_style": "casual",
            "signature": "\nBest regards,\nSupport Team"
        },
        key_points=[
            "Address AI capabilities",
            "Explain pricing plans",
            "Highlight support options"
        ]
    )

Information Tracking

# Define your schema
schema = {
    "type": "object",
    "properties": {
        "primary_symptom": {
            "type": "string",
            "enum": ["headache", "nausea", "dizziness"]
        },
        "duration": {
            "type": "string"
        }
    },
    "required": ["primary_symptom", "duration"]
}

# Process information
async with create_client(async_client=True) as client:
    result = await client.information_tracker.process(
        content="Your text content",
        context={
            "updateType": "medical_symptoms",
            "currentInfo": {
                "previous_symptoms": ["mild headache"]
            }
        },
        config={
            "modelId": "mistral-large-2411",
            "temperature": 0.1,
            "maxTokens": 2000,
            "schema": schema
        }
    )

Concurrent Processing

async with create_client(async_client=True) as client:
    tasks = [
        client.aisummary.summarize(
            text=f"Text {i}",
            summary_type="executive",
            model_id="mistral-small"
        ) for i in range(3)
    ]
    
    results = await asyncio.gather(*tasks)

Error Handling

The SDK provides detailed error information:

try:
    result = await client.aisummary.summarize(text="Your text")
except Exception as e:
    if hasattr(e, 'response'):
        print(f"Status code: {e.response.status_code}")
        print(f"Error details: {e.response.text}")
    print(f"Error: {str(e)}")

Available Services

1. AI Summary (client.aisummary)

   • summarize: Generate text summaries

2. Email Response (client.email_response)

   • analyze_thread: Analyze email threads
   • generate_response: Generate email responses

3. Information Tracker (client.information_tracker)

   • process: Extract structured information from text

Best Practices

1. Always use context managers (with or async with) to ensure proper resource cleanup
2. Choose between sync and async clients based on your application's needs
3. Set appropriate timeouts and model parameters for your use case
4. Handle errors appropriately in production code
5. Store API keys securely using environment variables

Running the Demo

A comprehensive demo script is included that showcases all features:

python demo.py

Support

For issues and feature requests, please contact office@breact.ai