wiba 0.2.2

Official Python client library for WIBA argument mining API

WIBA Python Client

Official Python client library for the WIBA (What Is Being Argued?) argument mining API.

🚀 Quick Start

Installation

pip install wiba

Basic Usage

from wiba import WIBA

# Initialize client
client = WIBA(api_token="your_api_token")

# Detect arguments
result = client.detect("Climate change requires immediate action.")
print(result.is_argument)  # True
print(result.confidence)   # 0.89

# Extract topics
topic = client.extract("We must invest in renewable energy now.")
print(topic.topic)  # "renewable energy"

# Analyze stance
stance = client.stance("Solar power is expensive", "renewable energy")
print(stance.stance)  # "Against"

# Discover arguments in longer text
segments = client.discover_arguments(long_text)
for segment in segments:
    print(f"Argument: {segment.text} (confidence: {segment.confidence})")

Batch Processing

# Process multiple texts
texts = [
    "Climate change is a serious threat.",
    "We need renewable energy sources.",
    "Nuclear power is too dangerous."
]

# Batch argument detection
results = client.detect(texts)
for result in results:
    print(f"{result.text}: {result.is_argument}")

# Process DataFrames
import pandas as pd
df = pd.DataFrame({'text': texts})
df_results = client.detect(df)
print(df_results[['text', 'is_argument', 'confidence']])

🔧 Configuration

API Token

Get your API token from wiba.dev:

# Using API token
client = WIBA(api_token="your_token_here")

# Using environment variable
import os
os.environ['WIBA_API_TOKEN'] = 'your_token_here'
client = WIBA()  # Will use WIBA_API_TOKEN automatically

Custom Configuration

from wiba import WIBA, ClientConfig

config = ClientConfig(
    api_url="https://custom.wiba.dev",
    api_token="your_token",
    log_level="DEBUG"
)

client = WIBA(config=config)

📊 Features

Core Functions

• detect(texts) - Argument detection in text
• extract(texts) - Topic extraction from arguments
• stance(texts, topics) - Stance analysis (favor/against/neutral)
• discover_arguments(text) - Find argumentative segments in longer texts

Input Formats

• Single text: "This is an argument"
• List of texts: ["Text 1", "Text 2", "Text 3"]
• pandas DataFrame: DataFrame with text column
• CSV string: Comma-separated values

Response Objects

All methods return structured response objects with:

• Results: List of prediction results
• Metadata: Request information and statistics
• Confidence scores: Model confidence for each prediction

Advanced Features

• Batch processing with progress bars
• Automatic retries with exponential backoff
• Connection pooling for better performance
• DataFrame integration for data science workflows
• Statistics tracking for usage monitoring

🧪 Examples

Check out the examples/ directory for:

• Basic usage examples
• Batch processing demonstrations
• DataFrame integration
• Advanced configuration
• Error handling patterns

🤝 Contributing

This package is part of the WIBA-ORG collaborative development setup:

1. Report bugs: GitHub Issues
2. Contribute code: See CONTRIBUTING.md
3. Documentation: Help improve our docs

📄 License

MIT License - see LICENSE file for details.

🔗 Links

• Homepage: wiba.dev
• API Documentation: wiba.dev/docs
• Research Paper: [Link to paper]
• GitHub Organization: WIBA-ORG