Spaces:
Running
Running
# analyze_sentiment.py | |
# This script analyzes the sentiment of the summarized content using the Hugging Face Transformers library. | |
from transformers import pipeline | |
# Load zero-shot classification pipeline | |
classifier = pipeline("zero-shot-classification", model="facebook/bart-large-mnli") | |
def analyze_summary(summary): | |
""" | |
Analyze the sentiment of the given summary using zero-shot classification. | |
Returns a tuple of (sentiment, score). | |
""" | |
try: | |
if not summary.strip(): | |
return "No input provided.", 0.0 | |
candidate_labels = ["positive", "neutral", "negative"] | |
result = classifier(summary, candidate_labels) | |
sentiment = result['labels'][0].capitalize() | |
score = float(result['scores'][0]) | |
return sentiment, score | |
except Exception as e: | |
return f"Error analyzing sentiment: {str(e)}", 0.0 |