CoolFace
Apppublic

ashwinpatti/Balanced-News-Perspective-App

sourceHugging Facemitupdated 2y agoView on Hugging Face
2likes
App README

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Balanced-News-Perspective

This repository contains the source code for the Balanced News Perspective App, an innovative tool designed to provide users with diverse perspectives on news headlines. The app takes a headline as input and delivers three distinct views: right-leaning, left-leaning, and center/unbiased. By presenting these perspectives, the app ensures users gain a well-rounded understanding of the news, fostering critical thinking and informed discussions.

Features:

  • Access news from multiple viewpoints to understand the full picture.
  • User-friendly interface for easy input and balanced perspective retrieval.
  • Enhance critical thinking by comparing and contrasting different viewpoints.
  • Stay updated with the latest news while avoiding echo chambers and bias.
  • Valuable tool for debate preparation, providing access to opposing viewpoints and sufficient information to participate effectively.

Technology Stack:

  • Large Language Models (LLMs): To generate diverse perspectives and summaries.
  • MultiAgents Framework Agno: A lightweight library for building multi-modal Agents.
  • Monitoring/Observalibility: Using agno's in build evaluator (screenshot attached).
  • Vector Database Qdrant: To efficiently store and retrieve vectorized data for fast and accurate searches.
Monitoring screenshot:

<img width="1465" alt="Screenshot 2025-02-25 at 11 13 06 PM" src="https://github.com/user-attachments/assets/079977e0-7a53-4bde-8cc3-4d0607e91fb8" />

Getting Started:

  1. 1.Clone the repository.
  2. 2.Install the necessary dependencies by running uv sync
  3. 3.Run the app locally or deploy it to your preferred platform.

Future enhancements:

  1. 1.Ability to share the final report as a pdf.
  2. 2.Ability to share the final report as a social media post.
  3. 3.Chat feature where users can ask more questions.
  4. 4.Fine-tuned embedding model to enhance chatting feature.
  5. 5.A validation agent to validate other agents.

Contributing:

We welcome contributions from the community. Please read the contributing guidelines before submitting a pull request.

Architecture diagram:

WhatsApp Image 2025-02-26 at 01 19 55

License:

This project is licensed under the MIT License.