Horizon Scope is a project comparison and analysis tool that leverages vector search and AI-powered comparisons to help users discover similar projects and gain valuable insights.
Experience Horizon Scope in action! Visit my live demo:
- Vector-based Project Search 🔍:
- Utilizes Pinecone for efficient similarity searches across a database of projects.
- AI-powered Project Comparison 🤖:
- Employs OpenAI's GPT models to generate detailed comparisons between user input and existing projects.
- Web Interface 💻:
- Offers a user-friendly interface built with Streamlit.
HorizonScopeClient: Main client class interfacing with core functionalities.VectorSearchService: Manages vector-based similarity searches using Pinecone.ComparisonService: Handles AI-powered project comparisons using OpenAI.- Streamlit Web Application: Interactive interface for project comparisons.
- Input Project Descriptions ✏️: Users input project descriptions via the web interface or Python client.
- Vector Similarity Search 📈: The system searches for similar projects using Pinecone.
- AI-powered Comparisons 🔗: Provides insights on similarities, differences, and overall analysis.
- Results and Metrics 📊: Includes AI Similarity score, Cosine Similarity, and Confidence.
---
config:
theme: neo-dark
---
flowchart TD
A(["User Input: <br> query='PROJECT DESCRIPTION' <br> k=3"]) --> B("Generate Embeddings")
B <-- IN: query <br> OUT: embedding --> X_B[("OpenAI <br> text-embedding-3-small")]
B --> C("Search similar Projects")
C <-- IN: embedding, k <br> OUT: List[Project] --> X_C[("Pinecone <br> projects-text-embedding-3-small")]
C --> D{"For each Project (k)"}
D --> E["Generate ComparisonResult"]
E <-- IN: query, existing_project <br> OUT: ComparisonResult --> X_E[("OpenAI <br> gpt-4o-mini")]
E -- List[ComparisonResult] --> F(["Display ComparisonResult's"])
This tool uses project data embedded from the following sources: