usage.md

Detailed Usage Guide

1. Basic Paper Analysis

Using Interactive Tool

python my_analysis.py

Example session:

ArXiv Paper Analysis
==================================================

Enter your search query: quantum computing
How many papers to analyze? [50]: 25

Fetching papers...
Generating embeddings...
Clustering papers...

Found 25 papers in 4 clusters

Cluster 0 (8 papers):
- Title: Quantum Computing with Neutral Atoms
  Authors: John Smith, Jane Doe
  Published: 2024-01-15
...

Search Query Examples

1. Basic topic search:

   machine learning
   

2. Multiple topics:

   quantum computing AND artificial intelligence
   

3. Title-specific search:

   ti:"deep learning"
   

4. Author search:

   au:Smith
   

2. Detailed Analysis

For more comprehensive analysis:

python scripts/analyze_papers.py

This provides:

• Detailed clustering analysis
• Visualizations
• Saved results in outputs directory

3. Quick Start

For a simple predefined analysis:

python quick_start.py

4. Output Files

Results are saved in:

outputs/
├── papers_[timestamp]/
│   ├── papers.json        # Raw paper data
│   └── metadata.json      # Query information
├── clusters_[timestamp]/
│   └── cluster_data.json  # Clustering results
└── visualizations/
    └── clusters.html      # Interactive visualization

5. Advanced Features

Custom Analysis

from src.api.arxiv_api import ArxivAPI
from src.embedding_generator import EnhancedEmbeddingGenerator

# Initialize
api = ArxivAPI()
generator = EnhancedEmbeddingGenerator()

# Fetch papers
papers = api.fetch_papers_batch("your query", max_papers=50)

# Process
embeddings = generator.generate_embeddings([p['summary'] for p in papers])

6. Best Practices

1. Start Small

   • Begin with 25-50 papers
   • Increase if needed

2. Refine Searches

   • Use specific terms
   • Add category filters
   • Combine search terms

3. Save Results

   • Results automatically saved
   • Check outputs directory

7. Troubleshooting

1. No Results

   • Broaden search terms
   • Check internet connection
   • Verify query syntax

2. Slow Performance

   • Reduce number of papers
   • Close other applications
   • Check memory usage

3. Installation Issues

   • Use Python 3.11
   • Create fresh virtual environment
   • Update pip and dependencies