v1.3.1 Nvidia NIM Migration

v1.3.1

Fixes NVIDIA NIM support and improves multi-provider runtime stability.

Added

• NVIDIA NIM integration
• Provider switching
• Llama 3.1 70B & 405B support
• Batch PDF processing

Improvements

• Fixed NIM runtime issues
• Faster pipeline and better fallback handling

v1.3.0 Nvidia NIM Migration

v1.3.0

Introduces NVIDIA NIM support with a configurable multi-provider AI runtime.

Added

• NVIDIA NIM integration
• Dynamic provider switching
• Llama 3.1 70B & 405B support
• Retry and rate limiting
• Batch PDF processing
• Integration tests and benchmarking utilities

Improvements

• Faster classification pipeline
• Better scalability and resilience
• Reduced vendor lock-in
• Config-driven deployment workflow

Configuration

LLM_PROVIDER=nim

or

LLM_PROVIDER=gemini

Version 1.2.0

✨ What's New

Model Configuration Updates

• Upgraded to Gemini 2.0 Experimental Models
  • Classifier now uses gemini-2.0-flash-exp for fast, cost-effective question classification
  • Analyzer now uses gemini-2.0-pro-exp for complex reasoning and analysis tasks
  • All sub-agents updated with correct model specifications

v1.1.0 - Enhanced Trend Analysis

v1.1.0 - Enhanced Trend Analysis

What's New:

Quality Assurance Pipeline

Implemented a comprehensive CI/CD pipeline to improve code quality, reliability, and security.

• Automated Validation: 11 checks covering syntax validation, formatting (Black/Isort), and static analysis (Flake8)
• Security Scanning: Integrated Bandit for vulnerability scanning and Pip-audit for dependency auditing
• Multi-Version Testing: Automated testing across Python 3.10, 3.11, 3.12, and 3.13
• Package Verification: Automated build and metadata validation for distribution

Standardized Folder Structure

Introduced a production-ready input/output structure for a cleaner and more organized workflow.

• Input: Dedicated input/ directory for exam PDFs
• Output: Structured output/ directory with charts/, logs/, and reports/ subfolders
• Utilities: Added paths.py for automatic path resolution
• Git Integration: Updated .gitignore to preserve folder structure while excluding sensitive exam data

Developer Experience

• New Scripts: Added create_sample_exams.py for instant test data generation
• Documentation: Added CI_CD_IMPLEMENTATION.md and FOLDER_IMPLEMENTATION.md for implementation and architecture reference

Dependency Updates

• Added reportlab for sample PDF generation
• Updated requirements.txt with testing, linting, and validation dependencies