documentation/system-architecture/technology-stack.md

# Technology Stack

Complete technology stack powering OPSIIE.

## 🐍 Core Language
**Python 3.8+**

## 🤖 AI & ML

### Language Models
- **Ollama** - Local LLM (Llama3 8B)
- **OpenAI** - GPT-3.5-turbo (Nyx agent)
- **Google Gemini** - 1.5 Flash (G1 agent)
- **ElevenLabs** - Conversational AI (Kronos)

### Vision & Voice
- **OpenCV** - Face recognition
- **SpeechRecognition** - Voice input (Google)
- **ElevenLabs API** - Text-to-speech
- **PyAudio** - Audio I/O

### ML Libraries
- **PyTorch** - Deep learning framework
- **Transformers** - Hugging Face models
- **Sentence Transformers** - Embeddings
- **CUDA** - GPU acceleration (optional)

### Generation
- **Hugging Face** - Image generation
- **Replicate** - Video & music generation
- **AudioCraft** - Music models (MusicGen)

## 💾 Data & Storage

### Databases
- **PostgreSQL** - Conversation storage
- **ChromaDB** - Vector database
- **psycopg2** - PostgreSQL adapter

### Data Processing
- **pandas** - Data analysis
- **numpy** - Numerical computing

## 📄 Document Processing

- **PyPDF2** - PDF reading
- **pdfplumber** - Advanced PDF extraction
- **python-docx** - Word documents
- **openpyxl** - Excel files
- **csv** - CSV parsing

## 🌐 Web & APIs

### HTTP
- **requests** - API calls
- **urllib** - URL handling
- **websockets** - Real-time communication

### Web3
- **web3.py** - Ethereum interaction
- **eth-account** - Key management
- **Base, Ethereum, Polygon** - Networks

### Financial
- **yfinance** - Yahoo Finance API
- **Real-time market data**

## 🧬 Bioinformatics

- **Biopython** - Sequence analysis
- **Bio.Blast** - Homology search
- **NCBI Entrez** - Database access
- **UniProt, Pfam** - Protein databases

## 📧 Communication

- **smtplib** - Email sending
- **imaplib** - Email receiving
- **email** - Message formatting
- **Gmail SMTP/IMAP**

## 🎨 Media

### Audio
- **pygame** - Audio playback
- **pydub** - Audio processing
- **soundfile** - File I/O

### Image
- **Pillow (PIL)** - Image processing
- **matplotlib** - Visualization

## 🎨 UI/UX

- **terminal_colors.py** - Custom theming

- **ASCII art** - Splash screens

- **Markdown rendering** - Formatted output

- **Pastel/Vibrant** - Color themes



## 🔧 Utilities



- **python-dotenv** - Environment variables

- **os, sys** - System operations

- **pathlib** - Path handling

- **json, pickle** - Serialization

- **datetime** - Time operations

- **re** - Regular expressions

- **hashlib** - Hashing



## 📦 Package Management



**requirements.txt**:

```

openai

google-generativeai

elevenlabs

psycopg2-binary

chromadb

sentence-transformers

opencv-python

SpeechRecognition

pyaudio

pygame

pydub

torch

transformers

biopython

web3

eth-account

yfinance

requests

pandas

openpyxl

PyPDF2

pdfplumber

python-docx

python-dotenv

replicate

pillow

```



## 🏗️ Architecture Patterns



**MVC-like**:

- Models: Data classes, API interfaces

- Views: Terminal output, formatting

- Controllers: Command parsers, handlers



**Service Layer**:

- Memory service (PostgreSQL + ChromaDB)

- Agent service (Nyx, G1, Kronos)

- Generation service (Images, videos, music)

- Web3 service (Blockchain operations)



**Repository Pattern**:

- Database interactions abstracted

- Consistent interface for data access



## 🚀 Performance



**Optimizations**:

- Async operations (where possible)

- Connection pooling (database)

- Caching (model outputs)

- Batch processing (embeddings)



**Scalability**:

- Stateless agent calls

- Modular architecture

- Configurable limits

- Resource-aware processing



## 🔄 Integration Flow



```

User Input

    ↓

Terminal/Voice

    ↓

Command Parser

    ↓

Service Layer

    ↓

APIs/Models/Database

    ↓

Response Processing

    ↓

Memory Storage

    ↓

Output Formatting

    ↓

Terminal/Voice Output

```



## 📊 Data Pipeline



**Memory Pipeline**:

```

Conversation → PostgreSQL

Conversation → Embeddings → ChromaDB

Query → Vector Search → Relevant Context

```



**Generation Pipeline**:

```

Prompt → Model API → Generation → Storage → Display

```



**Agent Pipeline**:

```

Query → Agent API → Response → Evaluation → Selection → Display

```



## 🔐 Security Stack



- **OpenCV** - Facial authentication

- **dotenv** - Secret management

- **HTTPS** - All API calls

- **Web3** - Checksum addresses

- **psycopg2** - Parameterized queries



## 🌍 External Services



**Required**:

- ElevenLabs API

- Google AI API (Gemini)

- OpenAI API



**Optional (R-Grade)**:

- Hugging Face Inference

- Replicate API

- Yahoo Finance

- NCBI (Entrez, BLAST)

- Blockchain RPC nodes



## 🖥️ System Requirements



**Minimum**:

- Python 3.8+

- 8GB RAM

- PostgreSQL

- Camera + Microphone

- Internet connection



**Recommended**:

- Python 3.10+

- 16GB RAM

- CUDA-capable GPU

- SSD storage

- High-speed internet



---



**Technology powering intelligence.** 🚀