How botta0oss/bigbrother works
Main Components
a. Data Ingestion
• Local CSV datasets (data/*.csv)
• Telegram API input (API.py – to be created; uses api_id/api_hash/phone)
b. Analysis Engine
• Entry point: main.py
• Core logic: Analysis.py (statistical analysis, sentiment classification via HuggingFace transformers)
c. Data Storage & Artifacts
• Raw & processed CSV files (data/)
• Generated charts (grafici/)
d. Visualization
• Graph generation modules (matplotlib/seaborn in Analysis.py)
e. External Dependencies & Models
• Python 3.9+ runtime
• Transformers library (HuggingFace)
• pandas, numpy, matplotlib, etc.
Component Relationships & Data Flow
• CSV files read by main.py/Analysis.py
• Telegram messages fetched by API.py
• Preprocessing, tokenization, statistical computations
• Calls HuggingFace transformer models for embedding/sentiment analysis
• Writes results back to CSV in data/
• Generates charts in grafici/
Architectural Patterns & Principles
• Modular separation: ingestion logic (API.py), analysis logic (Analysis.py), orchestration (main.py)
• Single-process monolith for simplicity, but modules can be refactored into microservices if needed
• Extensibility points: pluggable data sources (add new API modules), new analysis submodules
Technologies & Frameworks
• Python 3.9+
• HuggingFace Transformers (for NLP models)
• pandas/numpy (data manipulation)
• matplotlib/seaborn (plotting)
• Telethon or python-telegram-api (in API.py for Telegram ingestion)
Final Checklist
• All modules and folders clearly labeled
• An outer boundary that encapsulates the system