PGL Utils

A comprehensive library for Machine Learning, Deep Learning, and Generative AI utilities, designed for PUC and IBMEC post-graduation students.

Features

• Machine Learning (ML): Preprocessing, models, and utilities for classical ML
• Deep Learning: Architectures, training utilities, and pre-trained models
• Generative AI (GenAI): LLM utilities, RAG implementations, and prompt engineering tools
• Institution-specific extensions: Customized tools for PUC and IBMEC students

Installation

Basic Installation

pip install pgl-utils

Installation with specific features

# Machine Learning only
pip install pgl-utils[ml]

# Deep Learning only
pip install pgl-utils[deep_learning]

# Generative AI only
pip install pgl-utils[genai]

# All features
pip install pgl-utils[all]

# Development
pip install pgl-utils[dev]

Installation from source

git clone https://github.com/renansantosmendes/pgl_utils.git
cd pgl_utils
pip install -e .

Quick Start

Using Core Utilities

from pgl_utils import core

# Your code here

Using Machine Learning Tools

from pgl_utils.ml import preprocessing, models

# Your code here

Using Deep Learning Tools

from pgl_utils.deep_learning import draw_neural_network

# Your code here

Curated Stock Ticker Lists

pgl_utils.deep_learning ships curated lists of B3 (Brazil) and US stock tickers, grouped by sector, for use in time series / financial deep learning examples. They are bundled as JSON resources and loaded via load_brazil_tickers() / load_us_tickers(), so they work both in a dev checkout and after installing the package with pip.

from pgl_utils.deep_learning import load_brazil_tickers, load_us_tickers

brazil_tickers = load_brazil_tickers()
us_tickers = load_us_tickers()

# Each function returns a dict keyed by sector (snake_case), mapping to a list of tickers
brazil_tickers["bancos_servicos_financeiros"]
# ["ITUB4.SA", "ITUB3.SA", "BBDC4.SA", ...]

us_tickers["tecnologia"]
# ["AAPL", "MSFT", "GOOGL", ...]

The underlying JSON files live at pgl_utils/deep_learning/data/brazil_tickers.json and pgl_utils/deep_learning/data/us_tickers.json.

Using Generative AI Tools

from pgl_utils.genai import llm, rag

# Your code here

Using the PGL Search Gateway

from pgl_utils.genai.search_gateway import PGLSearchGateway

gateway = PGLSearchGateway()
results = gateway.search(
    "inteligência artificial generativa",
    search_depth="advanced",
    include_raw_content=True,
)

Institution-Specific Tools

For PUC Students

from pgl_utils.puc import config

puc_info = config.PUCConfig.get_info()

For IBMEC Students

from pgl_utils.ibmec import config

ibmec_info = config.IBMECConfig.get_info()

Project Structure

pgl_utils/
├── pgl_utils/          # Main package
│   ├── __init__.py
│   ├── core/                       # Shared utilities
│   │   ├── __init__.py
│   │   └── utils.py
│   ├── ml/                         # Machine Learning module
│   │   ├── __init__.py
│   │   ├── preprocessing.py
│   │   └── models.py
│   ├── deep_learning/              # Deep Learning module
│   │   ├── __init__.py
│   │   ├── architectures.py
│   │   ├── training.py
│   │   ├── tickers.py              # Brazil/US ticker loaders
│   │   └── data/                   # brazil_tickers.json, us_tickers.json
│   ├── genai/                      # Generative AI module
│   │   ├── __init__.py
│   │   ├── llm.py
│   │   ├── rag.py
│   │   └── search_gateway.py
│   ├── puc/                        # PUC-specific extensions
│   │   ├── __init__.py
│   │   └── config.py
│   └── ibmec/                      # IBMEC-specific extensions
│       ├── __init__.py
│       └── config.py
├── tests/                          # Unit tests
├── examples/                       # Example notebooks and scripts
├── docs/                           # Documentation
├── setup.py                        # Package configuration
├── requirements.txt                # Dependencies
├── README.md                       # This file
└── .gitignore                      # Git ignore rules

Requirements

• Python >= 3.8
• numpy >= 1.21.0
• pandas >= 1.3.0
• scikit-learn >= 1.0.0

Dependencies by Module

Machine Learning (ML)

• scikit-learn
• xgboost
• lightgbm

Deep Learning

• torch
• tensorflow
• keras

Generative AI (GenAI)

• openai
• langchain
• huggingface-hub

Examples

See the examples/ directory for jupyter notebooks and scripts demonstrating library usage.

Testing

Run tests with pytest:

pytest tests/

With coverage:

pytest tests/ --cov=pgl_utils

Contributing

Contributions are welcome! Please feel free to submit pull requests or open issues.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Support

For issues, questions, or suggestions, please open an issue on GitHub.

Changelog

Version 0.1.0

• Initial release
• Core functionality for ML, Deep Learning, and GenAI
• Institution-specific extensions for PUC and IBMEC