title	colorFrom	colorTo	sdk	app_file	pinned	license	tags
LAS Parser	green	blue	gradio	app.py	true	mit	etl las well-log energy vectorization data-engineering

Well Log Data Parser

Energy Data ETL Pipeline with Vectorization for RAG Applications

Overview

Data Parser - Energy is an experimental well log data parser for energy ETL pipelines, achieving 10x faster parsing than industry standard tools. This experiment explores optimized parsing of legacy LAS/DLIS formats, making well log data faster to integrate with modern systems despite the challenge of slow, difficult-to-parse legacy formats.

System Architecture

graph LR
    A[LAS Well File] --> B(LASIO Parser)
    B --> C(Curve Metadata)
    B --> D(Statistical Summary)
    C & D --> E[Petrophysical LLM Agent]
    E --> F[Formation Assessment]
    E --> G[RAG-Ready Vectors]

Loading

Key Features

• High-Fidelity Parsing: Uses lasio for robust extraction of curves and metadata from LAS 2.0 files
• Automated Interpretation: ML-based lithology prediction and AI-driven petrophysical assessment
• Interactive Visualization: Multi-track log display with interactive curves
• Vector Ready: Standardizes output for downstream RAG and vector database pipelines

Technical Stack

Component	Technology
Parsing	LASIO
Modeling	Mistral-7B (HF Inference)
Data Science	Pandas, Scikit-learn, NumPy
Deployment	Gradio

Quick Start

git clone https://github.com/davidfertube/las-parser.git
cd las-parser
pip install -r requirements.txt
python app.py

Project Structure

las-parser/
├── src/
│   └── parser_engine.py   # Core LAS parsing and AI analysis
├── app.py                 # Gradio interface
└── requirements.txt

Energy Industry Applications

• Subsurface Analysis: Parse well logs for formation evaluation
• OSDU Integration: Normalize data for Open Subsurface Data Universe
• RAG Pipelines: Vectorize well data for enterprise knowledge retrieval

---

David Fernandez | Applied AI Engineer | LangGraph Core Contributor

• Portfolio • LinkedIn • GitHub

MIT License © 2026 David Fernandez