Source2Lab
PDFs into traceable, executable notebooks.
Currently unavailable to view: this project is being extended with new functionality, so access is temporarily paused.
Tech Stack
About
Source2Lab takes research papers, documentation, and technical PDFs and extracts their content into structured Jupyter notebooks. Each cell is linked to its source passage, creating a traceable connection between reference material and executable code. The system is designed to work without any paid API, external LLM, or cloud service — it runs entirely locally.
The Problem
Converting research PDFs into actionable, executable learning material is tedious and disconnected — extracted text cannot be run, and the link back to source passages is lost.
The Approach
Source2Lab uses pypdf for extraction and nbformat/nbclient to construct and execute notebooks, keeping traceability links from each cell back to its source paragraph.
Key Features
- PDF extraction and chunking with pypdf
- Structured Jupyter notebook generation with nbformat
- Executable notebook validation with nbclient
- Source traceability: each cell links to its source
- Docker-based isolated execution environment
- No paid API or LLM key required
Highlights
- Fully local pipeline — no external API or LLM key required.
- Source traceability preserved from PDF extraction to notebook cell.