CRABS: A syntactic-semantic pincer strategy for bounding LLM interpretation of Python notebooks

TL;DR

CRABS combines syntactic analysis with LLMs to accurately interpret data flow and dependencies in Python notebooks, overcoming hallucination and context challenges for better understanding and reuse.

Contribution

It introduces a novel pincer strategy that integrates shallow syntactic parsing with LLMs to improve notebook comprehension without execution.

Findings

Achieves 98% accuracy in identifying cell-to-cell information flows.

Reaches 99% accuracy in determining transitive cell dependencies.

Successfully resolves 98% of ambiguities in a dataset of Kaggle notebooks.

Abstract

Recognizing the information flows and operations comprising data science and machine learning Python notebooks is critical for evaluating, reusing, and adapting notebooks for new tasks. Investigating a notebook via re-execution often is impractical due to the challenges of resolving data and software dependencies. While Large Language Models (LLMs) pre-trained on large codebases have demonstrated effectiveness in understanding code without running it, we observe that they fail to understand some realistic notebooks due to hallucinations and long-context challenges. To address these issues, we propose a notebook understanding task yielding an information flow graph and corresponding cell execution dependency graph for a notebook, and demonstrate the effectiveness of a pincer strategy that uses limited syntactic analysis to assist full comprehension of the notebook using an LLM. Our Capture and Resolve Assisted Bounding Strategy (CRABS) employs shallow syntactic parsing and analysis of the abstract syntax tree (AST) to capture the correct interpretation of a notebook between lower and upper estimates of the inter-cell I/O set$\unicode{x2014}$the flows of information into or out of cells via variables$\unicode{x2014}$then uses an LLM to resolve remaining ambiguities via cell-by-cell zero-shot learning, thereby identifying the true data inputs and outputs of each cell. We evaluate and demonstrate the effectiveness of our approach using an annotated dataset of 50 representative, highly up-voted Kaggle notebooks that together represent 3454 actual cell inputs and outputs. The LLM correctly resolves 1397 of 1425 (98%) ambiguities left by analyzing the syntactic structure of these notebooks. Across 50 notebooks, CRABS achieves average F1 scores of 98% identifying cell-to-cell information flows and 99% identifying transitive cell execution dependencies.