Code-Craft: Hierarchical Graph-Based Code Summarization for Enhanced Context Retrieval

Ashok Javvaji; David Sounthiraraj; Jared Hancock; Prabhat Singh; Shaila Shankar; Yassin Kortam

RECORD

Code-Craft: Hierarchical Graph-Based Code Summarization for Enhanced Context Retrieval

David Sounthiraraj, Jared Hancock, Yassin Kortam, Ashok Javvaji, Prabhat Singh, Shaila Shankar

Software Engineering · Submitted 11 Apr 2025

abstract compiled · 2026-09-23 · no language-model review

The authors' main claim The system's hierarchical approach consistently outperforms traditional code-only retrieval across all metrics, with particularly substantial gains in larger, more complex codebases where understanding function relationships is crucial.

how it is stated · asserted 1.00 read the record →

Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID2504.08975 (version 1)

Pith IDOI4QF7MK, Pith's permanent record of this paper and its review

CitationAshok Javvaji, David Sounthiraraj, Jared Hancock, Prabhat Singh, Shaila Shankar, Yassin Kortam. "Code-Craft: Hierarchical Graph-Based Code Summarization for Enhanced Context Retrieval." arXiv:2504.08975 (2025). https://pith.science/paper/OI4QF7MK

Claim paper

The compiled record

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Open the compiled record

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 5,033 input tokens.

What the abstract itself states, as probabilities: claims first 0.11 · states limitation 0.09 · human participants 0.04 · controlled comparison 0.86 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.94 · mentions code or data availability 0.03.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

Understanding and navigating large-scale codebases remains a significant challenge in software engineering. Existing methods often treat code as flat text or focus primarily on local structural relationships, limiting their ability to provide holistic, context-aware information retrieval. We present Hierarchical Code Graph Summarization (HCGS), a novel approach that constructs a multi-layered representation of a codebase by generating structured summaries in a bottom-up fashion from a code graph. HCGS leverages the Language Server Protocol for language-agnostic code analysis and employs a parallel level-based algorithm for efficient summary generation. Through extensive evaluation on five diverse codebases totaling 7,531 functions, HCGS demonstrates significant improvements in code retrieval accuracy, achieving up to 82 percentage relative improvement in top-1 retrieval precision for large codebases like libsignal (27.15 percentage points), and perfect Pass@3 scores for smaller repositories. The system's hierarchical approach consistently outperforms traditional code-only retrieval across all metrics, with particularly substantial gains in larger, more complex codebases where understanding function relationships is crucial.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Understanding and navigating large-scale codebases remains a significant challenge in software engineering.	background 1.00	0.03	0.02	asserted 1.00
s1	Existing methods often treat code as flat text or focus primarily on local structural relationships, limiting their ability to provide holistic, context-aware information retrieval.	gap 0.88	0.03	0.03	asserted 0.97
s2	We present Hierarchical Code Graph Summarization (HCGS), a novel approach that constructs a multi-layered representation of a codebase by generating structured summaries in a bottom-up fashion from a code graph.	contribution 0.91	0.92	0.10	asserted 1.00
s3	HCGS leverages the Language Server Protocol for language-agnostic code analysis and employs a parallel level-based algorithm for efficient summary generation.	method 1.00	0.75	0.03	asserted 1.00
s4	Through extensive evaluation on five diverse codebases totaling 7,531 functions, HCGS demonstrates significant improvements in code retrieval accuracy, achieving up to 82 percentage relative improvement in top-1 retrieval precision for large codebases like libsignal (27.15 percentage points), and perfect Pass@3 scores for smaller repositories.	result 1.00	0.97	0.54	asserted 1.00
s5	The system's hierarchical approach consistently outperforms traditional code-only retrieval across all metrics, with particularly substantial gains in larger, more complex codebases where understanding function relationships is crucial.	result 0.63	0.88	0.63	asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.99.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

The paper

Original abstract

Understanding and navigating large-scale codebases remains a significant challenge in software engineering. Existing methods often treat code as flat text or focus primarily on local structural relationships, limiting their ability to provide holistic, context-aware information retrieval. We present Hierarchical Code Graph Summarization (HCGS), a novel approach that constructs a multi-layered representation of a codebase by generating structured summaries in a bottom-up fashion from a code graph. HCGS leverages the Language Server Protocol for language-agnostic code analysis and employs a parallel level-based algorithm for efficient summary generation. Through extensive evaluation on five diverse codebases totaling 7,531 functions, HCGS demonstrates significant improvements in code retrieval accuracy, achieving up to 82 percentage relative improvement in top-1 retrieval precision for large codebases like libsignal (27.15 percentage points), and perfect Pass@3 scores for smaller repositories. The system's hierarchical approach consistently outperforms traditional code-only retrieval across all metrics, with particularly substantial gains in larger, more complex codebases where understanding function relationships is crucial.

Human review

How Pith checks papers → · Full audit of this paper →

References

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Code Comprehensor: Hierarchical Documentation and Cross-Modal Retrieval for Large Repositories
cs.SE 2026-10 conditional novelty 5.0 of 10

Using LLM-generated hierarchical documentation and synthetic questions as training signal improves text-to-code retrieval on large C/C++ repositories, per a two-repository evaluation.
Agent4cs: A Multi-agent System for Code Summarization in Large Hierarchical Codebases
cs.AI 2026-07 unverdicted novelty 5.0 of 10

A multi-agent bottom-up framework (summarizer + keyword extractor + QA) improves hierarchical code-summary consistency by ~8% and normalized keyword coverage by up to 38% over structured prompting baselines.

How the literature uses it

2 arXiv papers cite this in 3 sentences

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

Papers citing it · 2 sentences without a confident label

Cites it as background 1 paper

For instance, HCGS generated summaries of code elements and their relationships to achieve context-aware code retrieval [ref] , while another method leveraged LLMs to build hierarchical project summaries for bug localization, performing top-down inference to overcome domain mismatch and context limits [ref] .

Agent4cs: A Multi-agent System for Code Summarization in Large Hierarchical Codebases · Hierarchical Code Summarization · 0.95

What it cites · 3 sentences without a confident label

Cites as support 1 paper

Ma et al. [ref] emphasizes that comprehensive repository-level understanding is critical for effective automated software engineering, identifying key challenges including the impracticality of including thousands of files in LLM context windows and the difficulty of identifying code relevant to specific issues within extensive contexts.

Alibaba LingmaAgent: Improving Automated Issue Resolution via Comprehensive Repository Exploration · Introduction · 0.97

Cites as background 2 papers

Ma et al. [ref] introduced RepoUnderstander, which constructs a repository knowledge graph in a top-down manner and employs Monte Carlo Tree Search to efficiently explore it, prioritizing critical information discovery.

Alibaba LingmaAgent: Improving Automated Issue Resolution via Comprehensive Repository Exploration · Related Work · 0.84
Deep learning has emerged as the dominant approach in recent years [ref] , particularly for addressing the semantic gap between queries and code.

Geometrical Tools for Teaching Azeotropy Using Simplified Thermodynamic Models · Related Work · 0.95

Discussion

Know something about this paper? Leave a comment, or highlight any passage above to comment on it.