GitHub Repo Indexer: functions, calls and metrics
By (author)fizzah abdullah
Tech stack:
JS
TS
CSS
Point it at a public GitHub repo and get every Python, JavaScript and TypeScript function with exact line ranges, a call graph and simple metrics. Use...
read more
read more
AI Contributions:0
Published At:October 19, 2026
Files Included:7 file(s)
Updated At:October 22, 2026
Read Me
# GitHub Repo Indexer: functions, calls and metrics
## Description
Point it at a public GitHub repo and get every Python, JavaScript and TypeScript function with exact line ranges, a call graph and simple metrics. Uses tree-sitter. No cloning, no code execution.
## Skills
TypeScript, Node.js
## Custom tags
tree-sitter, Octokit, GitHub API, static analysis, call graph, code search
## What it does
- Inspects public GitHub repositories without cloning them locally.
- Traverses the file tree, filtering out binaries, vendor bundles, and build directories (`node_modules`, `dist`, `.next`).
- Uses `web-tree-sitter` WASM grammars to parse Python, JavaScript, TypeScript, and TSX files statically.
- Extracts named functions, class methods, variable-assigned arrow functions, and lambdas with exact start and end line ranges.
- Traces call expressions within function bodies and resolves callers and callees (same file first, then repository-wide unique names).
- Computes code complexity metrics: length in lines, decision branches, `data_access` detection, and an importance score.
- Categorizes each function into `entry`, `data`, `core`, or `helper`.
- Zero dependencies on application code or specific database schemas — drop it directly into any Node.js project.
## Public API
```ts
export type IndexRepoOptions = {
url: string;
token?: string;
folder?: string;
limits?: {
maxFiles?: number; // default: 400
maxFileSize?: number; // default: 200 KB
parallelRequests?: number; // default: 8
};
signal?: AbortSignal;
};
export type IndexRepoOutput = {
owner: string;
name: string;
commitSha: string;
files: string[];
functions: IndexedFunction[];
calls: ResolvedCall[];
skipped: Array<{ path: string; reason: string }>;
truncated: boolean;
};
export function indexRepo(options: IndexRepoOptions): Promise<Result<IndexRepoOutput>>;
export function indexSourceFiles(options: IndexFilesOptions): Promise<Result<IndexRepoOutput>>;
```
## Install
```bash
npm install web-tree-sitter@0.20.8 tree-sitter-wasms@0.1.11
```
## Setup
For Next.js App Router, ensure tree-sitter's WASM files are preserved in the bundle by adding to `next.config.ts`:
```ts
export default {
serverExternalPackages: ["web-tree-sitter"],
};
```
## Usage Example
```ts
import { indexRepo } from "./lib/repo-indexer";
const result = await indexRepo({
url: "https://github.com/expressjs/express",
token: process.env.GITHUB_TOKEN, // optional
});
if (result.ok) {
const { functions, calls, files } = result.data;
console.log(`Indexed ${functions.length} functions across ${files.length} files.`);
// Find top functions by importance
const topFunctions = [...functions].sort((a, b) => b.importance - a.importance).slice(0, 5);
for (const fn of topFunctions) {
console.log(`${fn.qualifiedName} [${fn.category}] - score: ${fn.importance}`);
}
} else {
console.error("Index failed:", result.reason);
}
```
## Options and Limits
- Default file size cap is 200 KB per file.
- Default file count cap is 400 supported files.
- Default concurrent GitHub API requests is 8.
- Flags repositories marked as `truncated` by GitHub when the tree exceeds API size limits.
- Supports both remote GitHub repositories (`indexRepo`) and in-memory source files (`indexSourceFiles`).
## Dependencies and Licences
- `web-tree-sitter` — MIT: https://github.com/tree-sitter/tree-sitter
- `tree-sitter-wasms` — MIT: https://github.com/SamSovereign/tree-sitter-wasms
## Contributors
Roshaan Tahir, M Hammad Khan, Fizzah Abdullah, Hammad Khan
---
## GIG READY block
GIG READY: GitHub Repo Indexer
Publish at Devmarketplace → Add your Gigs
1. Thumbnail
Upload docs/gigs/thumbnails/github-repo-indexer.png
2. Basic Info
Title (49/50): GitHub Repo Indexer: functions, calls and metrics
Description (195/200): Point it at a public GitHub repo and get every Python, JavaScript and TypeScript function with exact line ranges, a call graph and simple metrics. Uses tree-sitter. No cloning, no code execution.
3. Tech Stack
Skills (pick the closest ones in the list): TypeScript, Node.js
Custom tags (6/10, press Enter after each): tree-sitter, Octokit, GitHub API, static analysis, call graph, code search
4. Code Snippet
Language: TypeScript
From: lib/repo-indexer/index.ts, lines 177 to 207
```ts
export async function indexRepo(
options: IndexRepoOptions,
): Promise<IndexRepoResult> {
const parsed = parseRepoUrl(options.url);
if (!parsed.ok) return parsed;
const { owner, repo } = parsed.data;
const token = options.token ?? process.env.GITHUB_TOKEN ?? "";
// 1. Fetch file tree without cloning
const treeResult = await fetchDefaultBranchTree({
owner,
repo,
token,
signal: options.signal,
});
if (!treeResult.ok) return { ok: false, reason: treeResult.reason };
// 2. Filter supported source files (Python, JS, TS, TSX)
const sourcePaths = filterTree(treeResult.data.entries).filter(isSupportedSourcePath);
const fetchResult = await fetchFiles({
owner,
repo,
ref: treeResult.data.defaultBranch,
paths: sourcePaths.slice(0, options.limits?.maxFiles ?? 400),
token,
});
// 3. Extract functions, resolve call graph, and compute metrics
const allFunctions = await extractAllFunctions(fetchResult.files);
const resolvedCalls = resolveCalls(allFunctions.rawCalls, allFunctions.list);
const indexedFunctions = await computeAllMetrics(allFunctions.list, allFunctions.rawCalls, resolvedCalls);
return { ok: true, data: { owner, name: repo, functions: indexedFunctions, calls: resolvedCalls, files: fetchResult.files.map((f) => f.path) } };
}
```
5. Read ME
Open docs/gigs/github-repo-indexer.md on GitHub, copy the rendered page and paste it into the editor.
6. After you click Create
Add the gig link to the tracker in docs/GIGS.md and to the Devpost draft.
