Gather-ts NPM | npm.io

gather-ts

A powerful code analysis and packaging tool designed for creating AI-friendly code representations. Gather-ts analyzes your codebase, resolves dependencies, and generates a comprehensive single-file representation optimized for AI analysis.

Features

📊 Comprehensive code analysis with dependency resolution
🔍 Intelligent file filtering and ignore patterns
📝 Token-aware processing with multiple model support
🚀 High performance with batch processing and caching
⚡️ Support for TypeScript and JavaScript projects
🛠️ Configurable output formats
📈 Detailed metrics and analytics

Installation

Install globally:

npm install -g gather-ts

Or as a project dependency:

npm install --save-dev gather-ts

Quick Start

Navigate to your project directory:

cd your-project

Generate a default configuration:

gather-ts --init

Run the analysis:

gather-ts src/index.ts --output analysis.txt

Usage

Basic Command Structure

gather-ts <files...> --output <output> [options]

Arguments

files: Entry files to analyze (comma or space separated)
--output, -o: Output file path
--root, -r: Project root directory (default: current directory)
--depth, -d: Maximum depth for dependency analysis
--debug: Enable debug logging
--batch-size: Batch size for processing files
--metrics: Include performance metrics in output
--config, -c: Path to custom config file
--encoding: File encoding (default: utf8)
--ignore: Additional patterns to ignore
--require: Required files to include
--init: Initialize configuration in current directory

Examples

Analyze a single file:

gather-ts src/app/page.tsx --output output.txt

Analyze multiple files with metrics:

gather-ts src/app/page.tsx,src/components/Button.tsx --output output.txt --metrics

Custom configuration and depth:

gather-ts src/index.ts --output analysis.txt --config custom-config.json --depth 3

Configuration

Gather-ts can be configured using a gather-ts.config.json file in your project root. Generate a default configuration using:

gather-ts --init

Configuration Options

{
  "maxDepth": 5,
  "topFilesCount": 5,
  "showTokenCount": true,
  "tokenizer": {
    "model": "gpt-4",
    "showWarning": true
  },
  "outputFormat": {
    "includeSummaryInFile": true,
    "includeGenerationTime": true,
    "includeUsageGuidelines": true
  },
  "debug": false,
  "cacheTokenCounts": true
}

Configuration Fields

Field	Type	Description	Default
maxDepth	number	Maximum depth for dependency analysis	5
topFilesCount	number	Number of top files to show in summary	5
showTokenCount	boolean	Show token counts in output	true
tokenizer.model	string	Model to use for tokenization	"gpt-4"
tokenizer.showWarning	boolean	Show warnings for token limits	true

File Ignoring

Gather-ts uses both .gitignore and .gather-ts-ignore files for determining which files to process. Create a .gather-ts-ignore file in your project root to specify additional ignore patterns:

# Ignore test files
**/*.test.ts
**/*.spec.ts

# Ignore documentation
docs/
*.md

# Ignore specific directories
temp/
build/

API Usage

You can also use gather-ts programmatically in your Node.js applications:

import { configureContainer } from 'gather-ts';

async function analyzeCode() {
  const container = await configureContainer(process.cwd(), {
    debug: true,
    maxCacheAge: 7 * 24 * 60 * 60 * 1000, // 7 days
    fileExtensions: ['ts', 'tsx', 'js', 'jsx']
  });

  const compiler = container.resolve(ServiceTokens.COMPILER);
  
  const result = await compiler.compile({
    entryFiles: ['src/index.ts'],
    outputFile: 'analysis.txt',
    includeMetrics: true
  });

  console.log(`Analysis complete! Output: ${result.outputPath}`);
}