Skip to content

CLAUDE.md

Latest commit

 

History

History
242 lines (191 loc) · 7.46 KB

CLAUDE.md

File metadata and controls

242 lines (191 loc) · 7.46 KB
description Use Bun instead of Node.js, npm, pnpm, or vite.
globs *.ts, *.tsx, *.html, *.css, *.js, *.jsx, package.json
alwaysApply false

Default to using Bun instead of Node.js.

  • Use bun <file> instead of node <file> or ts-node <file>
  • Use bun test instead of jest or vitest
  • Use bun build <file.html|file.ts|file.css> instead of webpack or esbuild
  • Use bun install instead of npm install or yarn install or pnpm install
  • Use bun run <script> instead of npm run <script> or yarn run <script> or pnpm run <script>
  • Bun automatically loads .env, so don't use dotenv.

APIs

  • Bun.serve() supports WebSockets, HTTPS, and routes. Don't use express.
  • bun:sqlite for SQLite. Don't use better-sqlite3.
  • Bun.redis for Redis. Don't use ioredis.
  • Bun.sql for Postgres. Don't use pg or postgres.js.
  • WebSocket is built-in. Don't use ws.
  • Prefer Bun.file over node:fs's readFile/writeFile
  • Bun.$ls instead of execa.

Testing

Use bun test to run tests.

import { test, expect } from "bun:test";

test("hello world", () => {
  expect(1).toBe(1);
});

Frontend

Use HTML imports with Bun.serve(). Don't use vite. HTML imports fully support React, CSS, Tailwind.

Server:

import index from "./index.html"

Bun.serve({
  routes: {
    "/": index,
    "/api/users/:id": {
      GET: (req) => {
        return new Response(JSON.stringify({ id: req.params.id }));
      },
    },
  },
  // optional websocket support
  websocket: {
    open: (ws) => {
      ws.send("Hello, world!");
    },
    message: (ws, message) => {
      ws.send(message);
    },
    close: (ws) => {
      // handle close
    }
  },
  development: {
    hmr: true,
    console: true,
  }
})

HTML files can import .tsx, .jsx or .js files directly and Bun's bundler will transpile & bundle automatically. <link> tags can point to stylesheets and Bun's CSS bundler will bundle.

<html>
  <body>
    <h1>Hello, world!</h1>
    <script type="module" src="./frontend.tsx"></script>
  </body>
</html>

With the following frontend.tsx:

import React from "react";

// import .css files directly and it works
import './index.css';

import { createRoot } from "react-dom/client";

const root = createRoot(document.body);

export default function Frontend() {
  return <h1>Hello, world!</h1>;
}

root.render(<Frontend />);

Then, run index.ts

bun --hot ./index.ts

For more information, read the Bun API docs in node_modules/bun-types/docs/**.md.

AI Interaction Guidelines

User Environment & Preferences

  • OS: macOS (Darwin 23.6.0) with limited sudo access for system diagnostics
  • Communication style: Expert level, concise responses preferred
  • Boundaries: Will not kill critical system processes, prefers non-destructive debugging methods

Critical Instructions

  • NEVER suggest killing system processes if user explicitly states they don't want to kill them
  • Start with non-privileged tools (ps, top, Activity Monitor) before attempting sudo-required diagnostics
  • If system diagnostic commands fail with permissions, immediately pivot to user-accessible alternatives
  • For macOS system issues, always provide context about normal vs abnormal behavior before suggesting fixes

System Debugging Approach

  1. Use basic process info commands first: ps, top
  2. If detailed inspection is needed and fails due to permissions, suggest Activity Monitor GUI
  3. Always explain both immediate fixes and long-term preventive measures
  4. Focus on monitoring and diagnostic approaches rather than destructive fixes

TodoWrite Usage

  • Create structured plans with TodoWrite for complex system debugging tasks
  • Track multiple diagnostic approaches systematically

Claude Story Analyzer Development Notes

Project Architecture

src/
├── parsers/           # Claude Code JSONL file parsing
├── analyzers/         # AI insight extraction (pattern-matching + AI-powered)
├── exporters/         # Report generation in multiple formats
├── cli/              # Command-line interface
└── types/            # TypeScript type definitions

Key Implementation Insights

1. Claude Code Storage Format

  • Conversations stored in ~/.claude/projects/[project-hash]/[session-id].jsonl
  • Project names encoded as -Users-username-path-to-project
  • Each JSONL line is a message with type, timestamp, message, uuid, etc.
  • Need to handle user, assistant, summary, and tool_result message types

2. Project Detection Strategy

  • Exact path match: Most reliable, matches encoded paths exactly
  • Project name match: Fallback, matches last directory segment
  • Partial path match: Last resort, matches common path segments
  • Always pick most recently active project when multiple matches

3. AI-Powered vs Pattern-Matching Analysis

  • Pattern-matching: Fast, regex-based, good for basic insights
  • AI-powered: Requires API key, deep contextual analysis, slower but much better
  • Chunk conversations into ~45k token segments for API calls
  • Use single comprehensive prompt for both insights and user patterns

4. Development & Testing Tips

Quick Testing

# Test project detection
bun test-detector.ts

# Test simple analysis  
bun test-simple-analyze.ts

# Test specific depth
bun run dev analyze --depth 20 --debug

Build & Type Issues

# Separate tsconfig for declarations to avoid conflicts
tsconfig.json          # Main config with noEmit: true
tsconfig.build.json     # Build config for declarations

# Fix TypeScript strict null issues with !
const first = messages[0]!;  # Assert non-null

Common Problems

  1. Indentation errors: Use consistent 2-space indentation
  2. API type issues: Cast Anthropic response content as any if needed
  3. Progress bars: Use cli-progress for long-running operations
  4. Rate limiting: Add delays between API calls (1000ms recommended)

CLI Structure

  • Separate command functions in cli/commands/ for cleaner code
  • Use Commander.js options properly with string defaults
  • Handle missing API keys gracefully with helpful error messages

Testing Strategy

  1. Unit tests: Test parsers and analyzers separately
  2. Integration tests: Test with real JSONL files
  3. Manual testing: Use current project as test case (meta!)
  4. Debug mode: Always provide debug options for troubleshooting

Performance Optimization

  • Limit message analysis with --depth parameter
  • Use --recent flag to analyze only recent sessions
  • Chunk large conversations for memory efficiency
  • Add progress indicators for long operations

Environment Setup

# Required for AI-powered analysis
export ANTHROPIC_API_KEY=your_key_here

# Optional: faster package manager
export BUN_INSTALL="$HOME/.bun"

Useful Commands

# Development
bun run dev analyze --debug              # Debug project detection
bun run dev analyze --depth 50          # Quick analysis
bun run dev analyze --ai-powered         # Full AI analysis

# Build & Deploy
bun run build                           # Build for production
bun test                               # Run tests
node dist/cli.js analyze               # Test built version

Future Improvements

  • Add conversation threading analysis
  • Implement conversation search/filtering
  • Export to more formats (Excel, PDF)
  • Add real-time analysis of active sessions
  • Implement conversation comparison between projects