combicode 1.0.2 → 1.1.0

package/README.md

@@ -4,15 +4,16 @@
 [![PyPI Version](https://img.shields.io/pypi/v/combicode.svg)](https://pypi.org/project/combicode/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
-<img align="center" src="https://github.com/aaurelions/combicode/raw/main/screenshot.png" width="600"/>
+<img align="center" src="https://github.com/aaurelions/combicode/raw/main/screenshot.png" width="800"/>
 
 **Combicode** is a zero-dependency CLI tool that intelligently combines your project's source code into a single, LLM-friendly text file.
 
-Paste the contents of `combicode.txt` into ChatGPT, Claude, or any other LLM to give it the full context of your repository instantly.
+The generated file starts with a system prompt and a file tree overview, priming the LLM to understand your project's complete context instantly. Paste the contents of `combicode.txt` into ChatGPT, Claude, or any other LLM to get started.
 
 ## Why use Combicode?
 
-- **Maximum Context:** Give your LLM a complete picture of your project structure and code.
+- **Maximum Context:** Gives your LLM a complete picture of your project structure and code.
+- **Intelligent Priming:** Starts the output with a system prompt and a file tree, directing the LLM to analyze the entire codebase before responding.
 - **Intelligent Ignoring:** Automatically skips `node_modules`, `.venv`, `dist`, `.git`, binary files, and other common junk.
 - **`.gitignore` Aware:** Respects your project's existing `.gitignore` rules out of the box.
 - **Zero-Install Usage:** Run it directly with `npx` or `pipx` without polluting your environment.
@@ -34,7 +35,7 @@ npx combicode
 pipx run combicode
 ```
 
-This will create a `combicode.txt` file in your project directory.
+This will create a `combicode.txt` file in your project directory, complete with the context-setting header.
 
 ## Usage and Options
 
@@ -88,6 +89,7 @@ npx combicode -e "**/*_test.py,docs/**"
 | `--include-ext`  | `-i`  | Comma-separated list of extensions to exclusively include.   | (include all)   |
 | `--exclude`      | `-e`  | Comma-separated list of additional glob patterns to exclude. | (none)          |
 | `--no-gitignore` |       | Do not use patterns from the project's `.gitignore` file.    | `false`         |
+| `--no-header`    |       | Omit the introductory prompt and file tree from the output.  | `false`         |
 | `--version`      | `-v`  | Show the version number.                                     |                 |
 | `--help`         | `-h`  | Show the help message.                                       |                 |
 

package/config/ignore.json

@@ -0,0 +1,53 @@
+[
+  "**/node_modules/**",
+  "**/.git/**",
+  "**/.vscode/**",
+  "**/.idea/**",
+  "**/*.log",
+  "**/.env",
+  "**/*.lock",
+  "**/.venv/**",
+  "**/venv/**",
+  "**/env/**",
+  "**/__pycache__/**",
+  "**/*.pyc",
+  "**/*.egg-info/**",
+  "**/build/**",
+  "**/dist/**",
+  "**/.pytest_cache/**",
+  "**/.npm/**",
+  "**/pnpm-lock.yaml",
+  "**/package-lock.json",
+  "**/.next/**",
+  "**/.DS_Store",
+  "**/Thumbs.db",
+  "**/*.png",
+  "**/*.jpg",
+  "**/*.jpeg",
+  "**/*.gif",
+  "**/*.ico",
+  "**/*.svg",
+  "**/*.webp",
+  "**/*.mp3",
+  "**/*.wav",
+  "**/*.flac",
+  "**/*.mp4",
+  "**/*.mov",
+  "**/*.avi",
+  "**/*.zip",
+  "**/*.tar.gz",
+  "**/*.rar",
+  "**/*.pdf",
+  "**/*.doc",
+  "**/*.docx",
+  "**/*.xls",
+  "**/*.xlsx",
+  "**/*.dll",
+  "**/*.exe",
+  "**/*.so",
+  "**/*.a",
+  "**/*.lib",
+  "**/*.o",
+  "**/*.bin",
+  "**/*.iso"
+]

package/index.js

@@ -8,60 +8,29 @@ const glob = require("fast-glob");
 
 const { version } = require("./package.json");
 
-const DEFAULT_IGNORE_PATTERNS = [
-  "**/node_modules/**",
-  "**/.git/**",
-  "**/.vscode/**",
-  "**/.idea/**",
-  "**/*.log",
-  "**/.env",
-  "**/*.lock",
-  "**/.venv/**",
-  "**/venv/**",
-  "**/env/**",
-  "**/__pycache__/**",
-  "**/*.pyc",
-  "**/*.egg-info/**",
-  "**/build/**",
-  "**/dist/**",
-  "**/.pytest_cache/**",
-  "**/.npm/**",
-  "**/pnpm-lock.yaml",
-  "**/package-lock.json",
-  "**/.next/**",
-  "**/.DS_Store",
-  "**/Thumbs.db",
-  // Common binary file extensions
-  "**/*.png",
-  "**/*.jpg",
-  "**/*.jpeg",
-  "**/*.gif",
-  "**/*.ico",
-  "**/*.svg",
-  "**/*.webp",
-  "**/*.mp3",
-  "**/*.wav",
-  "**/*.flac",
-  "**/*.mp4",
-  "**/*.mov",
-  "**/*.avi",
-  "**/*.zip",
-  "**/*.tar.gz",
-  "**/*.rar",
-  "**/*.pdf",
-  "**/*.doc",
-  "**/*.docx",
-  "**/*.xls",
-  "**/*.xlsx",
-  "**/*.dll",
-  "**/*.exe",
-  "**/*.so",
-  "**/*.a",
-  "**/*.lib",
-  "**/*.o",
-  "**/*.bin",
-  "**/*.iso",
-];
+const SYSTEM_PROMPT = `You are an expert software architect. The user is providing you with the complete source code for a project, contained in a single file. Your task is to meticulously analyze the provided codebase to gain a comprehensive understanding of its structure, functionality, dependencies, and overall architecture.
+
+A file tree is provided below to give you a high-level overview. The subsequent sections contain the full content of each file, clearly marked with "// FILE: <path>".
+
+Your instructions are:
+1.  **Analyze Thoroughly:** Read through every file to understand its purpose and how it interacts with other files.
+2.  **Identify Key Components:** Pay close attention to configuration files (like package.json, pyproject.toml), entry points (like index.js, main.py), and core logic.
+`;
+
+function loadDefaultIgnorePatterns() {
+  const configPath = path.resolve(__dirname, "config", "ignore.json");
+  try {
+    const rawConfig = fs.readFileSync(configPath, "utf8");
+    return JSON.parse(rawConfig);
+  } catch (err) {
+    console.error(
+      `❌ Critical: Could not read or parse bundled ignore config at ${configPath}`
+    );
+    process.exit(1);
+  }
+}
+
+const DEFAULT_IGNORE_PATTERNS = loadDefaultIgnorePatterns();
 
 function isLikelyBinary(file) {
   const buffer = Buffer.alloc(512);
@@ -69,18 +38,52 @@ function isLikelyBinary(file) {
   try {
     fd = fs.openSync(file, "r");
     const bytesRead = fs.readSync(fd, buffer, 0, 512, 0);
-    // Check for null bytes, a strong indicator of a binary file
     return buffer.slice(0, bytesRead).includes(0);
   } catch (e) {
-    // If we can't read it, treat it as something to skip
     return true;
   } finally {
     if (fd) fs.closeSync(fd);
   }
 }
 
+function generateFileTree(files, root) {
+  let tree = `${path.basename(root)}/\n`;
+  const structure = {};
+
+  files.forEach((file) => {
+    const parts = file.split(path.sep);
+    let currentLevel = structure;
+    parts.forEach((part) => {
+      if (!currentLevel[part]) {
+        currentLevel[part] = {};
+      }
+      currentLevel = currentLevel[part];
+    });
+  });
+
+  const buildTree = (level, prefix) => {
+    const entries = Object.keys(level);
+    entries.forEach((entry, index) => {
+      const isLast = index === entries.length - 1;
+      tree += `${prefix}${isLast ? "└── " : "├── "}${entry}\n`;
+      if (Object.keys(level[entry]).length > 0) {
+        buildTree(level[entry], `${prefix}${isLast ? "    " : "│   "}`);
+      }
+    });
+  };
+
+  buildTree(structure, "");
+  return tree;
+}
+
 async function main() {
-  const argv = yargs(hideBin(process.argv))
+  const rawArgv = hideBin(process.argv);
+  if (rawArgv.includes("--version") || rawArgv.includes("-v")) {
+    console.log(`Combicode (JavaScript), version ${version}`);
+    process.exit(0);
+  }
+
+  const argv = yargs(rawArgv)
     .scriptName("combicode")
     .usage("$0 [options]")
     .option("o", {
@@ -110,7 +113,12 @@ async function main() {
       type: "boolean",
       default: false,
     })
-    .version(version) // Read version from package.json
+    .option("no-header", {
+      describe: "Omit the introductory prompt and file tree from the output",
+      type: "boolean",
+      default: false,
+    })
+    .version(version)
     .alias("v", "version")
     .help()
     .alias("h", "help").argv;
@@ -155,7 +163,8 @@ async function main() {
 
   const includedFiles = allFiles
     .filter((file) => {
-      if (fs.statSync(file).isDirectory()) return false;
+      const stats = fs.statSync(file, { throwIfNoEntry: false });
+      if (!stats || stats.isDirectory()) return false;
       if (isLikelyBinary(file)) return false;
       if (allowedExtensions && !allowedExtensions.has(path.extname(file)))
         return false;
@@ -168,22 +177,40 @@ async function main() {
     process.exit(1);
   }
 
+  const relativeFiles = includedFiles.map((file) =>
+    path.relative(projectRoot, file)
+  );
+
   if (argv.dryRun) {
-    console.log("\n📋 Files to be included (Dry Run):");
-    includedFiles.forEach((file) =>
-      console.log(`  - ${path.relative(projectRoot, file)}`)
-    );
+    console.log("\n📋 Files to be included (Dry Run):\n");
+    const tree = generateFileTree(relativeFiles, projectRoot);
+    console.log(tree);
     console.log(`\nTotal: ${includedFiles.length} files.`);
     return;
   }
 
   const outputStream = fs.createWriteStream(argv.output);
+
+  if (!argv.noHeader) {
+    outputStream.write(SYSTEM_PROMPT + "\n");
+    outputStream.write("## Project File Tree\n\n");
+    outputStream.write("```\n");
+    const tree = generateFileTree(relativeFiles, projectRoot);
+    outputStream.write(tree);
+    outputStream.write("```\n\n");
+    outputStream.write("---\n\n");
+  }
+
   for (const file of includedFiles) {
     const relativePath = path.relative(projectRoot, file).replace(/\\/g, "/");
     outputStream.write(`// FILE: ${relativePath}` + "\n");
     outputStream.write("```\n");
-    const content = fs.readFileSync(file, "utf8");
-    outputStream.write(content);
+    try {
+      const content = fs.readFileSync(file, "utf8");
+      outputStream.write(content);
+    } catch (e) {
+      outputStream.write(`... (error reading file: ${e.message}) ...`);
+    }
     outputStream.write("\n```\n\n");
   }
   outputStream.end();

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "combicode",
-  "version": "1.0.2",
+  "version": "1.1.0",
   "description": "A CLI tool to combine a project's codebase into a single file for LLM context.",
   "main": "index.js",
   "bin": {
@@ -30,7 +30,6 @@
   "license": "MIT",
   "dependencies": {
     "fast-glob": "^3.3.1",
-    "ignore": "^5.2.4",
     "yargs": "^17.7.2"
   }
 }