nigig-org/crates/copy_code.rs

use std::collections::HashSet;
use std::fs::{self, File};
use std::io::{self, Read, Write};
use std::path::{Path, PathBuf};
use std::process::Command;

/// Directories to always skip, regardless of .gitignore
const IGNORED_DIRS: &[&str] = &[
    "target",
    "node_modules",
    // "resources",
    ".git",
    "dist",
    "build",
    ".next",
    ".nuxt",
    ".cache",
    ".turbo",
    "out",
    "pkg",
    "vendor",
    "coverage",
    "__pycache__",
];

/// Files to always skip
const IGNORED_FILES: &[&str] = &[
    "package-lock.json",
    "yarn.lock",
    "pnpm-lock.yaml",
    "*.txt",           // All cad files
    "Cargo.lock",
    "map.txt",
    "security_tests.rs",
    "security_suite.rs",
];

/// Specific files/directories to always include (with specific extensions)
const ADD_FILES: &[&str] = &[
    "*.java",           // All Java files
    "*.cad",           // All cad files
    "*.splash",           // All cad files
    "*.py",           // All cad files
    "*.obj",           // All obj files
    "*.kt",             // Kotlin files
    "*.gradle",         // Gradle build files
    "*.xml",            // XML configuration files
    "*.properties",     // Properties files
    "pom.xml",          // Maven configuration
    "build.gradle",     // Gradle configuration
    "settings.gradle",  // Gradle settings
];

fn is_ignored_dir(path: &Path) -> bool {
    path.file_name()
        .and_then(|name| name.to_str())
        .map(|name| IGNORED_DIRS.contains(&name))
        .unwrap_or(false)
}

fn is_ignored_file(path: &Path) -> bool {
    path.file_name()
        .and_then(|name| name.to_str())
        .map(|name| IGNORED_FILES.contains(&name))
        .unwrap_or(false)
}

fn is_hidden(path: &Path) -> bool {
    path.file_name()
        .and_then(|name| name.to_str())
        .map(|name| name.starts_with('.'))
        .unwrap_or(false)
}

/// Matches this tool's own output files, e.g. "pdf_concatenated_code.txt" or the
/// older unprefixed "concatenated_code.txt". Any such file found while scanning
/// (left over from an earlier run, possibly with a different folder ordering) is
/// skipped so its contents never get sucked back in as if it were source code.
fn is_previous_output_file(path: &Path) -> bool {
    path.file_name()
        .and_then(|name| name.to_str())
        .map(|name| name == "concatenated_code.txt" || name.ends_with("_concatenated_code.txt"))
        .unwrap_or(false)
}

fn is_git_ignored(root: &Path, path: &Path) -> bool {
    let relative_path = path.strip_prefix(root).unwrap_or(path);
    let relative_path_str = match relative_path.to_str() {
        Some(s) => s,
        None => return false,
    };

    let output = Command::new("git")
        .args(["check-ignore", "-q", relative_path_str])
        .current_dir(root)
        .output();

    match output {
        Ok(output) => output.status.success(),
        Err(_) => false, // git not available — don't skip
    }
}

fn matches_add_files_pattern(path: &Path) -> bool {
    let file_name = match path.file_name().and_then(|n| n.to_str()) {
        Some(name) => name,
        None => return false,
    };

    let path_str = path.to_str().unwrap_or("");

    for pattern in ADD_FILES {
        if pattern.starts_with("*.") {
            // Handle wildcard extension patterns like "*.java"
            let ext = &pattern[2..]; // Remove the "*."
            if path.extension()
                .and_then(|e| e.to_str())
                .map(|e| e == ext)
                .unwrap_or(false) {
                return true;
            }
        } else {
            // Handle exact file name matches
            if file_name == *pattern || path_str.ends_with(pattern) {
                return true;
            }
        }
    }
    false
}

fn should_skip(root: &Path, path: &Path) -> bool {
    // Check if file should be included based on ADD_FILES patterns
    if path.is_file() && matches_add_files_pattern(path) {
        return false; // Don't skip files that match ADD_FILES patterns
    }

    if is_hidden(path) {
        println!("  skip (hidden):       {}", path.display());
        return true;
    }

    if path.is_dir() && is_ignored_dir(path) {
        println!("  skip (ignored dir):  {}", path.display());
        return true;
    }

    if path.is_file() && is_ignored_file(path) {
        println!("  skip (ignored file): {}", path.display());
        return true;
    }

    if path.is_file() && is_previous_output_file(path) {
        println!("  skip (old output):   {}", path.display());
        return true;
    }

    if is_git_ignored(root, path) {
        println!("  skip (gitignored):   {}", path.display());
        return true;
    }

    false
}

const VALID_EXTENSIONS: &[&str] = &[
    "rs", "toml", "mjs", "md", "txt", "js", "ts", "tsx", "html", "css", "json", "yaml", "yml",
    "env", "sh", "sql", "graphql", "proto", "dockerfile", "java", "kt", "gradle", "xml", "properties",
];

fn has_valid_extension(path: &Path) -> bool {
    // Also allow files with no extension but known names
    let known_extensionless = &["Makefile", "Dockerfile", "Justfile", ".env", ".env.example", "pom.xml", "build.gradle", "settings.gradle"];

    if let Some(name) = path.file_name().and_then(|n| n.to_str()) {
        if known_extensionless.contains(&name) {
            return true;
        }
    }

    path.extension()
        .and_then(|ext| ext.to_str())
        .map(|ext| VALID_EXTENSIONS.contains(&ext.to_lowercase().as_str()))
        .unwrap_or(false)
}

/// Short label for a root folder, used both to build the output file name and to
/// prefix each "// File:" header so it's obvious which folder a block came from
/// once multiple folders are combined into one output.
fn folder_label(path: &Path) -> String {
    path.file_name()
        .and_then(|n| n.to_str())
        .map(|s| s.to_string())
        .unwrap_or_else(|| "root".to_string())
}

fn copy_files_to_single_file(
    source_dir: &Path,
    output_file: &mut File,
    root: &Path,
    output_file_abs_path: &Path,
    seen_files: &mut HashSet<PathBuf>,
) -> Result<(), io::Error> {
    // Read and sort entries for deterministic output
    let mut entries: Vec<_> = fs::read_dir(source_dir)?
        .filter_map(|e| e.ok())
        .collect();
    entries.sort_by_key(|e| e.path());

    for entry in entries {
        let path = entry.path();

        if should_skip(root, &path) {
            continue;
        }

        if path.is_dir() {
            copy_files_to_single_file(&path, output_file, root, output_file_abs_path, seen_files)?;
        } else if path.is_file() {
            let file_abs_path = fs::canonicalize(&path)?;

            // Skip the output file itself
            if file_abs_path == output_file_abs_path {
                println!("  skip (output file):  {}", path.display());
                continue;
            }

            if !has_valid_extension(&path) && !matches_add_files_pattern(&path) {
                println!("  skip (extension):    {}", path.display());
                continue;
            }

            // If two provided root folders overlap (e.g. one nested inside another),
            // make sure the same file never gets written into the output twice.
            if !seen_files.insert(file_abs_path) {
                println!("  skip (duplicate):    {}", path.display());
                continue;
            }

            let relative_path = path.strip_prefix(root).unwrap_or(&path);
            let label = folder_label(root);
            writeln!(output_file, "// File: {}/{}\n", label, relative_path.display())?;

            let mut file = File::open(&path)?;
            let mut contents = String::new();
            file.read_to_string(&mut contents)?;
            writeln!(output_file, "{}\n", contents)?;

            println!("  ✓ included:          {}/{}", label, relative_path.display());
        }
    }
    Ok(())
}

fn copy_single_file(
    path: &Path,
    output_file: &mut File,
    output_file_abs_path: &Path,
    seen_files: &mut HashSet<PathBuf>,
) -> io::Result<()> {
    if should_skip(path.parent().unwrap_or(Path::new(".")), path) {
        return Ok(());
    }

    let file_abs_path = fs::canonicalize(path)?;

    if file_abs_path == output_file_abs_path {
        println!("  skip (output file): {}", path.display());
        return Ok(());
    }

    if !has_valid_extension(path) && !matches_add_files_pattern(path) {
        println!("  skip (extension): {}", path.display());
        return Ok(());
    }

    if !seen_files.insert(file_abs_path) {
        println!("  skip (duplicate): {}", path.display());
        return Ok(());
    }

    writeln!(output_file, "// File: {}\n", path.display())?;

    let mut file = File::open(path)?;
    let mut contents = String::new();
    file.read_to_string(&mut contents)?;
    writeln!(output_file, "{}\n", contents)?;

    println!("  ✓ included: {}", path.display());

    Ok(())
}

fn main() -> Result<(), io::Error> {
    // Add as many folders as you like here — they all get combined into one output file.
    // let project_roots: Vec<PathBuf> = vec![
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/nigig-org/crates/apps/nigig-pay"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/nigig-org/crates/apps/nigig-pay-ui"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/nigig-org/crates/nigig-core"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/robius-sms"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/robius-fingerprinting"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/robius-trigger"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/robius-ussd"),
    //     PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/robius-contacts"),
    // ];
    let project_roots: Vec<PathBuf> = vec![
        PathBuf::from("/Users/aok/Projects/rustdev/CratesCode/nigig-org/crates/apps/nigig-build/src/construction_frame/pages/workspace/cost_estimator"),
    ];
    

    if project_roots.is_empty() {
        eprintln!("No project roots specified — nothing to do.");
        return Ok(());
    }

    // The output file is named after the first (or only) folder, and is written
    // inside that folder. Change `output_dir` below if you'd rather it land somewhere else.
    let first_root = &project_roots[0];
    let output_dir = first_root.clone();
    let output_file_name = format!("{}_concatenated_code.txt", folder_label(first_root));
    let output_file_path = output_dir.join(&output_file_name);

    // Ensure the output file exists so canonicalize works
    if !output_file_path.exists() {
        File::create(&output_file_path)?;
    }
    let output_file_abs_path = fs::canonicalize(&output_file_path)?;

    let mut output_file = File::create(&output_file_path)?;

    println!("╔══════════════════════════════════════════╗");
    println!("║  Code Concatenation Tool                 ║");
    println!("╚══════════════════════════════════════════╝");
    println!();
    println!("Roots ({}):", project_roots.len());
    for r in &project_roots {
        println!("  - {}", r.display());
    }
    println!("Output: {}", output_file_abs_path.display());
    println!();
    println!("Ignored dirs:  {}", IGNORED_DIRS.join(", "));
    println!("Ignored files: {}", IGNORED_FILES.join(", "));
    println!("Add files:     {}", ADD_FILES.join(", "));
    println!("Valid exts:    {}", VALID_EXTENSIONS.join(", "));
    println!();

    // Shared across every root so the same file can't end up written twice even if
    // two provided folders overlap.
    let mut seen_files: HashSet<PathBuf> = HashSet::new();

    for root in &project_roots {
        println!("── Scanning: {} ──", root.display());

        if root.is_dir() {
            copy_files_to_single_file(
                root,
                &mut output_file,
                root,
                &output_file_abs_path,
                &mut seen_files,
            )?;
        } else if root.is_file() {
            copy_single_file(
                root,
                &mut output_file,
                &output_file_abs_path,
                &mut seen_files,
            )?;
        } else {
            println!("  skip (missing): {}", root.display());
        }
    }

    // Print summary
    let output_size = fs::metadata(&output_file_path)?.len();
    println!();
    println!("════════════════════════════════════════════");
    println!(
        "Done. Output: {} ({} bytes)",
        output_file_path.display(),
        output_size
    );

    Ok(())
}