Voiced/src-tauri/src/commands/assets.rs

use crate::AppState;
use base64::{engine::general_purpose::STANDARD, Engine};
use std::path::{Component, Path, PathBuf};
use tauri::State;

const FORMAT_ERROR: &str = "Use a PNG or JPEG image";

fn safe_kind(kind: &str) -> String {
    kind.chars()
        .filter(|c| c.is_ascii_alphanumeric() || *c == '-' || *c == '_')
        .collect::<String>()
        .to_lowercase()
}

/// Identify an image by its magic bytes. react-pdf only renders PNG and JPEG and
/// silently drops everything else, so nothing else is accepted.
/// Returns (extension, MIME type).
pub fn sniff_image(bytes: &[u8]) -> Result<(&'static str, &'static str), String> {
    if bytes.starts_with(&[0x89, b'P', b'N', b'G', 0x0D, 0x0A, 0x1A, 0x0A]) {
        Ok(("png", "image/png"))
    } else if bytes.starts_with(&[0xFF, 0xD8, 0xFF]) {
        Ok(("jpg", "image/jpeg"))
    } else {
        Err(FORMAT_ERROR.to_string())
    }
}

fn assets_dir(data_dir: &Path) -> Result<PathBuf, String> {
    let dir = data_dir.join("assets");
    std::fs::create_dir_all(&dir).map_err(|e| e.to_string())?;
    Ok(dir)
}

/// Resolve a stored asset path (relative to the data dir, or a legacy absolute one)
/// to a real file inside `<data_dir>/assets`. `..`, symlink escapes and anything
/// outside the assets directory are rejected.
pub fn resolve_asset(data_dir: &Path, path: &str) -> Result<PathBuf, String> {
    let raw = Path::new(path);
    if path.trim().is_empty() {
        return Err("Asset path is empty".into());
    }
    if raw.components().any(|c| matches!(c, Component::ParentDir)) {
        return Err("Asset path must not contain '..'".into());
    }
    let full = if raw.is_absolute() { raw.to_path_buf() } else { data_dir.join(raw) };
    let root = data_dir
        .join("assets")
        .canonicalize()
        .map_err(|e| format!("Asset store is unavailable: {e}"))?;
    let canonical = full
        .canonicalize()
        .map_err(|e| format!("Could not open asset {path}: {e}"))?;
    if canonical.starts_with(&root) && canonical != root {
        Ok(canonical)
    } else {
        Err("Asset path is outside the asset store".into())
    }
}

/// The form stored in the database: `assets/<file>` relative to the data dir, with
/// forward slashes on every platform.
pub fn relative_asset_path(data_dir: &Path, path: &str) -> Result<String, String> {
    let canonical = resolve_asset(data_dir, path)?;
    let base = data_dir.canonicalize().map_err(|e| e.to_string())?;
    let rel = canonical.strip_prefix(&base).map_err(|e| e.to_string())?;
    Ok(rel
        .components()
        .map(|c| c.as_os_str().to_string_lossy().into_owned())
        .collect::<Vec<_>>()
        .join("/"))
}

fn write_asset(data_dir: &Path, kind: &str, bytes: &[u8]) -> Result<String, String> {
    let (ext, _) = sniff_image(bytes)?;
    let dir = assets_dir(data_dir)?;
    let file = format!("{}-{}.{}", safe_kind(kind), uuid::Uuid::new_v4().simple(), ext);
    std::fs::write(dir.join(&file), bytes).map_err(|e| e.to_string())?;
    Ok(format!("assets/{file}"))
}

/// Write `assets/<file_name>` unless it already exists. For content-addressed derivatives, where the
/// same name always means the same bytes. The name must be a plain `[a-z0-9._-]` file name.
pub fn write_named_asset(data_dir: &Path, file_name: &str, bytes: &[u8]) -> Result<String, String> {
    let plain = !file_name.is_empty()
        && !file_name.starts_with('.')
        && file_name
            .chars()
            .all(|c| c.is_ascii_lowercase() || c.is_ascii_digit() || matches!(c, '.' | '-' | '_'));
    if !plain {
        return Err(format!("Invalid asset name: {file_name}"));
    }
    let target = assets_dir(data_dir)?.join(file_name);
    if !target.exists() {
        std::fs::write(&target, bytes).map_err(|e| e.to_string())?;
    }
    Ok(format!("assets/{file_name}"))
}

fn asset_data_uri(data_dir: &Path, path: &str) -> Result<String, String> {
    let file = resolve_asset(data_dir, path)?;
    let bytes = std::fs::read(&file).map_err(|e| e.to_string())?;
    let (_, mime) = sniff_image(&bytes)?;
    Ok(format!("data:{};base64,{}", mime, STANDARD.encode(bytes)))
}

fn remove_asset_file(data_dir: &Path, path: &str) -> Result<(), String> {
    let raw = Path::new(path);
    let full = if raw.is_absolute() { raw.to_path_buf() } else { data_dir.join(raw) };
    // Removing something that is already gone is fine, but an escape attempt is not.
    if std::fs::symlink_metadata(&full).is_err() {
        return Ok(());
    }
    let target = resolve_asset(data_dir, path)?;
    std::fs::remove_file(target).map_err(|e| e.to_string())
}

/// Copy a user-picked file into the app's asset store and return the stored path.
#[tauri::command]
pub fn import_asset(
    state: State<AppState>,
    source_path: String,
    kind: String,
) -> Result<String, String> {
    let bytes = std::fs::read(&source_path).map_err(|e| e.to_string())?;
    write_asset(&state.data_dir, &kind, &bytes)
}

/// Used when the frontend already holds the bytes (base64) instead of a path.
/// The file name is ignored: the type comes from the bytes.
#[tauri::command]
pub fn save_asset_bytes(
    state: State<AppState>,
    kind: String,
    #[allow(unused_variables)] file_name: String,
    data_base64: String,
) -> Result<String, String> {
    let bytes = STANDARD
        .decode(data_base64.as_bytes())
        .map_err(|e| e.to_string())?;
    write_asset(&state.data_dir, &kind, &bytes)
}

/// Read a stored asset back as a data URI so it can be embedded in the PDF.
#[tauri::command]
pub fn read_asset_data_uri(state: State<AppState>, path: String) -> Result<String, String> {
    asset_data_uri(&state.data_dir, &path)
}

#[tauri::command]
pub fn remove_asset(state: State<AppState>, path: String) -> Result<(), String> {
    remove_asset_file(&state.data_dir, &path)
}

#[cfg(test)]
mod tests {
    use super::*;
    use tempfile::tempdir;

    const PNG: &[u8] = &[0x89, b'P', b'N', b'G', 0x0D, 0x0A, 0x1A, 0x0A, 0, 0, 0, 0];
    const JPEG: &[u8] = &[0xFF, 0xD8, 0xFF, 0xE0, 0, 0x10, b'J', b'F', b'I', b'F'];

    #[test]
    fn sniffing_accepts_only_png_and_jpeg() {
        assert_eq!(sniff_image(PNG).unwrap(), ("png", "image/png"));
        assert_eq!(sniff_image(JPEG).unwrap(), ("jpg", "image/jpeg"));
        assert_eq!(sniff_image(b"GIF89a....").unwrap_err(), "Use a PNG or JPEG image");
        assert!(sniff_image(b"RIFF\x00\x00\x00\x00WEBPVP8 ").is_err());
        assert!(sniff_image(b"<svg xmlns='http://www.w3.org/2000/svg'/>").is_err());
        assert!(sniff_image(b"").is_err());
    }

    #[test]
    fn write_ignores_the_name_and_stores_a_relative_path() {
        let dir = tempdir().unwrap();
        let stored = write_asset(dir.path(), "Logo", JPEG).unwrap();
        assert!(stored.starts_with("assets/logo-") && stored.ends_with(".jpg"), "{stored}");
        assert!(dir.path().join(&stored).is_file());
        assert!(write_asset(dir.path(), "logo", b"GIF89a").is_err());
        let uri = asset_data_uri(dir.path(), &stored).unwrap();
        assert!(uri.starts_with("data:image/jpeg;base64,"));
    }

    #[test]
    fn named_assets_are_plain_names_and_not_overwritten() {
        let dir = tempdir().unwrap();
        let stored = write_named_asset(dir.path(), "logo-ab12cd34-print.png", b"one").unwrap();
        assert_eq!(stored, "assets/logo-ab12cd34-print.png");
        write_named_asset(dir.path(), "logo-ab12cd34-print.png", b"two").unwrap();
        assert_eq!(std::fs::read(dir.path().join(&stored)).unwrap(), b"one");
        for bad in ["", ".hidden", "../x.png", "a/b.png", "A.png", "x y.png"] {
            assert!(write_named_asset(dir.path(), bad, b"x").is_err(), "{bad}");
        }
    }

    #[test]
    fn data_uri_mime_comes_from_the_bytes_not_the_extension() {
        let dir = tempdir().unwrap();
        let assets = dir.path().join("assets");
        std::fs::create_dir_all(&assets).unwrap();
        std::fs::write(assets.join("liar.jpg"), PNG).unwrap();
        let uri = asset_data_uri(dir.path(), "assets/liar.jpg").unwrap();
        assert!(uri.starts_with("data:image/png;base64,"));
    }

    #[test]
    fn resolver_accepts_relative_and_inside_absolute_paths() {
        let dir = tempdir().unwrap();
        let stored = write_asset(dir.path(), "logo", PNG).unwrap();
        let by_relative = resolve_asset(dir.path(), &stored).unwrap();
        let absolute = dir.path().join(&stored);
        let by_absolute = resolve_asset(dir.path(), absolute.to_str().unwrap()).unwrap();
        assert_eq!(by_relative, by_absolute);
        assert_eq!(relative_asset_path(dir.path(), absolute.to_str().unwrap()).unwrap(), stored);
    }

    #[test]
    fn resolver_rejects_escapes() {
        let dir = tempdir().unwrap();
        let stored = write_asset(dir.path(), "logo", PNG).unwrap();
        std::fs::write(dir.path().join("voiced.db"), b"secret").unwrap();
        let outside = tempdir().unwrap();
        std::fs::write(outside.path().join("x.png"), PNG).unwrap();

        assert!(resolve_asset(dir.path(), "../voiced.db").is_err());
        assert!(resolve_asset(dir.path(), "assets/../voiced.db").is_err());
        assert!(resolve_asset(dir.path(), "voiced.db").is_err());
        assert!(resolve_asset(dir.path(), "assets").is_err());
        assert!(resolve_asset(dir.path(), "").is_err());
        let abs_outside = outside.path().join("x.png");
        assert!(resolve_asset(dir.path(), abs_outside.to_str().unwrap()).is_err());
        let abs_db = dir.path().join("voiced.db");
        assert!(resolve_asset(dir.path(), abs_db.to_str().unwrap()).is_err());
        assert!(resolve_asset(dir.path(), "assets/missing.png").is_err());
        assert!(resolve_asset(dir.path(), &stored).is_ok());
    }

    #[cfg(unix)]
    #[test]
    fn resolver_rejects_symlink_escapes() {
        let dir = tempdir().unwrap();
        let _ = write_asset(dir.path(), "logo", PNG).unwrap();
        let outside = tempdir().unwrap();
        let target = outside.path().join("x.png");
        std::fs::write(&target, PNG).unwrap();
        std::os::unix::fs::symlink(&target, dir.path().join("assets/link.png")).unwrap();
        std::os::unix::fs::symlink(outside.path(), dir.path().join("assets/dir")).unwrap();

        assert!(resolve_asset(dir.path(), "assets/link.png").is_err());
        assert!(resolve_asset(dir.path(), "assets/dir/x.png").is_err());
        assert!(remove_asset_file(dir.path(), "assets/link.png").is_err());
        assert!(target.exists());
    }

    #[test]
    fn remove_deletes_inside_and_tolerates_missing() {
        let dir = tempdir().unwrap();
        let stored = write_asset(dir.path(), "logo", PNG).unwrap();
        remove_asset_file(dir.path(), &stored).unwrap();
        assert!(!dir.path().join(&stored).exists());
        remove_asset_file(dir.path(), &stored).unwrap();
        std::fs::write(dir.path().join("voiced.db"), b"x").unwrap();
        assert!(remove_asset_file(dir.path(), "voiced.db").is_err());
        assert!(dir.path().join("voiced.db").exists());
    }
}