youtube-legend-cli 0.2.9

//! TTL-keyed local file cache for fetched subtitles.
//!
//! The cache lives under the user's standard cache directory
//! (e.g. `~/.cache/youtube-legend-cli/`) and is keyed on the
//! `(video_id, language, format)` triple. Each entry's freshness is
//! decided by the file's modification time compared to the TTL passed
//! to `cache_path`; expired entries are removed on read.

use crate::error::{AppError, AppResult};
use directories::ProjectDirs;
use std::path::PathBuf;
use std::time::{Duration, SystemTime, UNIX_EPOCH};

const DEFAULT_TTL_HOURS: u64 = 24;
const ENV_QUALIFIER: &str = "YOUTUBE_LEGEND_CLI_AUTHOR";
const FALLBACK_QUALIFIER: &str = "youtube-legend-cli";

/// Build the absolute cache file path for a `(video_id, language, format)`
/// triple under the user's cache directory, creating the parent
/// directory if necessary.
///
/// # Errors
///
/// - [`AppError::InvalidInput`] when any of the components is empty or
///   the TTL is zero.
/// - [`AppError::Internal`] when the platform's project directory cannot
///   be determined.
/// - [`AppError::Io`] when the parent directory cannot be created.
#[tracing::instrument(level = "debug", err, skip(video_id, lang, format), fields(video_id, lang, format, ttl_secs = ttl.as_secs()))]
pub fn cache_path(video_id: &str, lang: &str, format: &str, ttl: Duration) -> AppResult<PathBuf> {
    if video_id.is_empty() || lang.is_empty() || format.is_empty() {
        return Err(AppError::InvalidInput(
            "cache_path requires non-empty video_id, lang, and format".to_string(),
        ));
    }
    if ttl.is_zero() {
        return Err(AppError::InvalidInput(
            "cache_path requires a non-zero ttl".to_string(),
        ));
    }

    let qualifier = qualifier_from_env();
    let proj = ProjectDirs::from("com", &qualifier, FALLBACK_QUALIFIER)
        .ok_or_else(|| AppError::Internal("could not determine cache directory".to_string()))?;

    let dir = proj
        .cache_dir()
        .join("subtitles")
        .join(sanitize(video_id)?)
        .join(sanitize(lang)?);

    std::fs::create_dir_all(&dir)
        .map_err(|e| AppError::Io(std::io::Error::other(format!("creating cache dir: {e}"))))?;

    Ok(dir.join(format!("{}.bin", sanitize(format)?)))
}

fn qualifier_from_env() -> String {
    if let Ok(value) = std::env::var(ENV_QUALIFIER) {
        let trimmed = value.trim();
        if !trimmed.is_empty() {
            return sanitize_qualifier(trimmed);
        }
    }
    if let Ok(home) = std::env::var("HOME") {
        let trimmed = home.trim();
        if !trimmed.is_empty() {
            return sanitize_qualifier(trimmed);
        }
    }
    FALLBACK_QUALIFIER.to_string()
}

fn sanitize_qualifier(input: &str) -> String {
    input
        .chars()
        .map(|c| {
            if c.is_ascii_alphanumeric() || c == '_' || c == '-' || c == '.' {
                c
            } else {
                '_'
            }
        })
        .collect()
}

fn sanitize(input: &str) -> AppResult<String> {
    if input
        .chars()
        .all(|c| c.is_ascii_alphanumeric() || c == '_' || c == '-' || c == '.')
    {
        Ok(input.to_string())
    } else {
        Err(AppError::InvalidInput(format!(
            "invalid path component: {input}"
        )))
    }
}

/// Read a cached entry if it exists and is still fresh.
///
/// Returns `Ok(None)` when the file does not exist or is older than `ttl`
/// (in which case the stale file is also removed).
///
/// # Errors
///
/// - [`AppError::Io`] on any filesystem or metadata read failure.
#[tracing::instrument(level = "debug", err, skip(path), fields(path = %path.display(), ttl_secs = ttl.as_secs()))]
pub async fn read_cache(path: &PathBuf, ttl: Duration) -> AppResult<Option<Vec<u8>>> {
    if !path.exists() {
        return Ok(None);
    }

    let metadata = tokio::fs::metadata(path).await.map_err(AppError::Io)?;
    let modified = metadata
        .modified()
        .map_err(|e| AppError::Io(std::io::Error::other(e.to_string())))?;
    let elapsed = modified
        .duration_since(UNIX_EPOCH)
        .map_err(|e| AppError::Io(std::io::Error::other(e.to_string())))?;
    let now = SystemTime::now()
        .duration_since(UNIX_EPOCH)
        .map_err(|e| AppError::Io(std::io::Error::other(e.to_string())))?;

    if now.saturating_sub(elapsed) > ttl {
        let _ = tokio::fs::remove_file(path).await;
        return Ok(None);
    }

    let bytes = tokio::fs::read(path).await.map_err(AppError::Io)?;
    Ok(Some(bytes))
}

/// Persist `content` to `path`, creating the parent directory if needed.
///
/// # Errors
///
/// - [`AppError::Io`] on any filesystem write failure.
#[tracing::instrument(level = "debug", err, skip(path, content), fields(path = %path.display(), bytes = content.len()))]
pub async fn write_cache(path: &PathBuf, content: &[u8]) -> AppResult<()> {
    if let Some(parent) = path.parent() {
        tokio::fs::create_dir_all(parent)
            .await
            .map_err(AppError::Io)?;
    }
    tokio::fs::write(path, content)
        .await
        .map_err(AppError::Io)?;
    Ok(())
}

/// Remove a cache entry if it exists. A missing entry is not an error.
///
/// # Errors
///
/// - [`AppError::Io`] on filesystem remove failure.
#[tracing::instrument(level = "debug", err, skip(path), fields(path = %path.display()))]
pub async fn invalidate_cache(path: &PathBuf) -> AppResult<()> {
    if path.exists() {
        tokio::fs::remove_file(path).await.map_err(AppError::Io)?;
    }
    Ok(())
}

/// Default TTL of 24 hours, used when the user does not pass `--cache-ttl`.
#[tracing::instrument(level = "debug")]
pub fn default_ttl() -> Duration {
    Duration::from_secs(DEFAULT_TTL_HOURS * 3600)
}

#[cfg(test)]
mod tests {
    use super::*;
    use serial_test::serial;

    #[test]
    fn default_ttl_is_24_hours() {
        assert_eq!(default_ttl(), Duration::from_secs(24 * 3600));
    }

    #[test]
    #[serial]
    fn qualifier_prefers_env_var() {
        // SAFETY: `std::env::set_var` is `unsafe` in Rust 2024 because it
        // mutates process-global state. The `#[serial]` attribute from
        // `serial_test` ensures no other test in this binary runs
        // concurrently; the env var is removed at the end of the test.
        unsafe {
            std::env::set_var(ENV_QUALIFIER, "test-author");
        }
        let q = qualifier_from_env();
        // SAFETY: same as above, paired with the set_var above.
        unsafe {
            std::env::remove_var(ENV_QUALIFIER);
        }
        assert_eq!(q, "test-author");
    }

    #[test]
    #[serial]
    fn qualifier_falls_back_to_home() {
        let original = std::env::var(ENV_QUALIFIER).ok();
        // SAFETY: `std::env::set_var` / `remove_var` mutate process-global
        // state. The `#[serial]` attribute ensures exclusive access; the
        // original value is captured and restored before returning.
        unsafe {
            std::env::remove_var(ENV_QUALIFIER);
            std::env::set_var("HOME", "/home/test-user");
        }
        let q = qualifier_from_env();
        // SAFETY: restoration paired with the set_var above.
        unsafe {
            std::env::remove_var("HOME");
            if let Some(v) = original {
                std::env::set_var(ENV_QUALIFIER, v);
            }
        }
        assert!(q.contains("test-user") || q == FALLBACK_QUALIFIER);
    }

    #[test]
    fn qualifier_sanitizes_invalid_chars() {
        let s = sanitize_qualifier("hello world/foo");
        assert_eq!(s, "hello_world_foo");
    }

    #[test]
    fn cache_path_rejects_zero_ttl() {
        let res = cache_path("vid12345678", "en", "txt", Duration::ZERO);
        assert!(matches!(res, Err(AppError::InvalidInput(_))));
    }

    #[test]
    fn cache_path_rejects_empty_components() {
        let res = cache_path("", "en", "txt", default_ttl());
        assert!(matches!(res, Err(AppError::InvalidInput(_))));
    }

    #[test]
    fn sanitize_accepts_safe_chars() {
        assert_eq!(sanitize("video_123-abc.txt").unwrap(), "video_123-abc.txt");
    }

    #[test]
    fn sanitize_rejects_unsafe_chars() {
        assert!(sanitize("../etc/passwd").is_err());
        assert!(sanitize("with space").is_err());
    }
}