use anyhow::Result; use std::path::Path; use tokio::time; use tracing::{info, warn}; pub struct WatcherConfig { pub directories: Vec, pub poll_interval_ms: u64, } impl Default for WatcherConfig { fn default() -> Self { let default_dir = std::env::var("MOMENTRY_SFTP_ROOT") .unwrap_or_else(|_| "/Users/accusys/momentry/var/sftpgo/data/demo/".to_string()); Self { directories: vec![default_dir], poll_interval_ms: 60000, } } } /// Starts the file watcher in the background. /// Detects new files and logs them. Does NOT auto-register or auto-pre-process. pub async fn run_watcher() -> Result<()> { let config = WatcherConfig::default(); let dirs = config.directories.clone(); if dirs.is_empty() { warn!("No directories configured for watching."); return Err(anyhow::anyhow!("No watch directories")); } info!("Starting File Watcher (detection only, no auto-modification)..."); info!("Watch directories: {:?}", dirs); tokio::spawn(async move { let mut interval = time::interval(std::time::Duration::from_millis(config.poll_interval_ms)); // Track known files across cycles let mut known = std::collections::HashSet::new(); loop { interval.tick().await; report_new_files(&dirs, &mut known).await; } }); Ok(()) } async fn report_new_files(directories: &[String], known: &mut std::collections::HashSet) { for dir in directories { let dir_path = Path::new(dir); if !dir_path.is_dir() { continue; } let entries = match std::fs::read_dir(dir_path) { Ok(e) => e, _ => continue, }; let mut current_files = std::collections::HashSet::new(); for entry in entries.flatten() { let file_path = entry.path(); if !file_path.is_file() { continue; } let fname = match file_path.file_name().and_then(|n| n.to_str()) { Some(n) if !n.starts_with('.') && !n.ends_with(".pre.json") => n.to_string(), _ => continue, }; current_files.insert(fname.clone()); if !known.contains(&fname) { info!("[WATCHER] New file detected: {} in {}", fname, dir); known.insert(fname); } } // Remove files that no longer exist (deleted between cycles) known.retain(|k| current_files.contains(k)); } } /// Pre-process a single file: compute SHA256 + probe + UUID → .pre.json /// This is called explicitly from register API, NOT from the watcher. pub async fn pre_process_file(file_path: &str) -> Option { let path = std::path::Path::new(file_path); if !path.is_file() { return None; } let canonical = path.canonicalize().ok()?; let canonical_str = canonical.to_string_lossy().to_string(); let filename = path.file_name()?.to_string_lossy().to_string(); let output_dir = std::env::var("MOMENTRY_OUTPUT_DIR") .unwrap_or_else(|_| "/Users/accusys/momentry/output_dev".to_string()); let birthday = std::fs::metadata(&path).ok() .and_then(|m| m.created().ok()) .map(|t| { let secs = t.duration_since(std::time::UNIX_EPOCH).unwrap_or_default().as_secs(); chrono::DateTime::from_timestamp(secs as i64, 0) .map(|dt| dt.to_rfc3339()) .unwrap_or_else(|| chrono::Utc::now().to_rfc3339()) }) .unwrap_or_else(|| chrono::Utc::now().to_rfc3339()); let mac = crate::core::storage::uuid::get_mac_address(); let file_uuid = crate::core::storage::uuid::compute_birth_uuid( &mac, &birthday, &canonical_str, &filename, ); let pre_path = std::path::PathBuf::from(&output_dir).join(format!("{}.pre.json", file_uuid)); if pre_path.exists() { info!("[PRE-PROCESS] Already pre-processed: {}", filename); return Some(file_uuid); } info!("[PRE-PROCESS] Pre-processing: {} → {}", filename, file_uuid); let content_hash = crate::core::storage::content_hash::compute_sha256(&path).unwrap_or_default(); let probe_json: serde_json::Value = if let Ok(result) = crate::core::probe::probe_video(&canonical_str) { serde_json::to_value(&result).unwrap_or_default() } else { let size = std::fs::metadata(&path).ok().map(|m| m.len()).unwrap_or(0); serde_json::json!({ "format": {"filename": canonical_str, "size": size.to_string(), "format_name": "unknown"}, "streams": [] }) }; let file_type = if probe_json.get("streams").and_then(|s| s.as_array()) .map_or(false, |streams| streams.iter().any(|st| st.get("codec_type").and_then(|c| c.as_str()) == Some("video"))) { "video" } else { "unknown" }; let pre_data = serde_json::json!({ "file_name": filename, "file_path": canonical_str, "content_hash": content_hash, "probe_json": probe_json, "birthday": birthday, "file_uuid": file_uuid, "file_size": std::fs::metadata(&path).ok().map(|m| m.len()).unwrap_or(0), "file_type": file_type, "pre_processed_at": chrono::Utc::now().to_rfc3339(), }); if let Ok(content) = serde_json::to_string_pretty(&pre_data) { if std::fs::write(&pre_path, content).is_ok() { info!("[PRE-PROCESS] {} → {}.pre.json", filename, file_uuid); } } Some(file_uuid) }