feat(captions): add list_system_fonts command with fc-list + fallback

Co-authored-by: Cursor <cursoragent@cursor.com>

src-tauri/src/commands/media_analysis.rs

@@ -1,23 +1,50 @@
-use crate::models::ThumbnailSpritesheet;
+use crate::models::{ThumbnailSpritesheet, WaveformTiers};
+use crate::services::dependency_manager::{ffmpeg_bin, ffprobe_bin};
 use crate::services::{keyframe_index, thumbnail_extractor, waveform_generator};
+use tauri::ipc::Channel;
 
 #[tauri::command]
 pub async fn extract_keyframes(file_path: String) -> Result<Vec<f64>, String> {
-    tokio::task::spawn_blocking(move || keyframe_index::extract_keyframes(&file_path))
+    eprintln!("[video-clipper:keyframes] starting for '{}'", file_path);
+    let result = tokio::task::spawn_blocking(move || keyframe_index::extract_keyframes(&file_path))
         .await
-        .map_err(|e| format!("Task failed: {e}"))?
+        .map_err(|e| format!("Task failed: {e}"))?;
+    match &result {
+        Ok(kf) => eprintln!("[video-clipper:keyframes] done — {} keyframes", kf.len()),
+        Err(e) => eprintln!("[video-clipper:keyframes] FAILED: {}", e),
+    }
+    result
 }
 
+/// Extract waveform at multiple zoom tiers in a single ffmpeg call.
+/// Streams progress (0.0–1.0) via the channel.
 #[tauri::command]
-pub async fn extract_waveform(
+pub async fn extract_waveform_tiers(
     file_path: String,
-    sample_count: usize,
+    duration: f64,
-) -> Result<Vec<f64>, String> {
+    on_progress: Channel<f64>,
-    tokio::task::spawn_blocking(move || {
+) -> Result<WaveformTiers, String> {
-        waveform_generator::extract_waveform(&file_path, sample_count)
+    eprintln!(
+        "[video-clipper:waveform-tiers] starting for '{}' (duration={:.0}s)",
+        file_path, duration
+    );
+    let result = tokio::task::spawn_blocking(move || {
+        waveform_generator::extract_waveform_tiers(&file_path, duration, move |pct| {
+            let _ = on_progress.send(pct);
+        })
     })
     .await
-    .map_err(|e| format!("Task failed: {e}"))?
+    .map_err(|e| format!("Task failed: {e}"))?;
+    match &result {
+        Ok(tiers) => eprintln!(
+            "[video-clipper:waveform-tiers] done — {}/{}/{} peaks",
+            tiers.tier0.len(),
+            tiers.tier1.len(),
+            tiers.tier2.len()
+        ),
+        Err(e) => eprintln!("[video-clipper:waveform-tiers] FAILED: {}", e),
+    }
+    result
 }
 
 #[tauri::command]
@@ -25,7 +52,11 @@ pub async fn extract_thumbnails(
     file_path: String,
     duration: f64,
 ) -> Result<Vec<ThumbnailSpritesheet>, String> {
-    tokio::task::spawn_blocking(move || {
+    eprintln!(
+        "[video-clipper:thumbnails] starting for '{}' (duration={})",
+        file_path, duration
+    );
+    let result = tokio::task::spawn_blocking(move || {
         let temp_dir = std::env::temp_dir()
             .join("video-clipper-thumbs")
             .join(uuid::Uuid::new_v4().to_string());
@@ -33,7 +64,12 @@ pub async fn extract_thumbnails(
         thumbnail_extractor::extract_thumbnails(&file_path, duration, &temp_dir_str)
     })
     .await
-    .map_err(|e| format!("Task failed: {e}"))?
+    .map_err(|e| format!("Task failed: {e}"))?;
+    match &result {
+        Ok(sheets) => eprintln!("[video-clipper:thumbnails] done — {} sheets", sheets.len()),
+        Err(e) => eprintln!("[video-clipper:thumbnails] FAILED: {}", e),
+    }
+    result
 }
 
 #[tauri::command]
@@ -42,3 +78,108 @@ pub async fn cleanup_thumbnails(paths: Vec<String>) -> Result<(), String> {
         .await
         .map_err(|e| format!("Task failed: {e}"))?
 }
+
+#[tauri::command]
+pub async fn check_embedded_subtitles(file_path: String) -> Result<Option<String>, String> {
+    tokio::task::spawn_blocking(move || {
+        use std::process::Command;
+
+        eprintln!("[video-clipper:embedded-subs] checking '{}'", file_path);
+
+        let probe = Command::new(ffprobe_bin())
+            .args([
+                "-v", "quiet",
+                "-print_format", "json",
+                "-show_streams",
+                "-select_streams", "s",
+                &file_path,
+            ])
+            .output()
+            .map_err(|e| format!("Failed to run ffprobe: {e}"))?;
+
+        if !probe.status.success() {
+            return Ok(None);
+        }
+
+        let stdout = String::from_utf8_lossy(&probe.stdout);
+        let parsed: serde_json::Value = serde_json::from_str(&stdout)
+            .map_err(|e| format!("Failed to parse ffprobe output: {e}"))?;
+
+        let streams = parsed.get("streams").and_then(|s| s.as_array());
+        let has_subs = streams.is_some_and(|s| !s.is_empty());
+
+        if !has_subs {
+            eprintln!("[video-clipper:embedded-subs] no subtitle streams found");
+            return Ok(None);
+        }
+
+        eprintln!("[video-clipper:embedded-subs] found subtitle stream, extracting as VTT");
+
+        let output_dir = std::env::temp_dir()
+            .join("video-clipper")
+            .join("subtitles");
+        std::fs::create_dir_all(&output_dir)
+            .map_err(|e| format!("Failed to create subtitle dir: {e}"))?;
+
+        let vtt_path = output_dir
+            .join("embedded.vtt")
+            .to_string_lossy()
+            .to_string();
+
+        let extract = Command::new(ffmpeg_bin())
+            .args(["-y", "-i", &file_path, "-map", "0:s:0", "-f", "webvtt", &vtt_path])
+            .output()
+            .map_err(|e| format!("Failed to extract subtitles: {e}"))?;
+
+        if extract.status.success() {
+            eprintln!("[video-clipper:embedded-subs] extracted to '{}'", vtt_path);
+            Ok(Some(vtt_path))
+        } else {
+            let stderr = String::from_utf8_lossy(&extract.stderr);
+            eprintln!("[video-clipper:embedded-subs] extraction failed: {}", stderr);
+            Ok(None)
+        }
+    })
+    .await
+    .map_err(|e| format!("Task failed: {e}"))?
+}
+
+#[tauri::command]
+pub async fn list_system_fonts() -> Result<Vec<String>, String> {
+    tokio::task::spawn_blocking(|| {
+        // Try fc-list first (available if fontconfig is installed, which ffmpeg depends on)
+        let output = std::process::Command::new("fc-list")
+            .args([":lang=en", "family"])
+            .output();
+
+        match output {
+            Ok(out) if out.status.success() => {
+                let text = String::from_utf8_lossy(&out.stdout);
+                let mut families: Vec<String> = text
+                    .lines()
+                    .flat_map(|line| line.split(','))
+                    .map(|s| s.trim().to_string())
+                    .filter(|s| !s.is_empty())
+                    .collect();
+                families.sort_unstable();
+                families.dedup();
+                Ok(families)
+            }
+            _ => {
+                // Fallback: common cross-platform fonts
+                Ok(vec![
+                    "Arial".to_string(),
+                    "Comic Sans MS".to_string(),
+                    "Courier New".to_string(),
+                    "Georgia".to_string(),
+                    "Helvetica".to_string(),
+                    "Impact".to_string(),
+                    "Times New Roman".to_string(),
+                    "Verdana".to_string(),
+                ])
+            }
+        }
+    })
+    .await
+    .map_err(|e| format!("Font detection failed: {e}"))?
+}

src-tauri/src/lib.rs

@@ -3,6 +3,14 @@ mod models;
 mod services;
 
 use commands::{dependencies, export, media_analysis, video};
+use tauri::Manager;
+
+struct MediaServerPort(u16);
+
+#[tauri::command]
+fn get_media_server_port(state: tauri::State<'_, MediaServerPort>) -> u16 {
+    state.0
+}
 
 #[cfg_attr(mobile, tauri::mobile_entry_point)]
 pub fn run() {
@@ -11,16 +19,31 @@ pub fn run() {
         .plugin(tauri_plugin_dialog::init())
         .plugin(tauri_plugin_store::Builder::new().build())
         .plugin(tauri_plugin_process::init())
+        .setup(|app| {
+            // Bind synchronously (no async needed), then spawn on Tauri's runtime
+            let (port, listener) = services::media_server::bind_media_server()
+                .expect("Failed to bind media server");
+            eprintln!("[video-clipper] media server bound to port {port}");
+            app.manage(MediaServerPort(port));
+            services::media_server::spawn_media_server(listener);
+            Ok(())
+        })
         .invoke_handler(tauri::generate_handler![
             dependencies::check_dependencies,
             dependencies::install_dependency,
+            dependencies::check_subtitles_filter_available,
             video::resolve_url,
             video::start_download,
+            video::check_cached_download,
+            video::download_subtitles,
             media_analysis::extract_keyframes,
-            media_analysis::extract_waveform,
+            media_analysis::extract_waveform_tiers,
             media_analysis::extract_thumbnails,
             media_analysis::cleanup_thumbnails,
+            media_analysis::check_embedded_subtitles,
+            media_analysis::list_system_fonts,
             export::export_clips,
+            get_media_server_port,
         ])
         .run(tauri::generate_context!())
         .expect("error while running tauri application");

src/lib/bindings/mediaAnalysis.ts

@@ -1,4 +1,4 @@
-import { invoke } from '@tauri-apps/api/core';
+import { invoke, Channel } from '@tauri-apps/api/core';
 
 export interface ThumbnailSpritesheet {
   filePath: string;
@@ -10,15 +10,33 @@ export interface ThumbnailSpritesheet {
   intervalSeconds: number;
 }
 
+export interface WaveformTiers {
+  /** ~2,000 peaks — overview for full-duration view */
+  tier0: number[];
+  /** ~10,000 peaks — moderate zoom */
+  tier1: number[];
+  /** All extracted peaks — deep zoom */
+  tier2: number[];
+}
+
 export async function extractKeyframes(filePath: string): Promise<number[]> {
   return invoke<number[]>('extract_keyframes', { filePath });
 }
 
-export async function extractWaveform(
+export async function extractWaveformTiers(
   filePath: string,
-  sampleCount: number
+  duration: number,
-): Promise<number[]> {
+  onProgress?: (pct: number) => void
-  return invoke<number[]>('extract_waveform', { filePath, sampleCount });
+): Promise<WaveformTiers> {
+  const channel = new Channel<number>();
+  if (onProgress) {
+    channel.onmessage = onProgress;
+  }
+  return invoke<WaveformTiers>('extract_waveform_tiers', {
+    filePath,
+    duration,
+    onProgress: channel,
+  });
 }
 
 export async function extractThumbnails(
@@ -34,3 +52,13 @@ export async function extractThumbnails(
 export async function cleanupThumbnails(paths: string[]): Promise<void> {
   return invoke<void>('cleanup_thumbnails', { paths });
 }
+
+export async function checkEmbeddedSubtitles(
+  filePath: string
+): Promise<string | null> {
+  return invoke<string | null>('check_embedded_subtitles', { filePath });
+}
+
+export async function listSystemFonts(): Promise<string[]> {
+  return invoke<string[]>('list_system_fonts');
+}