Remove --release flag (#22)

Also don't generate the full length videos during `watch`. Only during
`build`.

Author: rikhuijzer

.github/workflows/examples.yml

@@ -45,17 +45,17 @@ jobs:
             convert -size 500x500 xc:none -fill white -draw "polygon 150,100 150,400 400,250" play_button.png
 
             ./examples/demo.sh
-            mv _out/release.mp4 _public/demo.mp4
+            mv _out/out.mp4 _public/demo.mp4
             convert _out/image/1.png play_button.png -gravity center -composite tmp.png
             mv tmp.png _public/demo.png
 
             ./examples/google.sh
-            mv _out/release.mp4 _public/google.mp4
+            mv _out/out.mp4 _public/google.mp4
             convert _out/image/1.png play_button.png -gravity center -composite tmp.png
             mv tmp.png _public/google.png
 
             ./examples/first.sh
-            mv _out/release.mp4 _public/first.mp4
+            mv _out/out.mp4 _public/first.mp4
             convert _out/image/1.png play_button.png -gravity center -composite tmp.png
             mv tmp.png _public/first.png
 

README.md

@@ -32,7 +32,7 @@ To create a video, create a Typst presentation with speaker notes (we show only
 Next, run the following command:
 
 ```raw
-$ trv --release build examples/first.typ
+$ trv build examples/first.typ
 ```
 
 This generates the following video:
@@ -93,7 +93,7 @@ Google has some high-quality voices available via their API:
 ```raw
 $ export GOOGLE_KEY="<YOUR KEY>"
 
-$ trv --release build examples/google.typ
+$ trv build examples/google.typ
 ```
 
 [![Google demo video](https://transformrs.github.io/trv/google.png)](https://transformrs.github.io/trv/google.mp4)
@@ -121,7 +121,7 @@ So it's probably easiest to use DeepInfra:
 ```raw
 $ export DEEPINFRA_KEY="<YOUR KEY>"
 
-$ trv --release build presentation.typ
+$ trv build presentation.typ
 ```
 
 Do note that Zonos is way more unstable than Kokoros at the time of writing.

examples/demo.sh

@@ -4,4 +4,4 @@
 
 export GOOGLE_KEY=$(cat keys.env | grep GOOGLE_KEY | cut -d '=' -f 2)
 
-trv --release build examples/demo.typ
+trv build examples/demo.typ

examples/demo.typ

@@ -79,6 +79,13 @@
       #import "@preview/polylux:0.4.0": *
       #set page(paper: "presentation-16-9")
 
+      // --- trv config:
+      // provider = "openai-compatible(kokoros.transformrs.org)"
+      // model = "tts-1"
+      // voice = "af_sky"
+      // audio_format = "wav"
+      // ---
+
       #slide[
         Hello
 
@@ -103,13 +110,7 @@
     #set page(margin: 3em)
     #set text(size: 24pt)
     ```bash
-    $ trv \
-      --input presentation.typ \
-      --provider='openai-compatible(kokoros.transformrs.org)' \
-      --model=tts-1 \
-      --voice=af_sky \
-      --audio-format=wav \
-      --release
+    $ trv build presentation.typ
     ```
 
     #toolbox.pdfpc.speaker-note("

examples/first.sh

@@ -2,4 +2,4 @@
 
 # Run via `./examples/first.sh`
 
-trv --release build examples/first.typ
+trv build examples/first.typ

examples/google.sh

@@ -4,4 +4,4 @@
 
 export GOOGLE_KEY=$(cat keys.env | grep GOOGLE_KEY | cut -d '=' -f 2)
 
-trv --release build examples/google.typ
+trv build examples/google.typ

src/main.rs

@@ -91,15 +91,32 @@ fn test_parse_config() {
 }
 
 #[derive(Clone, Debug, Parser)]
-struct BuildArgs {
+pub(crate) struct BuildArgs {
     /// Path to the Typst input file.
     input: PathBuf,
+
+    /// Audio codec.
+    ///
+    /// This setting is passed to ffmpeg.
+    ///
+    /// Opus generally gives the best quality for the lowest file size, but is
+    /// not supported by all platforms. For example, Whatsapp Web and X don't
+    /// accept it.
+    ///
+    /// So therefore on MacOS set the value to `aac_at` and on Linux to
+    /// `libfdk_aac`.
+    #[arg(long, default_value = "opus")]
+    audio_codec: String,
 }
 
 #[derive(Clone, Debug, Parser)]
-struct WatchArgs {
+pub(crate) struct WatchArgs {
     /// Path to the Typst input file.
     input: PathBuf,
+
+    /// Port to run the server on.
+    #[arg(long, default_value = "8080")]
+    port: u16,
 }
 
 #[derive(Clone, Debug, clap::Subcommand)]
@@ -128,26 +145,6 @@ pub(crate) struct Arguments {
     /// Enable caching.
     #[arg(long, default_value = "true")]
     cache: Option<bool>,
-
-    /// Release.
-    ///
-    /// If true, attempt to convert the output video into a format that is more
-    /// widely supported.
-    #[arg(long, default_value = "false")]
-    release: bool,
-
-    /// Audio codec.
-    ///
-    /// This setting is passed to ffmpeg.
-    ///
-    /// Opus generally gives the best quality for the lowest file size, but is
-    /// not supported by all platforms. For example, Whatsapp Web and X don't
-    /// accept it.
-    ///
-    /// So therefore on MacOS set the value to `aac_at` and on Linux to
-    /// `libfdk_aac`.
-    #[arg(long, default_value = "opus")]
-    audio_codec: String,
 }
 
 // TODO: This logic should be in the transformrs crate as `Provider::from_str`.
@@ -218,7 +215,12 @@ fn copy_input_with_includes(dir: &str, input: &PathBuf) -> PathBuf {
     output_path
 }
 
-pub(crate) async fn build(input: PathBuf, args: &Arguments) -> Vec<Slide> {
+pub(crate) async fn build(
+    input: PathBuf,
+    args: &Arguments,
+    release: bool,
+    audio_codec: Option<String>,
+) -> Vec<Slide> {
     let out_dir = &args.out_dir;
     let copied_input = copy_input_with_includes(out_dir, &input);
     let config = parse_config(&copied_input);
@@ -258,9 +260,10 @@ pub(crate) async fn build(input: PathBuf, args: &Arguments) -> Vec<Slide> {
     )
     .await;
     let output = "out.mp4";
-    video::generate_video(out_dir, &slides, cache, &tts_config, output, &audio_ext);
-    if args.release {
-        video::generate_release_video(out_dir, output, "release.mp4", &args.audio_codec);
+    video::create_video_clips(out_dir, &slides, cache, &tts_config, &audio_ext);
+    if release {
+        let audio_codec = audio_codec.unwrap();
+        video::combine_video(out_dir, &slides, output, &audio_codec);
     }
     slides
 }
@@ -282,8 +285,10 @@ async fn main() {
 
     match args.task {
         Task::Build(ref build_args) => {
-            let _ = build(build_args.input.clone(), &args).await;
+            let release = true;
+            let audio_codec = Some(build_args.audio_codec.clone());
+            let _ = build(build_args.input.clone(), &args, release, audio_codec).await;
         }
-        Task::Watch(ref watch_args) => watch(watch_args.input.clone(), &args).await,
+        Task::Watch(ref watch_args) => watch(watch_args, &args).await,
     };
 }

src/video.rs

@@ -78,6 +78,19 @@ fn generate_concat_list(dir: &str, slides: &Vec<Slide>) -> String {
     lines.join("\n")
 }
 
+fn set_default_ffmpeg_video_args(cmd: &mut std::process::Command) {
+    cmd.arg("-c:v")
+        .arg("libx264")
+        .arg("-crf")
+        .arg("23")
+        .arg("-preset")
+        .arg("fast")
+        .arg("-vf")
+        .arg(format!("scale=-1:{HEIGHT},format=yuv420p"))
+        .arg("-pix_fmt")
+        .arg("yuv420p");
+}
+
 fn write_concat_list(dir: &str, path: &str, slides: &Vec<Slide>) {
     let concat_list = generate_concat_list(dir, slides);
     std::fs::write(path, concat_list).expect("couldn't write concat list");
@@ -101,34 +114,24 @@ fn create_video_clip(dir: &str, slide: &Slide, cache: bool, config: &TTSConfig,
         return;
     }
     let output_video = crate::path::video_path(dir, slide);
-    let output = std::process::Command::new("ffmpeg")
-        .arg("-y")
+    let mut cmd = std::process::Command::new("ffmpeg");
+    cmd.arg("-y")
         .arg("-loop")
         .arg("1")
         .arg("-i")
         .arg(input_image)
         .arg("-i")
-        .arg(input_audio)
-        .arg("-c:v")
-        .arg("libx264")
-        .arg("-crf")
-        .arg("23")
-        .arg("-preset")
-        .arg("fast")
-        .arg("-vf")
-        .arg(format!("scale=-1:{HEIGHT},format=yuv420p"))
-        .arg("-pix_fmt")
-        .arg("yuv420p")
-        .arg("-c:a")
+        .arg(input_audio);
+    set_default_ffmpeg_video_args(&mut cmd);
+    cmd.arg("-c:a")
         .arg("opus")
         .arg("-strict")
         .arg("experimental")
         .arg("-shortest")
         .arg("-tune")
         .arg("stillimage")
-        .arg(output_video.clone())
-        .output()
-        .expect("Failed to run ffmpeg command");
+        .arg(output_video.clone());
+    let output = cmd.output().expect("Failed to run ffmpeg command");
     if !output.status.success() {
         let stderr = String::from_utf8_lossy(&output.stderr);
         tracing::error!("Failed to create video clip: {stderr}");
@@ -139,7 +142,7 @@ fn create_video_clip(dir: &str, slide: &Slide, cache: bool, config: &TTSConfig,
     }
 }
 
-fn create_video_clips(
+pub(crate) fn create_video_clips(
     dir: &str,
     slides: &Vec<Slide>,
     cache: bool,
@@ -159,7 +162,13 @@ fn create_video_clips(
     }
 }
 
-fn concat_video_clips(concat_list: &str, output_path: &str) {
+pub(crate) fn combine_video(dir: &str, slides: &Vec<Slide>, output: &str, audio_codec: &str) {
+    let output = Path::new(dir).join(output);
+    let output_path = output.to_str().unwrap();
+    let concat_list = Path::new(dir).join("concat_list.txt");
+    let concat_list = concat_list.to_str().unwrap();
+    write_concat_list(dir, concat_list, slides);
+
     let output = std::process::Command::new("ffmpeg")
         .arg("-y")
         .arg("-f")
@@ -168,65 +177,18 @@ fn concat_video_clips(concat_list: &str, output_path: &str) {
         .arg(concat_list)
         .arg("-c")
         .arg("copy")
-        .arg(output_path)
-        .output()
-        .expect("Failed to run ffmpeg command");
-    if !output.status.success() {
-        let stderr = String::from_utf8_lossy(&output.stderr);
-        tracing::error!("Failed to concat video clips: {stderr}");
-        std::process::exit(1);
-    } else {
-        tracing::info!("Concatenated video clips into {output_path}");
-    }
-}
-
-pub fn generate_video(
-    dir: &str,
-    slides: &Vec<Slide>,
-    cache: bool,
-    config: &TTSConfig,
-    output: &str,
-    audio_ext: &str,
-) {
-    create_video_clips(dir, slides, cache, config, audio_ext);
-    let output = Path::new(dir).join(output);
-    let output = output.to_str().unwrap();
-    let concat_list = Path::new(dir).join("concat_list.txt");
-    let concat_list = concat_list.to_str().unwrap();
-    write_concat_list(dir, concat_list, slides);
-    concat_video_clips(concat_list, output);
-}
-
-pub fn generate_release_video(dir: &str, input: &str, output: &str, audio_codec: &str) {
-    let input_path = Path::new(dir).join(input);
-    let output_path = Path::new(dir).join(output);
-    let output_path = output_path.to_str().unwrap();
-    let mut cmd = std::process::Command::new("ffmpeg");
-    let output = cmd
-        .arg("-y")
-        .arg("-i")
-        .arg(input_path)
-        .arg("-c:v")
-        .arg("libx264")
-        .arg("-crf")
-        .arg("23")
-        .arg("-preset")
-        .arg("fast")
-        .arg("-vf")
-        .arg(format!("scale=-1:{HEIGHT},format=yuv420p"))
         .arg("-c:a")
         .arg(audio_codec)
         .arg("-strict")
         .arg("experimental")
         .arg(output_path)
         .output()
         .expect("Failed to run ffmpeg command");
-
     if !output.status.success() {
         let stderr = String::from_utf8_lossy(&output.stderr);
-        tracing::error!("Failed to create release video: {stderr}");
+        tracing::error!("Failed to concat video clips: {stderr}");
         std::process::exit(1);
     } else {
-        tracing::info!("Created release video {}", output_path);
+        tracing::info!("Combined video clips into {output_path}");
     }
 }