implement machine translation; various fixes and refactorings HEAD main

1 files changed, 0 insertions, 118 deletions

diff --git a/src/subtitle_extraction/embedded.rs b/src/subtitle_extraction/embedded.rs
deleted file mode 100644
index 0ba6178..0000000
--- a/src/subtitle_extraction/embedded.rs
+++ /dev/null
@@ -1,118 +0,0 @@
-use std::sync::mpsc;
-
-use anyhow::Context;
-
-use crate::subtitle_extraction::*;
-
-pub fn extract_embedded_subtitles(
-    // stream index to use when storing extracted subtitles, this index already
-    // has to be in TRACKS when this function is called!
-    stream_ix: StreamIndex,
-    context: ffmpeg::codec::Context,
-    time_base: ffmpeg::Rational,
-    packet_rx: mpsc::Receiver<ffmpeg::Packet>,
-    sender: ComponentSender<SubtitleExtractor>,
-) -> anyhow::Result<()> {
-    let mut decoder = context
-        .decoder()
-        .subtitle()
-        .with_context(|| format!("error creating subtitle decoder for stream {}", stream_ix))?;
-
-    while let Ok(packet) = packet_rx.recv() {
-        let mut subtitle = ffmpeg::Subtitle::new();
-        match decoder.decode(&packet, &mut subtitle) {
-            Ok(true) => {
-                if let Some(cue) = parse_subtitle(&subtitle, &packet, time_base) {
-                    SUBTITLE_TRACKS
-                        .write()
-                        .get_mut(&stream_ix)
-                        .unwrap()
-                        .cues
-                        .push(cue.clone());
-                    sender
-                        .output(SubtitleExtractorOutput::NewCue(stream_ix, cue))
-                        .unwrap();
-                } else {
-                    log::error!("error parsing subtitle at pts {:?}", packet.pts())
-                }
-            }
-            Ok(false) => {
-                log::debug!("got empty (?) subtitle, not sure if this should ever happen");
-            }
-            Err(e) => {
-                log::error!("error decoding subtitle: {:?}", e)
-            }
-        }
-    }
-
-    Ok(())
-}
-
-fn parse_subtitle(
-    subtitle: &ffmpeg::Subtitle,
-    packet: &ffmpeg::Packet,
-    time_base: Rational,
-) -> Option<SubtitleCue> {
-    let pts_to_clock_time = |pts: i64| {
-        let nseconds: i64 =
-            (pts * time_base.numerator() as i64 * 1_000_000_000) / time_base.denominator() as i64;
-        gst::ClockTime::from_nseconds(nseconds as u64)
-    };
-
-    let text = subtitle
-        .rects()
-        .into_iter()
-        .map(|rect| match rect {
-            ffmpeg::subtitle::Rect::Text(text) => text.get().to_string(),
-            ffmpeg::subtitle::Rect::Ass(ass) => {
-                extract_dialogue_text(ass.get()).unwrap_or(String::new())
-            }
-            _ => String::new(),
-        })
-        .collect::<Vec<String>>()
-        .join("\n— ");
-
-    let start = pts_to_clock_time(packet.pts()?);
-    let end = pts_to_clock_time(packet.pts()? + packet.duration());
-
-    Some(SubtitleCue { start, end, text })
-}
-
-fn extract_dialogue_text(dialogue_line: &str) -> Option<String> {
-    // ASS dialogue format: ReadOrder,Layer,Style,Name,MarginL,MarginR,MarginV,Effect,Text
-    // we need the 9th field (Text), so split on comma but only take first 9 splits
-    // see also https://github.com/FFmpeg/FFmpeg/blob/a700f0f72d1f073e5adcfbb16f4633850b0ef51c/libavcodec/ass_split.c#L433
-    let text = dialogue_line.splitn(9, ',').last()?;
-
-    // remove ASS override codes (formatting tags) like {\b1}, {\i1}, {\c&Hffffff&}, etc.
-    let mut result = String::new();
-    let mut in_tag = false;
-    let mut char_iter = text.chars().peekable();
-
-    while let Some(c) = char_iter.next() {
-        if c == '{' && char_iter.peek() == Some(&'\\') {
-            in_tag = true;
-        } else if c == '}' {
-            in_tag = false;
-        } else if !in_tag {
-            // process line breaks and hard spaces
-            if c == '\\' {
-                match char_iter.peek() {
-                    Some(&'N') => {
-                        char_iter.next();
-                        result.push('\n');
-                    }
-                    Some(&'n') | Some(&'h') => {
-                        char_iter.next();
-                        result.push(' ');
-                    }
-                    _ => result.push(c),
-                }
-            } else {
-                result.push(c);
-            }
-        }
-    }
-
-    Some(result)
-}