Auto save

Author: 169

Diff for: config.py

@@ -2,8 +2,8 @@
 
 DEBUG = False
 HERE = Path(__file__).parent.absolute()
-DOWNLOAD_DIR = HERE / ".download"
-OUTPUT_DIR = HERE / "output"
+DOWNLOAD_DIR: str | Path = HERE / ".download"
+OUTPUT_DIR: str | Path = HERE / "output"
 WHISPER_MODEL = "small"
 FFMPEG_BIN = "ffmpeg"
 FFMPEG_PREFIX_OPTS = "-hide_banner -loglevel error -y"

Diff for: core/utils.py

@@ -24,21 +24,23 @@ def format_timestamp(
     return f"{minutes:02d}:{seconds:02d}.{milliseconds:03d}"
 
 
-def write_vtt(transcript: list[dict|Segment], audio: str, language: str | None = None) -> str:
+def write_vtt(
+    transcript: list[dict | Segment], audio: str, language: str | None = None
+) -> str:
     if language is not None:
         vtt_filename = f"{audio.removesuffix('.mp3')}.{language}.vtt"
     else:
         vtt_filename = f"{audio.removesuffix('.mp3')}.vtt"
     return _write_vtt(transcript, vtt_filename)
 
 
-def _write_vtt(transcript: list[dict|Segment], vtt_filename: str) -> str:
-
+def _write_vtt(transcript: list[dict | Segment], vtt_filename: str) -> str:
     with open(vtt_filename, "w") as f:
         f.write("WEBVTT\n")
         for segment in transcript:
             if isinstance(segment, Segment):
                 segment = segment.model_dump()
+            print(segment, "segment")
             start = format_timestamp(segment["start"])
             end = format_timestamp(segment["end"])
             f.write(f"\n{start} --> {end}\n{segment['text'].strip()}\n")
@@ -100,8 +102,15 @@ def parse_vtt(text: str) -> list[Segment]:
             timestamp, text, _ = list(items)
         except ValueError:
             timestamp, text = list(items)
-        start, end = timestamp.split('-->')
-        texts.append(Segment(timestamp=timestamp, start=get_seconds(start), end=get_seconds(end), text=text))
+        start, end = timestamp.split("-->")
+        texts.append(
+            Segment(
+                timestamp=timestamp,
+                start=get_seconds(start),
+                end=get_seconds(end),
+                text=text,
+            )
+        )
     return texts
 
 

Diff for: webui.py

@@ -1,3 +1,5 @@
+# TODO:
+# 1. auto save
 import mimetypes
 import os.path
 
@@ -19,13 +21,14 @@
 )
 from core.audio import generate_audio, generate_subtitle
 from core.downloader import download
+from core.utils import _write_vtt, parse_vtt
 from core.video import generate_video
-from core.utils import parse_vtt, _write_vtt
 
 if "subtitle_path" not in st.session_state:
     st.session_state.update(
         {
             "segments": [],
+            "replaced_segments": [],
             "current_seg_index": 0,
             "current_time": 0,
             "vtt_content": "",
@@ -46,6 +49,7 @@ def wrapper(label, *args, **kwargs):
         widget_value = f(*args, **kwargs)
         st.session_state.widget_values[label] = widget_value
         return widget_value
+
     return wrapper
 
 
@@ -80,8 +84,6 @@ def wrapper(label, *args, **kwargs):
 
     st.write("## Subtitle")
 
-    subtitle_auto_save = st.checkbox('Audo Save')
-
     language: str = (
         st.selectbox(
             "Language",
@@ -109,12 +111,14 @@ def wrapper(label, *args, **kwargs):
 
 video_path = st.text_input("Video Path or URL")
 subtitle_path = st.text_input("VTT Path or URL", value=st.session_state.subtitle_path)
-if subtitle_path:
+
+if subtitle_path and not st.session_state.segments:
     with open(subtitle_path) as f:
         content = f.read()
         st.session_state.vtt_content = content
         if not st.session_state.segments:
             st.session_state.segments = parse_vtt(content)
+            st.session_state.replaced_segments = parse_vtt(content)
 
 
 def get_current_vtt_content() -> str:
@@ -123,11 +127,12 @@ def get_current_vtt_content() -> str:
         return ""
     if not values["video_component"]:
         return ""
-    current_time = values["video_component"]['current_time']
+    current_time = values["video_component"]["current_time"]
     for index, seg in enumerate(st.session_state.segments):
         if seg.start <= current_time <= seg.end:
-            st.session_state['current_seg_index'] = index
+            st.session_state["current_seg_index"] = index
             return seg.text
+    return ""
 
 
 def subtitle_callback(path: str) -> None:
@@ -152,8 +157,8 @@ def save_callback() -> None:
     if not subtitle_path:
         st.error("Subtitle is required.")
     else:
-        _write_vtt(st.session_state.segments, subtitle_path)
-        st.success(f"Subtitle saved")
+        _write_vtt(st.session_state.replaced_segments, subtitle_path)
+        st.session_state.segments = st.session_state.replaced_segments
 
 
 def preview_callback() -> None:
@@ -170,6 +175,7 @@ def preview_callback() -> None:
             current_time="",
         )
 
+
 def generate_callback() -> None:
     if not video_path:
         st.error("Video Path or URL is required.")
@@ -216,7 +222,4 @@ def generate_callback() -> None:
 
 current_vtt = st.text_input("Subtitle", value=get_current_vtt_content())
 if st.session_state.segments and current_vtt:
-    st.session_state.segments[st.session_state.current_seg_index].text = current_vtt
-
-if subtitle_auto_save and subtitle_path and st.session_state.segments:
-    _write_vtt(st.session_state.segments, subtitle_path)
+    st.session_state.replaced_segments[st.session_state.current_seg_index].text = current_vtt