V0

Signed-off-by: Valentin De Matos <[email protected]>

Dockerfile

@@ -0,0 +1,13 @@
+FROM python:3.9.16
+
+WORKDIR /code
+
+COPY ./requirements.txt /code/requirements.txt
+
+RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+
+RUN cat /etc/ImageMagick-6/policy.xml | sed "s|[none]|[write|read]|g" > /etc/ImageMagick-6/policy.xml
+
+COPY . .
+
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,10 +1,12 @@
 ---
 title: Youtube Subtitler
-emoji: 🐠
-colorFrom: indigo
-colorTo: yellow
-sdk: docker
-pinned: false
+emoji: 🎥
+colorFrom: blue
+colorTo: red
+sdk: gradio
+sdk_version: 3.19.1
+app_file: app.py
+pinned: true
 license: unknown
 ---
 

app.py

@@ -0,0 +1,138 @@
+import os
+import requests
+import gradio as gr
+
+from time import time
+from pytube import YouTube
+from moviepy.editor import *
+from moviepy.video.tools.subtitles import SubtitlesClip
+
+
+GLADIA_API_KEY = os.environ.get("GLADIA_API_KEY")
+
+
+headers = {
+    "accept": "application/json",
+    "x-gladia-key": GLADIA_API_KEY,
+}
+
+ACCEPTED_LANGUAGE_BEHAVIOUR = [
+    "manual",
+    "automatic single language",
+    "automatic multiple languages",
+]
+
+
+def __download_youtube_video(youtube_url: str, path):
+
+    youtube_video = YouTube(youtube_url)
+
+    video_stream = youtube_video.streams.filter(res="720p").first()
+
+    if not video_stream:
+        video_stream = youtube_video.streams.filter(res="480p").first()
+
+    if not video_stream:
+        video_stream = youtube_video.streams.filter(res="360p").first()
+
+    video_stream.download(filename=path)
+
+
+def __apply_srt_to_video(path_to_srt: str, path_to_video: str, output_path: str):
+
+    def __generator(raw_txt):
+
+        MIN_LINE_SIZE = 80
+        MAX_LINE_SIZE = 140
+
+        text_to_display = []
+
+        while raw_txt:
+            idx = MIN_LINE_SIZE if MIN_LINE_SIZE < len(raw_txt) else len(raw_txt)
+
+            while idx <= len(raw_txt) and raw_txt[idx - 1] not in ('\n', ' ', ',', ';', '.', '-') and idx < MAX_LINE_SIZE:
+                idx += 1
+
+            text_to_display.append(raw_txt[:idx])
+            raw_txt = raw_txt[idx:]
+
+        text_to_display = "\n".join(text_to_display)
+
+        return TextClip(text_to_display, font='Arial', fontsize=30, color='white')
+
+    subs = SubtitlesClip(path_to_srt, __generator)
+    subtitles = SubtitlesClip(subs, __generator)
+
+    video = VideoFileClip(path_to_video)
+
+    # Combine video and subtitle clips
+    result = CompositeVideoClip([video, subtitles.set_pos(('center','bottom'))])
+
+    # Write the result to a new video file
+    result.write_videofile(output_path)
+
+
+def youtube_subtitler(youtube_url: str):
+    files_to_remove = []
+
+    try:
+        path_to_srt = f"/tmp/{str(time())}.srt"
+
+        files = {
+            'audio_url': (None, youtube_url),
+            'output_format': (None, "srt"),
+            'language_behaviour': (None, 'automatic multiple languages'),
+        }
+
+        response = requests.post(
+            'https://api.gladia.io/audio/text/audio-transcription/?output_format=srt',
+            headers=headers,
+            files=files
+        )
+
+        if response.status_code != 200:
+            raise RuntimeError(f"Failed to transcribe the video: {response.content}")
+
+        with open(path_to_srt, "w+") as f:
+            f.write(response.json()["prediction"])
+
+        files_to_remove.append(path_to_srt)
+
+        path_to_video = f"/tmp/{str(time())}.mp4"
+
+        __download_youtube_video(youtube_url=youtube_url, path=path_to_video)
+
+        files_to_remove.append(path_to_video)
+
+        output_path = f"/tmp/{str(time())}.mp4"
+
+        __apply_srt_to_video(
+            path_to_srt=path_to_srt,
+            path_to_video=path_to_video,
+            output_path=output_path
+        )
+
+    finally:
+        for file in files_to_remove:
+            os.remove(file)
+
+    return output_path
+
+
+iface = gr.Interface(
+    title="Gladia.io Youtube Subtitler",
+    description="""Burn subtitles into your youtube video using Gladia's Audio-Transcription solution.
+    <br/><br/>
+    You are more than welcome to join us on [Slack](https://gladia-io.slack.com)
+    and don't forget to get your own API key on [Gladia.io](https://gladia.io/) during the free alpha !
+    """,
+    fn=youtube_subtitler,
+    inputs=[
+        gr.Text(label="Youtube URL"),
+    ], outputs=[
+        gr.Video(format="mp4"),
+    ],
+)
+
+iface.queue()
+iface.launch()

requirements.txt

@@ -0,0 +1,2 @@
+pytube
+moviepy