add build_videos file

build_videos.py

@@ -0,0 +1,108 @@
+import database
+import youtube_short
+import transcribe_video
+import twitch_chat_vod
+
+DB = None
+
+def build_chat_video():
+    chats = DB.get_unuploaded_chats()
+    
+    for chat in chats:
+        (
+            id,
+            title,
+            created_at,
+            view_count,
+            duration,
+            url,
+            thumbnail_url,
+            game_id,
+            game_name,
+            stream_id,
+            creator_name,
+            clip_is,
+            downloaded,
+            uploaded_yt,
+            uploaded_yt_chats,
+            uploaded_yt_shorts,
+        ) = chat
+        
+        print("====================================================")
+        print(f"🚀 Transcribe: {title}")
+        print("====================================================")
+
+        twitch_chat_vod.combine_twitch_vod_and_chat(f"download/videos/{id}/{id}.mp4", "overlay")
+
+
+def build_transcribe():
+    unuploadeds = DB.get_unuploaded()
+
+    for unuploaded in unuploadeds:
+        (
+            id,
+            title,
+            created_at,
+            view_count,
+            duration,
+            url,
+            thumbnail_url,
+            game_id,
+            game_name,
+            stream_id,
+            creator_name,
+            clip_is,
+            downloaded,
+            uploaded_yt,
+            uploaded_yt_chats,
+            uploaded_yt_shorts,
+        ) = unuploaded
+
+        if clip_is:
+            target_dir = f"download/clips"
+        else:
+            target_dir = f"download/videos"
+
+        print("====================================================")
+        print(f"🚀 Transcribe: {title}")
+        print("====================================================")
+        
+        transcribe_video.transcribe_to_srt(f"{target_dir}/{id}/{id}.mp4")
+
+
+def build_shorts():
+    shorts = DB.get_unuploaded_shorts()
+
+    top_txt = "TeamPGP Live on Twitch...\n Every Friday and Sunday @7:30 ET.\n twitch.tv/teampgp"
+
+    for short in shorts:
+        (
+            id,
+            title,
+            created_at,
+            view_count,
+            duration,
+            url,
+            thumbnail_url,
+            game_id,
+            game_name,
+            stream_id,
+            creator_name,
+            clip_is,
+            downloaded,
+            uploaded_yt,
+            uploaded_yt_chats,
+            uploaded_yt_shorts,
+        ) = short
+
+        target_dir = f"download/clips/{id}/{id}.mp4"
+
+        print("====================================================")
+        print(f"🚀 Processing: Clip to Youtube Short {title}")
+        print("====================================================")
+
+        youtube_short.fit_to_9_16_letterbox(target_dir, top_txt, f"Clipped By: {creator_name}.", True)   
+
+if __name__ =="__main__":
+    DB = database.Database()
+    build_shorts()

database.py

@@ -34,7 +34,7 @@ class Database:
             duration TEXT NOT NULL,
             url TEXT NOT NULL,
             thumbnail_url TEXT NOT NULL,
-            game_id INTEGER NOT NULL,
+            game_id TEXT NOT NULL,
             game_name TEXT NOT NULL,
             stream_id TEXT NOT NULL,
             creator_name TEXT NOT NULL,

transcribe_video.py

@@ -8,15 +8,16 @@ from whisper.utils import get_writer
 
 model = whisper.load_model("base")
 
-def extract_audio(video_path: str, audio_temp_path: str):
+def extract_audio(video_path: str):
     # Create a temporary path for a sanitized copy of the video
     sanitized_video_path = video_path.replace(".mp4", "_clean.mp4")
+    audio_temp_path = video_path.replace(".mp4", ".wav")
     
     print("Sanitizing video metadata for MoviePy parser...")
     # -map_chapters -1 removes chapter layouts that break the parser.
     # -sn strips text/subtitle streams that crash MoviePy.
     # -c copy copies video and audio instantly without quality loss.
-    linux.run_command(f"ffmpeg -y -i {video_path} -map_chapters -1 -sn -c copy {sanitized_video_path}")
+    linux.run_command_ffmpeg(f"ffmpeg -y -i {video_path} -map_chapters -1 -sn -c copy {sanitized_video_path}")
     
     print("Extracting uncompressed WAV audio...")
     try:
@@ -34,15 +35,20 @@ def extract_audio(video_path: str, audio_temp_path: str):
             os.remove(sanitized_video_path)
 
 
-def transcribe_to_srt(audio_path: str, output_directory: str, output_filename: str):    
+def transcribe_to_srt(video_path: str):
+
+    extract_audio(video_path)
+
+    transcribe_path = video_path.replace(".mp4", ".str")
+    audio_path = video_path.replace(".mp4", ".wav")
     print("Transcribing audio...")
     result = model.transcribe(audio_path)
     
     print("Creating SRT file...")
-    srt_writer = get_writer("srt", output_directory)
+    srt_writer = get_writer("srt", transcribe_path)
-    srt_writer(result, output_filename, {})
+    srt_writer(result, transcribe_path, {})
     
-    print(f"SRT subtitle file saved in: {output_directory}")
+    print(f"SRT subtitle file saved in: {transcribe_path}")
     if os.path.exists(audio_path):
         os.remove(audio_path)