trying romanian

app.py

@@ -1,7 +1,7 @@
 from transformers import pipeline
 import gradio as gr
 import time
-from video_downloader import download_video1, download_youtube_video
+from video_downloader import download_video, download_video1, download_youtube_video
 from moviepy.editor import AudioFileClip, VideoFileClip
 from moviepy.video.io.ffmpeg_tools import ffmpeg_extract_subclip
 import datetime
@@ -10,7 +10,7 @@ from pydub import AudioSegment
 from pydub.silence import split_on_silence
 import re
 
-pipe = pipeline("automatic-speech-recognition", model="Artanis1551/whisper_swedish")
+pipe = pipeline("automatic-speech-recognition", model="Artanis1551/whisper_romanian")
 
 
 def process_video1(date):
@@ -62,19 +62,78 @@ def process_video1(date):
     return video_path, transcription
 
 
+def process_video(date):
+    # If the date is not in YYYY-MM-DD format, return an error message
+    date_pattern = re.compile(r"\b\d{4}\d{2}\d{2}\b")
+    if not date_pattern.match(date):
+        video_path = download_youtube_video(
+            "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
+        )
+        transcription = "Please enter a date in the format YYYY-MM-DD."
+        return video_path, transcription
+    try:
+        # Download the video
+        video_path = download_video(date)
+
+        # Extract the first 30 seconds of the video
+        short_video_path = f"short_{date}.mp4"
+        ffmpeg_extract_subclip(video_path, 0, 30, targetname=short_video_path)
+
+        # Extract audio from the short video
+        audio_path = f"audio_{date}.wav"
+        AudioFileClip(short_video_path).write_audiofile(audio_path)
+
+        # Split the audio into chunks
+        audio = AudioSegment.from_wav(audio_path)
+        chunks = split_on_silence(audio, min_silence_len=500, silence_thresh=-40)
+
+        # Transcribe each chunk
+        transcription = ""
+        for i, chunk in enumerate(chunks):
+            chunk.export(f"chunk{i}.wav", format="wav")
+            with open(f"chunk{i}.wav", "rb") as audio_file:
+                audio = audio_file.read()
+            transcription += pipe(audio)["text"] + " "
+            os.remove(f"chunk{i}.wav")
+
+        # Remove the audio file
+        os.remove(audio_path)
+    except:
+        video_path = download_youtube_video(
+            "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
+        )
+        transcription = "No decision was made on this date."
+
+    return short_video_path, transcription
+
+
 iface = gr.Interface(
-    fn=process_video1,
-    inputs=[
-        gr.inputs.Textbox(label="Date with format YYYY-MM-DD"),
-    ],
+    fn=process_video,
+    inputs=gr.inputs.Textbox(label="Date with format YYYYMMDD"),
     outputs=[
         gr.outputs.Video(),
         gr.Textbox(lines=100, max_lines=100, interactive=True),
     ],
-    title="Transcribe Swedish Parliament Decisions",
-    description="This app transcribes the top Swedish Parliament decision"
-    + " video from the given date. Only the first 30 seconds of the "
+    title="Romanian Transcription Test",
+    description="This app transcribes videos from the Romanian Parliament"
+    + " on a given date. Only the first 30 seconds of the "
     + "video will be used if it is longer than that.",
 )
 
+
+# iface = gr.Interface(
+#     fn=process_video1,
+#     inputs=[
+#         gr.inputs.Textbox(label="Date with format YYYY-MM-DD"),
+#     ],
+#     outputs=[
+#         gr.outputs.Video(),
+#         gr.Textbox(lines=100, max_lines=100, interactive=True),
+#     ],
+#     title="Transcribe Swedish Parliament Decisions",
+# description="This app transcribes the top Swedish Parliament decision"
+# + " video from the given date. Only the first 30 seconds of the "
+# + "video will be used if it is longer than that.",
+# )
+
 iface.launch()

video_downloader.py

@@ -23,6 +23,26 @@ def get_response(url):
     return soup
 
 
+def download_video(date):
+    # Delete any existing .mp4 files
+    for mp4_file in glob.glob("*.mp4"):
+        os.remove(mp4_file)
+
+    year = date[:4]
+    url = f"https://www.cdep.ro/u02/comisii/{year}/cp46_{date}.mp4"
+    try:
+        urllib.request.urlretrieve(url, f"video_{date}.mp4")
+        print("Video downloaded successfully.")
+        return f"video_{date}.mp4"
+    except urllib.error.HTTPError as e:
+        if e.code == 404:
+            print("No video exists for the given date.")
+        else:
+            print(f"An error occurred while downloading the video: {e}")
+    except Exception as e:
+        print(f"An unexpected error occurred: {e}")
+
+
 def download_video1(date):
     # Get the webpage
     url = f"https://www.riksdagen.se/sv/sok/?avd=webbtv&from={date}&tom={date}&doktyp=kam-vo"
@@ -67,6 +87,3 @@ def download_youtube_video(url):
         return video_path
     except Exception as e:
         print(f"An error occurred while downloading the video: {e}")
-
-
-download_youtube_video("https://www.youtube.com/watch?v=dQw4w9WgXcQ")