Update pages/04_🔊_Upload_Audio_File.py

pages/04_🔊_Upload_Audio_File.py

@@ -2,6 +2,7 @@ import whisper
 import streamlit as st
 from streamlit_lottie import st_lottie
 from utils import write_vtt, write_srt
+from flores200_codes import flores_codes
 import ffmpeg
 import requests
 from typing import Iterator
@@ -36,11 +37,11 @@ with col1:
 
 with col2:
     st.write("""
-    ## Auto Transcriber
+    ## Sema Audio Transcriber
     ##### Input an audio file and get a transcript.
     ###### ➠ If you want to transcribe the audio in its original language, select the task as "Transcribe"
-    ###### ➠ If you want to translate the transcription to English, select the task as "Translate" 
-    ###### I recommend starting with the base model and then experimenting with the larger models, the small and medium models often work well. """)
+    ###### ➠ If you want to translate the subtitles to English, select the task as "Translate with Whisper" 
+    ###### ➠ If you want to translate the subtitles from English to any of the 200 supported languages, select the task as "Translate with Sema" """)
 
 loaded_model = whisper.load_model("base")
 current_size = "None"
@@ -92,6 +93,32 @@ def getSubs(segments: Iterator[dict], format: str, maxLineWidth: int) -> str:
     segmentStream.seek(0)
     return segmentStream.read()
 
+def translate(userinput, target_lang, source_lang=None):
+    if source_lang:
+       url = f"{Public_Url}/translate_enter/"
+       data = {
+           "userinput": userinput,
+           "source_lang": source_lang,
+           "target_lang": target_lang,
+        }
+       response = requests.post(url, json=data)
+       result = response.json()
+       print(type(result))
+       source_lange = source_lang
+       translation = result['translated_text']
+       
+    else:
+      url = f"{Public_Url}/translate_detect/"
+      data = {
+        "userinput": userinput,
+        "target_lang": target_lang,
+      }
+
+      response = requests.post(url, json=data)
+      result = response.json()
+      source_lange = result['source_language']
+      translation = result['translated_text']
+    return source_lange, translation
 
 def main():
     size = st.selectbox("Select Model Size (The larger the model, the more accurate the transcription will be, but it will take longer)", ["tiny", "base", "small", "medium", "large"], index=1)
@@ -103,7 +130,7 @@ def main():
         filename = input_file.name[:-4]
     else:
         filename = None
-    task = st.selectbox("Select Task", ["Transcribe", "Translate"], index=0)
+    task = st.selectbox("Select Task", ["Transcribe", "Translate with Whisper", "Translate with Sema"], index=0)
     if task == "Transcribe":
         if st.button("Transcribe"):
             results = inferecence(loaded_model, input_file, task)
@@ -150,7 +177,7 @@ def main():
             with col10:
                 st.info("Streamlit refreshes after the download button is clicked. The data is cached so you can download the transcript again without having to transcribe the video again.")
 
-    elif task == "Translate":
+    elif task == "Translate with Whisper":
         if st.button("Translate to English"):
             results = inferecence(loaded_model, input_file, task)
             col3, col4 = st.columns(2)
@@ -195,11 +222,17 @@ def main():
                 st.success("You can download the transcript in .srt format, edit it (if you need to) and upload it to YouTube to create subtitles for your video.")
             with col10:
                 st.info("Streamlit refreshes after the download button is clicked. The data is cached so you can download the transcript again without having to transcribe the video again.")
-
+    
+    elif task == "Translate with Sema":
+        default_language = "French"
+        target = st.selectbox("Select Language", list(flores_codes.keys()), index=list(flores_codes.keys()).index(default_language))
+        target_code = flores_codes[target]
+        
     else:
         st.error("Please select a task.")
 
 
 if __name__ == "__main__":
     main()
-    st.markdown("###### Made with :heart: by [@BatuhanYılmaz](https://github.com/BatuhanYilmaz26) [![this is an image link](https://i.imgur.com/thJhzOO.png)](https://www.buymeacoffee.com/batuhanylmz)")
+    st.markdown("###### ")
+    st.markdown("###### Powered by [sema © 2024](https://www.sema.wiki)")