Merge pull request #6 from JSchmie/gradio_app

Gradio_app

autotranscript/__init__.py

@@ -1,10 +1,13 @@
 from .autotranscript import *
-from .app.qtfaststart import *
 from .transcriber import *
 from .audio import *
 from .transcript_exporter import *
 from .diarisation import *
+
 from .version import get_version as _get_version
 from .misc import *
 
+from .app.gradio_app import *
+from .app.qtfaststart import *
+ 
 __version__ = _get_version()

autotranscript/app/__init__.py

@@ -1 +1,2 @@
 from .qtfaststart import *
+from .gradio_app import *

autotranscript/app/gradio_app.py

@@ -1,3 +1,20 @@
+"""
+Gradio Audio Transcription App.
+--------------------------------
+
+This module provides an interface to transcribe audio files using the 
+AutoTranscribe model. Users can either upload an audio file or record their speech 
+live for transcription. The application supports multiple languages and provides 
+options to specify the number of speakers and the language of the audio.
+
+Attributes:
+    LANGUAGES (list): A list of supported languages for transcription.
+
+Usage:
+    Run this script to start the Gradio web interface for audio transcription.
+    
+"""
+
 from autotranscript import AutoTranscribe
 import gradio as gr
 
@@ -18,14 +35,33 @@ LANGUAGES = [
 
 
 def gradio_server(model : AutoTranscribe):
+    """
+    Sets up and launches the Gradio interface for audio transcription.
 
+    Args:
+        model (AutoTranscribe): An instance of the AutoTranscribe model for transcription.
+    """
     def transcribe(audio, microphone, number_of_speakers, language):
+        """
+        Transcribes the provided audio input based on the given parameters.
+
+        Args:
+            audio (str): Filepath to the uploaded audio file.
+            microphone (str): Filepath to the recorded audio.
+            number_of_speakers (int): Number of speakers in the audio.
+            language (str): Language of the audio content.
+
+        Returns:
+            tuple: Transcribed text (str), JSON output (dict)
+        """
         kwargs = {}
         if number_of_speakers != 0:
             kwargs["num_speakers"] = number_of_speakers
         if language != "None":
             kwargs["language"] = language
             
+        print()
+        
         if audio is not None:
             out = model.transcribe(audio, **kwargs)
         elif microphone is not None:
@@ -33,30 +69,31 @@ def gradio_server(model : AutoTranscribe):
         else:
             out = "Please upload an audio file or record one."
         
-        
+        return str(out), out.get_json(), out.get_md()
-        return str(out)
 
     gr.Interface(
         fn=transcribe, 
         inputs=[
-            gr.Audio(source= "upload", type="filepath", label="Upload Your Audio File", interactive=True),
+            gr.Audio(source= "upload", type="filepath", label="Upload Your Audio File",
-            gr.Audio(source= "microphone", type="filepath", label="Record Your Audio", interactive=True),
+                     interactive=True),
-            gr.Number(value=0, label= "Number of speakers", 
+            gr.Audio(source= "microphone", type="filepath", label="Record Your Audio",
+                     interactive=True, container= False),
+            gr.Number(value=0, label= "Number of speakers (optional)", 
                       info = "Number of speakers in the audio file. If you don't know, leave it at 0."), 
-            # gr.Number(value=0, label= "Minimal number of speakers", 
-            #           info = "Minimal number of speakers in the audio file. If you don't know or you have specified Numspeakers, leave it at 0."),
             gr.Dropdown(LANGUAGES,
-                        label="Languages", default="None",
+                        label="Language (optional)", value = "None",
                         info="Language of the audio file. If you don't know, leave it at None.")
         ],
         outputs=[
-            "text"
+            gr.Textbox(label="Transcription"),
+            gr.JSON(label="Raw Output", container= False),
         ],
         title="Audio Transcription",
-        thumbnail = "Logo_KIDA.png",
         description="Upload an audio file to transcribe its content. Powered by AutoTranscribe!",
         theme="soft",       # Example of a more modern theme
-    ).launch(share=True)
+        server_port=7860,
+        server_name="autotranscribe",   
+    ).queue().launch() 
     
     
 if __name__ == "__main__":

autotranscript/autotranscript.py

@@ -353,7 +353,7 @@ def cli():
         pass
     
     if start_server:
-        from .gradio_app import gradio_app
+        from .app.gradio_app import gradio_app
         gradio_app(model)
 
 if __name__ == "__main__":

autotranscript/transcript_exporter.py

@@ -99,7 +99,7 @@ class Transcript:
             sseg = time.strftime("%H:%M:%S",time.gmtime(segm[0]))
             eseg = time.strftime("%H:%M:%S",time.gmtime(segm[1]))
             
-            fstring += f"{speaker} ({sseg} ; {eseg}): {seq['text']}\n"
+            fstring += f"{speaker} ({sseg} ; {eseg}):\t{seq['text']}\n"
         return fstring
     
     def __repr__(self) -> str:
@@ -128,7 +128,7 @@ class Transcript:
         :rtype: str
         """
         if "indent" not in kwargs:
-            kwargs["indent"] = 4
+            kwargs["indent"] = 3
         return json.dumps(self.transcript, *args, **kwargs)
     
     def get_html(self) -> str: