Update to separate outputs

app.py

@@ -89,42 +89,45 @@ def generate_soap(
         outputs.append(text)
         yield "".join(outputs)
 
-# Combine transcription and SOAP generation
-def transcribe_and_generate_soap(audio, task, system_prompt, max_new_tokens, temperature, top_p, top_k, repetition_penalty):
-    # First transcribe the audio
-    transcribed_text = transcribe(audio, task)
-    # Then generate SOAP notes based on the transcription and collect all generated text into a single string
-    generated_text = "".join(generate_soap(
-        transcribed_text,
-        system_prompt,
-        max_new_tokens,
-        temperature,
-        top_p,
-        top_k,
-        repetition_penalty
-    ))
-    return generated_text
-
 # Gradio Interface combining transcription and SOAP note generation
 demo = gr.Blocks(theme=gr.themes.Ocean())
 
 with demo:
     with gr.Tab("Clinical SOAP Note from Audio"):
-        audio_transcribe_and_soap = gr.Interface(
-            fn=transcribe_and_generate_soap,
+        # Transcription Interface
+        audio_input = gr.Audio(sources=["microphone", "upload"], type="filepath", label="Audio Input")
+        task_input = gr.Radio(["transcribe", "translate"], label="Task", value="transcribe")
+        transcription_output = gr.Textbox(label="Transcription Output")
+        
+        # Transcription button
+        transcribe_button = gr.Button("Transcribe")
+        transcribe_button.click(fn=transcribe, inputs=[audio_input, task_input], outputs=transcription_output)
+
+        # SOAP Generation Interface
+        transcribed_text_input = gr.Textbox(label="Edit Transcription before SOAP Generation", lines=5)
+        system_prompt_input = gr.Textbox(label="System Prompt", lines=2, value="You are a world class clinical assistant.")
+        max_new_tokens_input = gr.Slider(label="Max new tokens", minimum=1, maximum=2048, value=1024, step=1)
+        temperature_input = gr.Slider(label="Temperature", minimum=0.1, maximum=4.0, value=0.6, step=0.1)
+        top_p_input = gr.Slider(label="Top-p", minimum=0.05, maximum=1.0, value=0.9, step=0.05)
+        top_k_input = gr.Slider(label="Top-k", minimum=1, maximum=1000, value=50, step=1)
+        repetition_penalty_input = gr.Slider(label="Repetition penalty", minimum=1.0, maximum=2.0, value=1.2, step=0.05)
+
+        soap_output = gr.Textbox(label="Generated SOAP Note Output")
+
+        # SOAP generation button
+        generate_soap_button = gr.Button("Generate SOAP Note")
+        generate_soap_button.click(
+            fn=generate_soap,
             inputs=[
-                gr.Audio(sources=["microphone", "upload"], type="filepath", label="Audio Input"),
-                gr.Radio(["transcribe", "translate"], label="Task", value="transcribe"),
-                gr.Textbox(label="System Prompt", lines=2, value="You are a world class clinical assistant."),
-                gr.Slider(label="Max new tokens", minimum=1, maximum=2048, value=1024, step=1),
-                gr.Slider(label="Temperature", minimum=0.1, maximum=4.0, value=0.6, step=0.1),
-                gr.Slider(label="Top-p", minimum=0.05, maximum=1.0, value=0.9, step=0.05),
-                gr.Slider(label="Top-k", minimum=1, maximum=1000, value=50, step=1),
-                gr.Slider(label="Repetition penalty", minimum=1.0, maximum=2.0, value=1.2, step=0.05)
+                transcribed_text_input,
+                system_prompt_input,
+                max_new_tokens_input,
+                temperature_input,
+                top_p_input,
+                top_k_input,
+                repetition_penalty_input
             ],
-            outputs="text",
-            title="Generate Clinical SOAP Note from Audio",
-            description="Transcribe audio input and convert it into a structured clinical SOAP note."
+            outputs=soap_output
         )
 
 demo.queue().launch(ssr_mode=False)