Update app.py

app.py

@@ -402,13 +402,28 @@ def summarize_web_results(query: str, search_results: List[Dict[str, str]], conv
     except Exception as e:
         return f"An error occurred during summarization: {str(e)}"
 
+def process_input(message, history, audio, model, temperature, num_calls, use_web_search, selected_docs):
+    if audio is not None:
+        # If audio input is provided, transcribe it
+        try:
+            transcribed_text = transcribe(audio)
+            print(f"Transcribed audio: {transcribed_text}")
+            message = transcribed_text
+        except Exception as e:
+            print(f"Error transcribing audio: {str(e)}")
+            return f"Error transcribing audio: {str(e)}", history
+    
+    # Now call the existing respond function with the message (either text or transcribed audio)
+    return respond(message, history, model, temperature, num_calls, use_web_search, selected_docs)
+
 # Modify the existing respond function to handle both PDF and web search
-def respond(message, history, model, temperature, num_calls, use_web_search, selected_docs, audio_input):
-    if audio_input is not None:
+def respond(message, history, model, temperature, num_calls, use_web_search, selected_docs):
+    if audio is not None:
         # If audio input is provided, transcribe it
         try:
-            message = transcribe(audio_input)
-            print(f"Transcribed audio: {message}")
+            transcribed_text = transcribe(audio)
+            print(f"Transcribed audio: {transcribed_text}")
+            message = transcribed_text
         except Exception as e:
             print(f"Error transcribing audio: {str(e)}")
             return f"Error transcribing audio: {str(e)}", history
@@ -621,11 +636,12 @@ Write a detailed and complete response that answers the following user question:
         
         logging.info("Finished generating response")
 
-def transcribe(audio_path):
+def transcribe(audio):
+    if audio is None:
+        return ""
+    
     headers = {"Authorization": f"Bearer {huggingface_token}"}
-    with open(audio_path, "rb") as f:
-        data = f.read()
-    response = requests.post(WHISPER_API_URL, headers=headers, data=data)
+    response = requests.post(WHISPER_API_URL, headers=headers, data=audio)
     return response.json().get("text", "")
 
 def vote(data: gr.LikeData):
@@ -677,57 +693,60 @@ use_web_search = gr.Checkbox(label="Use Web Search", value=False)
 
 custom_placeholder = "Ask a question (Note: You can toggle between Web Search and PDF Chat in Additional Inputs below)"
 
-# Update the demo interface
-# Update the Gradio interface
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Dropdown(choices=MODELS, label="Select Model", value=MODELS[3]),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.2, step=0.1, label="Temperature"),
-        gr.Slider(minimum=1, maximum=5, value=1, step=1, label="Number of API Calls"),
-        gr.Checkbox(label="Use Web Search", value=True),
-        gr.CheckboxGroup(label="Select documents to query"),
-        gr.Audio(sources=["microphone"], streaming=True)
-    ],
-    title="AI-powered PDF Chat and Web Search Assistant",
-    description="Chat with your PDFs, use web search to answer questions, or speak your query.",
-    theme=gr.themes.Soft(
-        primary_hue="orange",
-        secondary_hue="amber",
-        neutral_hue="gray",
-        font=[gr.themes.GoogleFont("Exo"), "ui-sans-serif", "system-ui", "sans-serif"]
-    ).set(
-        body_background_fill_dark="#0c0505",
-        block_background_fill_dark="#0c0505",
-        block_border_width="1px",
-        block_title_background_fill_dark="#1b0f0f",
-        input_background_fill_dark="#140b0b",
-        button_secondary_background_fill_dark="#140b0b",
-        border_color_accent_dark="#1b0f0f",
-        border_color_primary_dark="#1b0f0f",
-        background_fill_secondary_dark="#0c0505",
-        color_accent_soft_dark="transparent",
-        code_background_fill_dark="#140b0b"
-    ),
-    css=css,
-    examples=[
-        ["Tell me about the contents of the uploaded PDFs."],
-        ["What are the main topics discussed in the documents?"],
-        ["Can you summarize the key points from the PDFs?"],
-        ["What's the latest news about artificial intelligence?"]
-    ],
-    cache_examples=False,
-    analytics_enabled=False,
-    textbox=gr.Textbox(placeholder="Ask a question about the uploaded PDFs or any topic", container=False, scale=7),
-    chatbot = gr.Chatbot(  
-        show_copy_button=True,
-        likeable=True,
-        layout="bubble",
-        height=400,
-        value=initial_conversation()
+# Create the Gradio interface
+with gr.Blocks() as demo:
+    gr.Markdown("# AI-powered PDF Chat and Web Search Assistant")
+    
+    audio_input = gr.Audio(sources=["microphone"], type="numpy", label="Speak your query")
+    
+    chat_interface = gr.ChatInterface(
+        fn=process_input,
+        additional_inputs=[
+            gr.Dropdown(choices=MODELS, label="Select Model", value=MODELS[3]),
+            gr.Slider(minimum=0.1, maximum=1.0, value=0.2, step=0.1, label="Temperature"),
+            gr.Slider(minimum=1, maximum=5, value=1, step=1, label="Number of API Calls"),
+            gr.Checkbox(label="Use Web Search", value=True),
+            gr.CheckboxGroup(label="Select documents to query"),
+            audio_input
+        ],
+        title="AI-powered PDF Chat and Web Search Assistant",
+        description="Chat with your PDFs, use web search to answer questions, or speak your query.",
+        theme=gr.themes.Soft(
+            primary_hue="orange",
+            secondary_hue="amber",
+            neutral_hue="gray",
+            font=[gr.themes.GoogleFont("Exo"), "ui-sans-serif", "system-ui", "sans-serif"]
+        ).set(
+            body_background_fill_dark="#0c0505",
+            block_background_fill_dark="#0c0505",
+            block_border_width="1px",
+            block_title_background_fill_dark="#1b0f0f",
+            input_background_fill_dark="#140b0b",
+            button_secondary_background_fill_dark="#140b0b",
+            border_color_accent_dark="#1b0f0f",
+            border_color_primary_dark="#1b0f0f",
+            background_fill_secondary_dark="#0c0505",
+            color_accent_soft_dark="transparent",
+            code_background_fill_dark="#140b0b"
+        ),
+        css=css,
+        examples=[
+            ["Tell me about the contents of the uploaded PDFs."],
+            ["What are the main topics discussed in the documents?"],
+            ["Can you summarize the key points from the PDFs?"],
+            ["What's the latest news about artificial intelligence?"]
+        ],
+        cache_examples=False,
+        analytics_enabled=False,
+        textbox=gr.Textbox(placeholder="Ask a question about the uploaded PDFs or any topic", container=False, scale=7),
+        chatbot = gr.Chatbot(  
+            show_copy_button=True,
+            likeable=True,
+            layout="bubble",
+            height=400,
+            value=initial_conversation()
+        )
     )
-)
-
 # Add file upload functionality
 # Add file upload functionality
 with demo: