Update app.py

app.py

@@ -1,12 +1,73 @@
+# import gradio as gr
+# from huggingface_hub import InferenceClient
+
+# """
+# For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
+# """
+# client = InferenceClient("Qwen/Qwen2.5-Coder-32B-Instruct")
+
+# def respond(message, history: list[tuple[str, str]]):
+#     system_message = "You are a friendly Chatbot. If the user query is product-related, provide structured product recommendations based on intent and relevance."
+#     max_tokens = 2048
+#     temperature = 0.7
+#     top_p = 0.95
+
+#     messages = [{"role": "system", "content": system_message}]
+
+#     for val in history:
+#         if val[0]:
+#             messages.append({"role": "user", "content": val[0]})
+#         if val[1]:
+#             messages.append({"role": "assistant", "content": val[1]})
+
+#     # Append product recommendation prompt if the user query is relevant
+#     product_prompt = ("Given a user's search query, recommend the most relevant products from the catalog. "
+#                       "Consider synonyms, user intent, and semantic meaning rather than just keyword matching. "
+#                       "If the query is vague, infer potential needs based on common shopping behavior. "
+#                       "Provide a ranked list of product recommendations with a short explanation for each suggestion. "
+#                       "Ensure the recommendations are diverse and cover multiple relevant categories if applicable. "
+#                       f"Now, based on the user query: '{message}', generate a well-structured product recommendation list.")
+    
+#     messages.append({"role": "user", "content": product_prompt})
+
+#     response = ""
+
+#     for message in client.chat_completion(
+#         messages,
+#         max_tokens=max_tokens,
+#         stream=True,
+#         temperature=temperature,
+#         top_p=top_p,
+#     ):
+#         token = message.choices[0].delta.content
+#         response += token
+#         yield response
+
+# """
+# For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
+# """
+# demo = gr.ChatInterface(respond)
+
+# if __name__ == "__main__":
+#     demo.launch()
 import gradio as gr
 from huggingface_hub import InferenceClient
+import tempfile
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("Qwen/Qwen2.5-Coder-32B-Instruct")
+# Initialize clients for both chat and speech-to-text
+chat_client = InferenceClient("Qwen/Qwen2.5-Coder-32B-Instruct")
+stt_client = InferenceClient("openai/whisper-large-v3")  # Using OpenAI's Whisper model for STT
+
+def transcribe_audio(audio_file):
+    """Convert audio to text using Whisper model"""
+    with open(audio_file, "rb") as f:
+        return stt_client.automatic_speech_recognition(f.read())
 
-def respond(message, history: list[tuple[str, str]]):
+def respond(message, history: list[tuple[str, str]], audio=None):
+    # If audio input is provided, transcribe it first
+    if audio:
+        message = transcribe_audio(audio)
+    
     system_message = "You are a friendly Chatbot. If the user query is product-related, provide structured product recommendations based on intent and relevance."
     max_tokens = 2048
     temperature = 0.7
@@ -20,33 +81,68 @@ def respond(message, history: list[tuple[str, str]]):
         if val[1]:
             messages.append({"role": "assistant", "content": val[1]})
 
-    # Append product recommendation prompt if the user query is relevant
-    product_prompt = ("Given a user's search query, recommend the most relevant products from the catalog. "
-                      "Consider synonyms, user intent, and semantic meaning rather than just keyword matching. "
-                      "If the query is vague, infer potential needs based on common shopping behavior. "
-                      "Provide a ranked list of product recommendations with a short explanation for each suggestion. "
-                      "Ensure the recommendations are diverse and cover multiple relevant categories if applicable. "
-                      f"Now, based on the user query: '{message}', generate a well-structured product recommendation list.")
-    
-    messages.append({"role": "user", "content": product_prompt})
+    product_prompt = ("Given a user's search query, recommend relevant products...")
+    messages.append({"role": "user", "content": f"{product_prompt}\n\nUser query: {message}"})
 
     response = ""
-
-    for message in client.chat_completion(
+    for chunk in chat_client.chat_completion(
         messages,
         max_tokens=max_tokens,
         stream=True,
         temperature=temperature,
         top_p=top_p,
     ):
-        token = message.choices[0].delta.content
+        token = chunk.choices[0].delta.content
         response += token
         yield response
 
+# Custom CSS for better visual appearance
+css = """
+.gradio-container {
+    background: linear-gradient(135deg, #f5f7fa 0%, #c3cfe2 100%);
+    min-height: 100vh;
+}
+.audio-input {
+    background: white !important;
+    border-radius: 10px !important;
+    padding: 20px !important;
+}
 """
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(respond)
+
+with gr.Blocks(css=css) as demo:
+    gr.Markdown("# Smart Product Assistant 🎤🛒")
+    with gr.Row():
+        with gr.Column(scale=2):
+            chatbot = gr.Chatbot(height=600)
+        with gr.Column(scale=1):
+            gr.Markdown("## Input Methods")
+            with gr.Tab("Text Input"):
+                text_input = gr.Textbox(placeholder="Type your query here...", label="Text Input")
+            with gr.Tab("Voice Input"):
+                audio_input = gr.Audio(
+                    sources="microphone",
+                    type="filepath",
+                    label="Speak your query",
+                    elem_classes="audio-input"
+                )
+            submit_btn = gr.Button("Submit", variant="primary")
+
+    def process_input(text, audio, history):
+        if audio:
+            return audio, history
+        elif text:
+            return text, history
+        return "", history
+
+    submit_btn.click(
+        process_input,
+        [text_input, audio_input, chatbot],
+        [text_input, chatbot]
+    ).then(
+        respond,
+        [text_input, chatbot, audio_input],
+        chatbot
+    )
 
 if __name__ == "__main__":
     demo.launch()