Update app.py

app.py

@@ -23,9 +23,24 @@ def predict(question_choice, audio):
     # Generate the system message based on the chosen question
     strategy, explanation = HongWenData.strategy_text["TREES"]
     
+   def predict(question_choice, audio):
+    # Transcribe the audio using Whisper
+    with open(audio, "rb") as audio_file:
+        transcript = openai.Audio.transcribe("whisper-1", audio_file)
+    message = transcript["text"]  # This is the transcribed message from the audio input
+    
+    # Generate the system message based on the chosen question
+    strategy, explanation = HongWenData.strategy_text["TREES"]
+    
     # Reference to the picture description from HongWenData.py
     picture_description = HongWenData.description
 
+    # Determine whether to include the picture description based on the question choice
+    picture_description_inclusion = f"""
+    For the first question, ensure your feedback refers to the picture description provided:
+    {picture_description}
+    """ if question_choice == HongWenData.questions[0] else ""
+
     # Construct the conversation with the system and user's message
     conversation = [
         {
@@ -33,8 +48,7 @@ def predict(question_choice, audio):
             "content": f"""
             You are an expert English Language Teacher in a Singapore Primary school, directly guiding a Primary 6 student in Singapore. 
             The student is answering the question: '{question_choice}'. 
-            For the first question, ensure your feedback refers to the picture description provided:
-            {picture_description}
+            {picture_description_inclusion}
             Point out areas they did well and where they can improve, following the {strategy}. 
             Encourage the use of sophisticated vocabulary and expressions.
             For the second and third questions, the picture is not relevant, so the student should not refer to it in their response.
@@ -50,7 +64,7 @@ def predict(question_choice, audio):
         model='gpt-3.5-turbo',
         messages=conversation,
         temperature=0.6,
-        max_tokens=1000,  # Limiting the response to 500 tokens
+        max_tokens=1000,  # Limiting the response to 1000 tokens
         stream=True
     )