Add Audio Input for Performance and Self-Reviews

[archie-ai-code-explain-pr-review[bot]]

• Add the ability for users to provide audio input for their performance reviews and self-reviews
• Previously, users had to type their input, but now they can record their audio
• Use the streamlit-audio-recorder library to enable audio recording in the app
• Implement the convert_speech_to_text function using the Whisper model on the Groq library to convert the audio to text
• Update the ReviewRequest and SelfReviewRequest models to include an optional audio_review field, and use this input for generating the review and self-review

[greptile-apps[bot]]

Implementation Steps

1. Install streamlit-audio-recorder:

   pip install streamlit-audio-recorder

2. Update app.py:

   import streamlit as st
   from streamlit_audio_recorder import audio_recorder
   from llm import convert_speech_to_text
   
   # Add audio recording option in the UI
   audio_data = audio_recorder()
   
   if audio_data:
      audio_text = convert_speech_to_text(audio_data)
      st.text_area('Converted Audio Text', audio_text, height=200)

3. Update ReviewRequest and SelfReviewRequest models:

   class ReviewRequest(BaseModel):
      your_role: str
      candidate_role: str
      perf_question: Optional[str] = None
      your_review: str
      audio_review: Optional[bytes] = None  # Add this line
      llm_type: str
      user_api_key: str
      model_size: str = "small"
   
   class SelfReviewRequest(BaseModel):
      text_dump: str
      questions: List[str]
      instructions: Optional[str] = None
      audio_review: Optional[bytes] = None  # Add this line
      llm_type: str
      user_api_key: str
      model_size: str = "medium"

4. Implement convert_speech_to_text in llm.py:

   from groq import Whisper
   
   def convert_speech_to_text(audio_data: bytes) -> str:
      whisper = Whisper(api_key=GROQ_API_KEY)
      response = whisper.transcribe(audio_data)
      return response['text']

5. Update review generation logic in app.py:

   if st.button('Generate Performance Review'):
      if not user_api_key:
          st.error("Please enter your API key in the sidebar.")
      elif not your_role or not candidate_role or not your_review:
          st.error("Please fill in all required fields.")
      else:
          try:
              questions = perf_question.split('\n') if perf_question else DEFAULT_QUESTIONS.split('\n')
              review_request = ReviewRequest(
                  your_role=your_role,
                  candidate_role=candidate_role,
                  perf_question="\n".join(questions),
                  your_review=your_review,
                  audio_review=audio_data,  # Add this line
                  llm_type=llm_type,
                  user_api_key=user_api_key,
                  model_size=model_size
              )
              review = generate_review(**review_request.model_dump())
              for qa in review:
                  st.markdown(f"**{qa['question']}**")
                  st.markdown(qa['answer'])
                  st.markdown("---")
          except Exception as e:
              st.error(f"An error occurred: {str(e)}")

References

/app.py /self_review.py /review.py /llm.py

#### About Greptile

This response provides a starting point for your research, not a precise solution. Help us improve! Please leave a 👍 if this is helpful and 👎 if it is irrelevant. [Ask Greptile](https://app.greptile.com/chat/github/ajitesh123/perf-review-ai/main) · [Edit Issue Bot Settings](https://app.greptile.com/apps/github)