Basic audio input integration to chatbot

app.py

@@ -4,7 +4,7 @@ from time import sleep
 from typing import Dict, List, Generator
 
 import gradio as gr
-import openai
+from openai import OpenAI
 from dotenv import load_dotenv
 
 load_dotenv()
@@ -12,19 +12,40 @@ load_dotenv()
 class MockInterviewer:
 
     def __init__(self) -> None:
-        self._client = openai.OpenAI(api_key=os.environ['OPENAI_API_KEY'])
+        self._client = OpenAI(api_key=os.environ['OPENAI_API_KEY'])
         self._assistant_id_cache: Dict[str, str] = {}
         self.clear_thread()
 
-    def chat(self, usr_message: Dict, history: List[List], job_role: str, company: str) -> Generator:
-        print('Started chat')
-        self._validate_fields(job_role, company)
-        assistant_id = self._init_assistant(job_role, company)
-        yield self._send_message(usr_message.get('text'), assistant_id)
+    def interface_chat(self, message: Dict, history: List[List], job_role: str, company: str) -> Generator:
+        yield self._chat(message.get('text'), job_role, company)
 
     def clear_thread(self) -> None:
         print('Initializing new thread')
         self._thread = self._client.beta.threads.create()
+    
+    def transcript(self, audio: str, job_role: str, company: str) -> str:
+        with open(audio, 'rb') as audio_file:
+            transcriptions = self._client.audio.transcriptions.create(
+                model='whisper-1',
+                file=audio_file,
+            )
+        os.remove(audio)
+        response = self._chat(transcriptions.text, job_role, company)
+        return [(transcriptions.text, response)]
+    
+    def _chat(self, message: str, job_role: str, company: str) -> str:
+        print('Started chat')
+        self._validate_fields(job_role, company)
+        assistant_id = self._init_assistant(job_role, company)
+        return self._send_message(message, assistant_id)
+
+    def _validate_fields(self, job_role: str, company: str) -> None:
+        if not job_role and not company:
+            raise gr.Error('Job Role and Company are required fields.')
+        if not job_role:
+            raise gr.Error('Job Role is a required field.')
+        if not company:
+            raise gr.Error('Company is a required field.')
 
     def _send_message(self, message: str, assistant_id: str) -> str:
         self._client.beta.threads.messages.create(thread_id=self._thread.id, role='user', content=message)
@@ -47,14 +68,6 @@ class MockInterviewer:
         print(f'Assistant response: {response}')
         return response
 
-    def _validate_fields(self, job_role: str, company: str) -> None:
-        if not job_role and not company:
-            raise gr.Error('Job Role and Company are required fields.')
-        if not job_role:
-            raise gr.Error('Job Role is a required field.')
-        if not company:
-            raise gr.Error('Company is a required field.')
-
     def _create_files(self, company: str) -> List[str]:
         if company.lower() == 'amazon':
             url = 'https://www.aboutamazon.com/about-us/leadership-principles'
@@ -95,20 +108,6 @@ class MockInterviewer:
     def _create_cache_key(self, job_role: str, company: str) -> str:
         return f'{job_role.lower()}+{company.lower()}'
 
-    def transcript(audio):
-        try:
-            print(audio)
-            audio_file = open(audio, "rb")
-            transcriptions = openai.audio.transcriptions.create(
-                model="whisper-1",
-                file=audio_file,
-            )
-        except Exception as error:
-            print(str(error))
-            raise gr.Error("An error occurred while generating speech. Please check your API key and come back try again.")
-
-        return transcriptions.text
-
 # Creating the Gradio interface
 with gr.Blocks() as demo:
     mock_interviewer = MockInterviewer()
@@ -116,27 +115,24 @@ with gr.Blocks() as demo:
     with gr.Row():
         job_role = gr.Textbox(label='Job Role', placeholder='Product Manager')
         company = gr.Textbox(label='Company', placeholder='Amazon')
-        audio = gr.Audio(sources=["microphone"], type="filepath")
-
-    submit_btn = gr.Button("Submit")
-
-    response_output = gr.Textbox(label="Interviewer Response")
-    stt_output = gr.Textbox(label="Speech-To-Text Transcription")
-
 
     chat_interface = gr.ChatInterface(
-        fn=lambda usr_message, history, job_role, company: mock_interviewer.chat(usr_message, history, job_role, company),
+        fn=mock_interviewer.interface_chat,
         additional_inputs=[job_role, company],
         title='I am your AI mock interviewer',
         description='Make your selections above to configure me.',
         multimodal=True,
         retry_btn=None,
-        undo_btn=None
-    ).queue()
+        undo_btn=None).queue()
 
     chat_interface.load(mock_interviewer.clear_thread)
     chat_interface.clear_btn.click(mock_interviewer.clear_thread)
-    audio.stop_recording(fn=MockInterviewer.transcript, inputs=[audio], outputs=stt_output, api_name=False)
+
+    audio = gr.Audio(sources=['microphone'], type='filepath', editable=False)
+    audio.stop_recording(fn=mock_interviewer.transcript,
+                         inputs=[audio, job_role, company],
+                         outputs=[chat_interface.chatbot],
+                         api_name=False)
 
 if __name__ == '__main__':
     demo.launch().queue()