added final touches

.env

@@ -1 +1,2 @@
-MODEL = ada:ft-personal-2023-02-17-17-56-33
+MODEL = ada:ft-personal-2023-02-17-17-56-33
+OPENAI_API_KEY = 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

app.py

@@ -1,45 +1,11 @@
-import os
 import gradio as gr
-import numpy as np
 import librosa
-import torch
-from transformers import Speech2TextProcessor, Speech2TextForConditionalGeneration
-from dotenv import load_dotenv
 import openai
-from run_command_responses import ResponseManager as rs
+from constants import *
 
-resoponses = {
-    "heated_seats_on": rs.activate_heated_seats,
-    "heated_seats_off": rs.deactivate_heated_seats,
-    "cooled_seats_on": rs.activate_cooled_seats,
-    "cooled_seats_off": rs.deactivate_cooled_seats,
-    "massage_seats_on": rs.activate_massage_seats,
-    "massage_seats_off": rs.deactivate_massage_seats,
-}
-
-id2label = {
-    1: "massage_seats_on",
-    2: "massage_seats_off",
-    3: "heated_seats_on",
-    4: "heated_seats_off",
-    5: "cooled_seats_on",
-    6: "cooled_seats_off",
-}
-
-load_dotenv()
-
-os.environ["PATH"] += ".\env\Lib\site-packages\ffprobe"
-
-OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
-MODEL = os.getenv("MODEL")
 
 openai.api_key = OPENAI_API_KEY
 
-model = Speech2TextForConditionalGeneration.from_pretrained(
-    "facebook/s2t-small-librispeech-asr"
-)
-processor = Speech2TextProcessor.from_pretrained("facebook/s2t-small-librispeech-asr")
-
 
 def get_command(command, model, id2label):
     completion = openai.Completion.create(
@@ -50,7 +16,11 @@ def get_command(command, model, id2label):
     return result
 
 
-def transcribe(audio):
+def transcribe(audio, text):
+    if text:
+        result = get_command(text, MODEL, id2label)
+        return "Text provided by the user", text_respnses[result], None
+
     input, rate = librosa.load(
         audio, sr=16000
     )  # Downsample original frequency to 16000hrz
@@ -60,13 +30,23 @@ def transcribe(audio):
     )
     transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)
     result = get_command(transcription, MODEL, id2label)
-    resoponses.get(result)()
-    return result
+    audio_res = resoponses.get(result)()
+    return transcription, text_respnses[result], audio_res
 
 
 if __name__ == "__main__":
     gr.Interface(
         fn=transcribe,
-        inputs=gr.Audio(source="microphone", type="filepath"),
-        outputs="text",
+        inputs=[
+            gr.Audio(label="", source="microphone", type="filepath"),
+            gr.Textbox(label="If you prefer type your command (more accurate)"),
+        ],
+        outputs=[
+            gr.Textbox(
+                label="Input Transcription (Please check that this matches what you've said)"
+            ),
+            gr.Textbox(label="Machine Response (Text Version)"),
+            gr.Audio(label="Machine Response (Audio Version)"),
+        ],
+        allow_flagging="auto",
     ).launch()

commands.py

@@ -0,0 +1,18 @@
+class ResponseManager:
+    def activate_heated_seats(self):
+        return "assets\command_responses\heated_activated.wav"
+
+    def deactivate_heated_seats(self):
+        return "assets\command_responses\heated_deactivated.wav"
+
+    def activate_cooled_seats(self):
+        return "assets\command_responses\cooled_activated.wav"
+
+    def deactivate_cooled_seats(self):
+        return "assets\command_responses\cooled_deactivated.wav"
+
+    def activate_massage_seats(self):
+        return "assets\command_responses\massage_activated.wav"
+
+    def deactivate_massage_seats(self):
+        return "assets\command_responses\massage_deactivated.wav"

constants.py

@@ -0,0 +1,44 @@
+from commands import ResponseManager
+from dotenv import load_dotenv
+from transformers import Speech2TextProcessor, Speech2TextForConditionalGeneration
+import os
+
+load_dotenv()
+
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+MODEL = os.getenv("MODEL")
+
+rs = ResponseManager()
+
+resoponses = {
+    "heated_seats_on": rs.activate_heated_seats,
+    "heated_seats_off": rs.deactivate_heated_seats,
+    "cooled_seats_on": rs.activate_cooled_seats,
+    "cooled_seats_off": rs.deactivate_cooled_seats,
+    "massage_seats_on": rs.activate_massage_seats,
+    "massage_seats_off": rs.deactivate_massage_seats,
+}
+
+id2label = {
+    1: "massage_seats_on",
+    2: "massage_seats_off",
+    3: "heated_seats_on",
+    4: "heated_seats_off",
+    5: "cooled_seats_on",
+    6: "cooled_seats_off",
+}
+
+text_respnses = {
+    "massage_seats_on": "Massage seats activated",
+    "massage_seats_off": "Massage seats deactivated",
+    "heated_seats_on": "Heated seats activated",
+    "heated_seats_off": "Heated seats deactivated",
+    "cooled_seats_on": "Cooled seats activated",
+    "cooled_seats_off": "Cooled seats deactivated",
+}
+
+model = Speech2TextForConditionalGeneration.from_pretrained(
+    "facebook/s2t-small-librispeech-asr"
+)
+
+processor = Speech2TextProcessor.from_pretrained("facebook/s2t-small-librispeech-asr")

flagged/log.csv

@@ -0,0 +1,7 @@
+audio,output,flag,username,timestamp
+C:\Users\admin\Desktop\DS JRP\Programming\Python\Portfolio\Voice_Commands\flagged\audio\tmp7zujjp6r.wav,,,,2023-02-18 19:27:17.163146
+C:\Users\admin\Desktop\DS JRP\Programming\Python\Portfolio\Voice_Commands\flagged\audio\tmpxmty_reu.wav,,,,2023-02-18 19:27:19.327085
+C:\Users\admin\Desktop\DS JRP\Programming\Python\Portfolio\Voice_Commands\flagged\component 0\tmp4mx6cxo6.wav,,,,2023-02-19 19:21:00.276808
+C:\Users\admin\Desktop\DS JRP\Programming\Python\Portfolio\Voice_Commands\flagged\component 0\tmpj5vph7ui.wav,,,,2023-02-19 19:22:46.434333
+,my neck is cold,,,2023-02-19 19:28:56.486805
+,my neck  need no more heating,,,2023-02-19 19:29:16.234959

packages.txt

@@ -1 +0,0 @@
-python3-gi

run_command_responses.py

@@ -1,21 +0,0 @@
-from playsound import playsound
-
-
-class ResponseManager:
-    def activate_heated_seats():
-        playsound("assets\command_responses\heated_activated.wav")
-
-    def deactivate_heated_seats():
-        playsound("assets\command_responses\heated_deactivated.wav")
-
-    def activate_cooled_seats():
-        playsound("assets\command_responses\cooled_activated.wav")
-
-    def deactivate_cooled_seats():
-        playsound("assets\command_responses\cooled_deactivated.wav")
-
-    def activate_massage_seats():
-        playsound("assets\command_responses\massage_activated.wav")
-
-    def deactivate_massage_seats():
-        playsound("assets\command_responses\massage_deactivated.wav")

script.py

@@ -1,12 +0,0 @@
-import os
-from pydub import AudioSegment
-
-path = "command_responses"
-file_names = os.listdir(path)
-res_path = "assets\command_responses"
-
-for file_name in file_names:
-    m4a_file = file_name
-    wav_filename = os.path.splitext(file_name)[0] + ".wav"
-    track = AudioSegment.from_file(os.path.join(path, m4a_file), format="m4a")
-    file_handle = track.export(os.path.join(res_path, wav_filename), format="wav")