Merge pull request 'Updates' (#21) from React-Group/interstellar_ai:main into main

Reviewed-on: https://interstellardevelopment.org/code/code/sageTheDm/interstellar_ai/pulls/21

py/api.py

@@ -99,16 +99,9 @@ class API:
 
         @self.app.route('/interstellar_ai/api/voice_recognition', methods=['POST'])
         def voice_recognition():
-            print(request.args)
-            recog_type = request.form.get('type')
-            print(recog_type)
             audio = request.files.get('audio')
-            option = request.form.get('option')
-            if recog_type == "basic":
-                text = self.voice.basic_recognition(audio, option)
-                return jsonify({'status': 200, 'response': text})
-            else:
-                return jsonify({'status': 401, 'response': "Invalid type"})
+            text = self.voice.recognition(audio)
+            return jsonify({'status': 200, 'response': text})
 
         @self.app.route('/interstellar_ai/api/weather', methods=['POST'])
         def get_weather():

py/requirements.txt

@@ -13,4 +13,5 @@ python-weather
 pycouchdb
 pyttsx3
 pip-licenses
-openai-whisper
+openai-whisper
+pydub

py/voice.py

@@ -1,24 +1,10 @@
-import speech_recognition as sr
+import whisper
 
 
 class VoiceRecognition:
     @staticmethod
-    def basic_recognition(audio, option):
-        print(type(audio))
-        print("preparing")
-        r = sr.Recognizer()
-        if option == "online":
-            with sr.AudioFile(audio) as source:
-                print("online")
-                text = r.recognize_google_cloud(source)
-                print("recognized as: " + text)
-                return text
-        elif option == "offline":
-            with sr.AudioFile(audio) as source:
-                print("offline")
-                text = r.recognize_sphinx(source)
-                print("recognized as: " + text)
-                return text
-
-        print("nothing")
-        return False
+    def recognition(audio):
+        model = whisper.load_model("base")
+        result = model.transcribe(audio)
+        print(result["text"])
+        return result["text"]