moved variables to env

added example env

main.py

@@ -6,14 +6,14 @@ asyncio.set_event_loop(loop)
 
 
 sp = speak.Speak()
-graph = agent.Agent("openai")
+graph = agent.Agent()
 
 
 while True:
+    print("Listening...")
     text = sp.listen()
     if text and "max" in text.lower():
         response = loop.run_until_complete(graph.invoke_agent(text))
         if response:
             sp.glitch_stream_output(response)
     
-    print("Listening again...")

modules/adapter.py

@@ -5,8 +5,8 @@ env = environ.Env()
 environ.Env.read_env()
 
 class Adapter:
-    def __init__(self, llm_type):
-        self.llm_text = llm_type
+    def __init__(self):
+        self.llm_text = env("LLM_TYPE")
         if self.llm_text.lower() == "openai":
             from langchain_openai import OpenAIEmbeddings, OpenAI
             from langchain_openai import ChatOpenAI
@@ -15,16 +15,15 @@ class Adapter:
                 "answer the following request: {topic}"
             )
             self.llm_chat = ChatOpenAI(
-                temperature=0.3, model="gpt-4o-mini", openai_api_key=env("OPENAI_API_KEY")
+                temperature=0.3, model=env("OPENAI_MODEL"), openai_api_key=env("OPENAI_API_KEY")
             )
             self.embedding = OpenAIEmbeddings(model="text-embedding-ada-002")
         elif self.llm_text.lower() == "local":
             from langchain_community.llms import Ollama
             from langchain_community.embeddings import HuggingFaceBgeEmbeddings
             from langchain_community.chat_models import ChatOllama
-            llm_model = "llama3"
-            # llm_model = "notus"
-            self.llm = Ollama(base_url="http://10.0.0.231:11434", model=llm_model)
+            llm_model = env("OLLAMA_MODEL")
+            self.llm = Ollama(base_url=env("OLLAMA_URL"), model=llm_model)
             self.prompt = ChatPromptTemplate.from_template(
                 "answer the following request: {topic}"
             )

modules/agent.py

@@ -1,5 +1,4 @@
 from typing import TypedDict, Annotated, List, Union
-import json
 import operator
 from modules import adapter, spotify
 from langchain_core.agents import AgentAction, AgentFinish
@@ -13,8 +12,8 @@ import asyncio
 
 
 class Agent:
-    def __init__(self, model):
-        self.ad = adapter.Adapter(model)
+    def __init__(self):
+        self.ad = adapter.Adapter()
         self.sp = spotify.Spotify()
         self.llm = self.ad.llm_chat
         # self.final_answer_llm = self.llm.bind_tools(

modules/example env

@@ -0,0 +1,15 @@
+OPENAI_API_KEY=''
+#LLM_TYPE will take openai, local, or hybrid
+LLM_TYPE='openai' 
+OPENAI_MODEL='gpt-4o-mini'
+#OLLAMA_MODEL will take any model you can load in ollama
+OLLAMA_MODEL='llama' 
+OLLAMA_URL='http://10.0.0.231:11434' 
+spotify_client_id = ''
+spotify_client_secret = ''
+spotify_redirect_uri = 'http://localhost:8888/callback'
+
+
+
+
+

modules/speak.py

@@ -56,20 +56,21 @@ class Speak:
       
     def listen(self):
         with self.microphone as source:
-            # Adjust for ambient noise
+            #! Adjust for ambient noise
             self.recognizer.adjust_for_ambient_noise(source, duration=1)
-            
-            print("Listening...")
-            audio = self.recognizer.listen(source)
-            
+            #! added 5 second timeout so ambient noise detection can compensate for music that started playing
+            audio = self.recognizer.listen(source, timeout=5)
             try:
+                
                 text = self.recognizer.recognize_google(audio)
                 print("You said: ", text)
                 return text
-            except sr.UnknownValueError:
-                print("Sorry, I didn't get that.")
-            except sr.RequestError as e:
-                print("Sorry, I couldn't request results; {0}".format(e)) 
+            except:
+                pass
+            # except sr.UnknownValueError:
+            #     print("Sorry, I didn't get that.")
+            # except sr.RequestError as e:
+            #     print("Sorry, I couldn't request results; {0}".format(e)) 
     
     def stream_output(self, text):
         import urllib.parse
@@ -152,7 +153,7 @@ class Speak:
 
             # Randomly adjust pitch
             # octaves = random.uniform(-0.5, 0.5)
-            octaves = random.uniform(-1, 1)
+            octaves = random.uniform(-0.5, 1)
             modified_chunk = change_pitch(audio_segment, octaves)
 
             if stream is None:
@@ -162,7 +163,7 @@ class Speak:
                                 rate=modified_chunk.frame_rate,
                                 output=True)
             
-            if random.random() < 0.01:  # 1% chance to trigger stutter
+            if random.random() < 0.001:  # 1% chance to trigger stutter
                 repeat_times = random.randint(2, 5)  # Repeat 2 to 5 times
                 for _ in range(repeat_times):
                     stream.write(modified_chunk.raw_data)