|
import ollama
|
|
import chromadb
|
|
import speech_recognition as sr
|
|
import requests
|
|
import pyttsx3
|
|
|
|
|
|
client = chromadb.Client()
|
|
|
|
message_history = [
|
|
|
|
{
|
|
'id' : 1,
|
|
'prompt' : 'What is your name?',
|
|
'response' : 'My name is TADBot, a bot to help with short term remedial help for mental purposes. '
|
|
|
|
},
|
|
|
|
{
|
|
'id' : 2,
|
|
'prompt' : 'Bye',
|
|
'response' : 'Good to see you get better. Hopefully you reach out to me if you have any problems.'
|
|
|
|
},
|
|
|
|
{
|
|
'id' : 3,
|
|
'prompt' : 'What is the essence of Life?',
|
|
'response' : 'The essence of life is to create what you want of yourself.'
|
|
|
|
}
|
|
]
|
|
convo = []
|
|
|
|
modelname = "ms"
|
|
|
|
|
|
def create_vector_db(conversations):
|
|
|
|
vector_db_name = 'conversations'
|
|
|
|
try:
|
|
client.delete_collection(vector_db_name)
|
|
except ValueError as e:
|
|
pass
|
|
|
|
vector_db = client.create_collection(name=vector_db_name)
|
|
for c in conversations:
|
|
|
|
serialized_convo = 'prompt: ' + c["prompt"] + ' response: ' + c["response"]
|
|
|
|
response = ollama.embeddings(model = "nomic-embed-text",prompt = serialized_convo)
|
|
embedding = response["embedding"]
|
|
|
|
vector_db.add(ids = [str(c['id'])], embeddings = [embedding], documents = [serialized_convo])
|
|
|
|
|
|
def stream_response(prompt):
|
|
|
|
convo.append({'role': "user", 'content': prompt})
|
|
output = ollama.chat(model = modelname, messages = convo)
|
|
response = output['message']['content']
|
|
|
|
print("TADBot: ")
|
|
print(response)
|
|
engine = pyttsx3.init()
|
|
engine.say(response)
|
|
engine.runAndWait()
|
|
|
|
convo.append({'role': "assistant", 'content': response})
|
|
|
|
def retrieve_embeddings(prompt):
|
|
response = ollama.embeddings(model = "nomic-embed-text", prompt = prompt)
|
|
propmt_embedding = response['embedding']
|
|
|
|
vector_db = client.get_collection(name = 'conversations')
|
|
results = vector_db.query(query_embeddings=[propmt_embedding], n_results = 1)
|
|
|
|
best_embedding = results['documents'][0][0]
|
|
return best_embedding
|
|
|
|
create_vector_db(message_history)
|
|
|
|
|
|
while True:
|
|
|
|
r = sr.Recognizer()
|
|
m = sr.Microphone()
|
|
|
|
try:
|
|
|
|
|
|
print("Say something!")
|
|
with m as source:
|
|
audio = r.listen(source)
|
|
|
|
|
|
try:
|
|
|
|
|
|
|
|
prompt = r.recognize_google(audio)
|
|
print("Tadbot thinks you said: " + prompt)
|
|
except sr.UnknownValueError:
|
|
print("Tadbot could not understand audio")
|
|
except sr.RequestError as e:
|
|
print("Could not request results from Google Speech Recognition service; {0}".format(e))
|
|
|
|
|
|
print("Please wait...")
|
|
with m as source:
|
|
r.adjust_for_ambient_noise(source)
|
|
|
|
if prompt == "bye" or prompt == "Bye":
|
|
print("TADBot: Hopefully I was able to help you out today. Have a Nice Day!")
|
|
break
|
|
"""
|
|
context = retrieve_embeddings(prompt)
|
|
|
|
prompt = prompt + "CONTEXT FROM EMBEDDING: " + context
|
|
"""
|
|
stream_response(prompt)
|
|
except KeyboardInterrupt:
|
|
pass
|
|
|
