File size: 1,673 Bytes
eb66dcb ba1cae1 e95de35 eb66dcb 5916e02 f631f81 eb66dcb 7fa286a eb66dcb 7fa286a eb66dcb 7fa286a f2e2bb3 ba1cae1 7fa286a ba1cae1 7fa286a eb66dcb e95de35 ef93b5e e95de35 eb66dcb e95de35 eb66dcb ef93b5e eb66dcb |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
from utils.convert_embedding import GetEmbedding
import random
import pickle
import os
from utils.rag import RAG
from faster_whisper import WhisperModel
def process(user_query:str):
# dump_user_question(user_query)
user_embedding = GetEmbedding([user_query]).user_query_emb()
with open(r"all_mix_embedding.pkl","rb") as f:
load_embedding = pickle.load(f)
with open(r"all_answers.pkl","rb") as f:
ans = pickle.load(f)
similarity_scores = cosine_similarity(user_embedding, load_embedding)
index = np.argmax(similarity_scores)
answer = ans[index]
score = similarity_scores[0,index]
print(f"Index : {index}:\tscore:{score} \tquery: {user_query}")
if float(score) > 0.60 :
final_output = random.choice(answer)
else:
final_output = RAG().pipeline(query=user_query)
return final_output
def audio_process(audio):
try:
model = WhisperModel(model_size_or_path="medium.en")
segments, info = model.transcribe(audio)
transcription = " ".join([seg.text for seg in segments])
result = process(user_query=transcription)
return result
except Exception as e:
print("Error:", e)
return str(e)
if __name__ == "__main__":
res = audio_process(r"C:\Users\lalit\Documents\Sound recordings\who_is_lalit.m4a")
print(res)
# for _ in range(3):
# user = input("How can i help you :? \n")
# result = process(user)
# print(result)
# with open(r"data\question_data.pkl","rb") as f:
# que = pickle.load(f)
# print(que) |