from sentence_transformers import SentenceTransformer
from A_Preprocess import load_pdf_data
from E_Model_utils import get_embeddings
import numpy as np
import faiss

# Load and preprocess data
data_file_path = r'C:\Users\serban.tica\Documents\tobi_llm_intent_recognition\data\Pager_Intents_Cleaned.csv'
data = load_pdf_data(data_file_path)


sentences = data['utterance'].tolist()
model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
model_name = 'all-MiniLM-L6-v2'
embeddings = get_embeddings(model, sentences)
print(f'Embeddings shape: {embeddings.shape}.')

#save embeddings as faiss index

# Convert embeddings to float32
embeddings = np.array(embeddings).astype('float32')

# Create a FAISS index
index = faiss.IndexFlatL2(embeddings.shape[1])  # L2 distance
index.add(embeddings)

# Save the FAISS index
faiss.write_index(index, f"{model_name}_faiss.index")

# Load the FAISS index (for later use)
index = faiss.read_index(f"{model_name}_faiss.index")

# To query the index, you can use the search method
# Example: Find the 5 nearest neighbors of a query embedding

query_embedding = 'cat am de platit la factura'
query_embedding = np.array([embeddings[0]]).astype('float32')  # Example query
D, I = index.search(query_embedding, 5)  # D: distances, I: indices
print("Indices of nearest neighbors:", I)
print("Distances of nearest neighbors:", D)




#print(embeddings[:10])