[Python] How can I improve the structure and readability of this Python RAG script? [closed]

Context: I'm a complete newbie trying to get good at structuring my code, I'm aware that I'm supposed to do my own research with docs, but I don't know what specific docs I should be reading for this specific use case.

Question: What kind of improvements can I make to this python code snippet?

**Context about code:**Uses .txt files for RAG and prints out response in terminal. LM Studio compatible.

import os
from dotenv import load_dotenv
from sentence_transformers import SentenceTransformer
import faiss
import numpy as np
import requests
import torch

# Envs
def load_config():
load_dotenv()
return {
"api_url": os.getenv("API_URL"),
"api_key": os.getenv("API_KEY"),
"model": os.getenv("LLM_MODEL", "deepseek-r1-distill-qwen-7b"),
"embedding_model": os.getenv("EMBEDDING_MODEL", "all-MiniLM-L6-v2"),
}

def chunk_text(text, chunk_size=500, overlap=50):
tokens = text.split()
return [
" ".join(tokens[i: i + chunk_size])
for i in range(0, len(tokens), chunk_size - overlap)
]

def build_embeddings(chunks, config):
device = 'cuda' if torch.cuda.is_available() else 'cpu'
embedder = SentenceTransformer(config["embedding_model"], device=device)
return embedder.encode(chunks, batch_size=32, show_progress_bar=True)

def build_faiss_index(embeddings):
dim = embeddings.shape[1]
quantizer = faiss.IndexFlatL2(dim)
index = faiss.IndexIVFFlat(quantizer, dim, min(50, len(embeddings)))
if len(embeddings) > 0:
index.train(np.array(embeddings).astype("float32"))
index.add(np.array(embeddings).astype("float32"))
return index

def save_index(index, filepath="index/index.faiss"):
os.makedirs(os.path.dirname(filepath), exist_ok=True)
faiss.write_index(index, filepath)

def load_index(dim, filepath="index/index.faiss"):
if os.path.exists(filepath):
return faiss.read_index(filepath, faiss.IO_FLAG_MMAP)
else:
index = faiss.IndexFlatL2(dim)
save_index(index, filepath)
return index


def retrieve_chunks(question, chunks, embedder, index, top_k =3):
q_emb = embedder.encode([question])
_, idxs = index.search(np.array(q_emb).astype("float32"), top_k)
return [chunks for i in idxs[0]]


def ask_with_context(question, config, chunks, embedder, index):
try:
snippets = retrieve_chunks(question, chunks, embedder, index)
prompt = (
"Use the following context to answer the question in a clear, concise, and friendly way.\n\n"
+ "\n\n".join(snippets)
+ f"\n\nQuestion: {question}\nAnswer:"
)
headers = {
"Authorization": f"Bearer {config['api_key']}",
"Content-Type": "application/json",
}
payload = {
"model" : config["model"],
"messages": [{"role": "user", "content": prompt}],
"max_tokens": 4096, # Max tokens
"temperature": 0.2, # Temperature
}
r = requests.post(config["api_url"], headers=headers, json=payload)
r.raise_for_status()
return r.json()["choices"][0]["message"]["content"]
except requests.exceptions.RequestException as e:
print(f"API Request Error: {e}")
return "Sorry, I couldn't get a response from the language model."
except Exception as e:
print(f"Error: {e}")
return "An error occurred while processing your Question."


if __name__ == "__main__":
cfg = load_config()
with open("data/message.txt", "r", encoding="utf-8") as f:
text = f.read()

chunks = chunk_text(text)
embeddings = build_embeddings(chunks, cfg)
index = build_faiss_index(embeddings)
embedder = SentenceTransformer(cfg["embedding_model"])

print("Ready! Ask your questions (type 'exit' to quit).")
while True:
q = input("Q: ")
if q.lower() in ("exit", "quit"):
break
ans = ask_with_context(q, cfg, chunks, embedder, index)
ans = ans.replace("<think>", "").replace("</think>", "").strip()
print("A:", ans)


**Question2:**Where can I learn how to get better at this? I'm a high school student and I don't have access to any paid courses yet, but if there are any book suggestions, that would be greatly appreciated.

Continue reading...