Spaces:
Sleeping
Sleeping
File size: 3,338 Bytes
47e4aa2 a45dfb0 facd13e 47e4aa2 facd13e 47e4aa2 a45dfb0 47e4aa2 a45dfb0 47e4aa2 a45dfb0 47e4aa2 a45dfb0 47e4aa2 facd13e 47e4aa2 facd13e 47e4aa2 facd13e 47e4aa2 |
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 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 |
import logging
import os
import shutil
from openai import OpenAI
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
import gradio as gr
from app.config import OPENAI_API_KEY
from app.functions.database_handling import BASE_DB_PATH # Aggiungi questo import
from app.configs.prompts import SYSTEM_PROMPTS
logging.basicConfig(level=logging.INFO)
def get_system_prompt(prompt_type="tutor"):
"""Seleziona il prompt di sistema appropriato"""
return SYSTEM_PROMPTS.get(prompt_type, SYSTEM_PROMPTS["tutor"])
def answer_question(question, db_name, prompt_type="tutor", chat_history=None):
"""
Risponde alla domanda 'question' usando i documenti del database 'db_name'.
Restituisce una lista di 2 messaggi in formato:
[
{"role": "user", "content": <domanda>},
{"role": "assistant", "content": <risposta>}
]
In questa versione, viene effettuato il log dei 'chunk' recuperati durante
la ricerca di similarità.
"""
if chat_history is None:
chat_history = []
logging.info(f"Inizio elaborazione domanda: {question} per database: {db_name}")
try:
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
db_path = os.path.join(BASE_DB_PATH, f"faiss_index_{db_name}") # Percorso corretto
logging.info(f"Verifico esistenza database in: {db_path}")
if not os.path.exists(db_path):
logging.warning(f"Database {db_name} non trovato in {db_path}")
return [
{"role": "user", "content": question},
{"role": "assistant", "content": f"Database non trovato in {db_path}"}
]
# Carica l'indice FAISS
vectorstore = FAISS.load_local(db_path, embeddings, allow_dangerous_deserialization=True)
# Cerca i documenti (chunk) più simili
relevant_docs = vectorstore.similarity_search(question, k=3)
# Logga i chunk recuperati
for idx, doc in enumerate(relevant_docs):
logging.info(f"--- Chunk {idx+1} ---")
logging.info(doc.page_content)
logging.info("---------------------")
# Prepara il contesto dai documenti
context = "\n".join([doc.page_content for doc in relevant_docs])
client = OpenAI(api_key=OPENAI_API_KEY)
messages = [
{"role": "system", "content": SYSTEM_PROMPTS[prompt_type].format(context=context)},
{"role": "user", "content": question}
]
# Esegui la chiamata a OpenAI
response = client.chat.completions.create(
model="gpt-4o-mini",
messages=messages,
temperature=0.5,
max_tokens=3048
)
answer = response.choices[0].message.content
return [
{"role": "user", "content": question},
{"role": "assistant", "content": answer}
]
except Exception as e:
logging.error(f"Errore durante la generazione della risposta: {e}")
return [
{"role": "user", "content": question},
{"role": "assistant", "content": f"Si è verificato un errore: {str(e)}"}
]
if __name__ == "__main__":
pass
|