File size: 2,166 Bytes
1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 9a0264a 1b3bcd8 |
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 |
import gradio as gr
from langchain_community.llms import HuggingFaceHub
from langchain_core.output_parsers import StrOutputParser
from langchain import hub
from langchain_community.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import Chroma
from langchain_huggingface import HuggingFaceEmbeddings
from huggingface_hub import InferenceClient
from rerankers import Reranker
import os
loader = PyPDFLoader("Constitucion_espa帽ola.pdf")
documents = loader.load()
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
docs_split = text_splitter.split_documents(documents)
embedding_function = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
vectordb = Chroma.from_documents(docs_split, embedding_function)
client = InferenceClient("google/flan-t5-base", token=os.getenv("HUGGINGFACEHUB_API_TOKEN"))
ranker = Reranker("answerdotai/answerai-colbert-small-v1", model_type='colbert')
def generate_text(context, query):
inputs = f"Context: {context} Question: {query}"
response = client.chat_completion(inputs=inputs, task="text2text-generation")
return response
def test_rag_reranking(query, ranker):
docs = vectordb.similarity_search_with_score(query)
context = []
for doc, score in docs:
if score < 7:
doc_details = doc.to_json()['kwargs']
context.append(doc_details['page_content'])
if len(context) > 0:
useful_context = context[0]
generation = generate_text(useful_context, query)
return generation
else:
return "No se encontr贸 informaci贸n suficiente para responder."
ranked = ranker.rerank(query=query, documents=raw_contexts, top_k=1)
best_context = ranked[0]["content"]
return generate_text(best_context, query)
def responder_chat(message, history):
respuesta = test_rag_reranking(message, ranker)
return respuesta
demo = gr.ChatInterface(
fn=responder_chat,
title="Chatbot sobre la constituci贸n espa帽ola",
theme="soft"
)
if __name__ == "__main__":
demo.launch()
|