Spaces:
Runtime error
Runtime error
File size: 4,234 Bytes
e1f19eb |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 |
import json
import logging
import os
import sys
import gradio as gr
import openai
from dotenv import load_dotenv
from llama_index import SimpleDirectoryReader, VectorStoreIndex, load_index_from_storage
from llama_index.indices.service_context import ServiceContext
from llama_index.llms import OpenAI
from llama_index.node_parser import SimpleNodeParser
from llama_index.node_parser.extractors import (
MetadataExtractor,
SummaryExtractor,
QuestionsAnsweredExtractor,
TitleExtractor,
KeywordExtractor,
)
from llama_index.storage.storage_context import StorageContext
from llama_index.text_splitter import TokenTextSplitter
# loads dotenv lib to retrieve API keys from .env file
load_dotenv()
openai.api_key = os.getenv("OPENAI_API_KEY")
# enable INFO level logging
logging.basicConfig(stream=sys.stdout, level=logging.INFO)
logging.getLogger().addHandler(logging.StreamHandler(stream=sys.stdout))
# define LLM service
llm = OpenAI(temperature=0.1, model_name="gpt-3.5-turbo-16k", max_tokens=512)
service_context = ServiceContext.from_defaults(llm=llm)
# construct text splitter to split texts into chunks for processing
text_splitter = TokenTextSplitter(separator=" ", chunk_size=512, chunk_overlap=128)
# set the global service context object, avoiding passing service_context when building the index
from llama_index import set_global_service_context
set_global_service_context(service_context)
# create metadata extractor
metadata_extractor = MetadataExtractor(
extractors=[
TitleExtractor(nodes=1, llm=llm),
QuestionsAnsweredExtractor(questions=3, llm=llm),
SummaryExtractor(summaries=["prev", "self"], llm=llm),
KeywordExtractor(keywords=10, llm=llm)
],
)
# create node parser to parse nodes from document
node_parser = SimpleNodeParser(
text_splitter=text_splitter,
metadata_extractor=metadata_extractor,
)
# loading documents
documents_2023 = SimpleDirectoryReader(input_files=["data/data.pdf"], filename_as_id=True).load_data()
print(f"loaded data with {len(documents_2023)} pages")
def load_index():
try:
# load storage context
storage_context = StorageContext.from_defaults(persist_dir="./storage")
# try to load the index from storage
index = load_index_from_storage(storage_context)
logging.info("Index loaded from storage.")
except FileNotFoundError:
# if index not found, create a new one
logging.info("Index not found. Creating a new one...")
nodes_2023 = node_parser.get_nodes_from_documents(documents_2023)
print(f"loaded nodes_2022 with {len(nodes_2023)} nodes")
# print metadata in json format
for node in nodes_2023:
metadata_json = json.dumps(node.metadata, indent=4) # Convert metadata to formatted JSON
print(metadata_json)
# based on the nodes and service_context, create index
index = VectorStoreIndex(nodes=nodes_2023, service_context=service_context)
# Persist index to disk
index.storage_context.persist()
logging.info("New index created and persisted to storage.")
return index
def data_querying(input_text):
# Load index
index = load_index()
# queries the index with the input text
response = index.as_query_engine().query(input_text)
return response.response
iface = gr.Interface(fn=data_querying,
inputs=gr.components.Textbox(lines=3, label="Inserte su pregunta aquí"),
outputs=gr.components.Textbox(lines=3, label="Respuesta"),
examples=[
["¿Quien eres?"],
["¿Cómo te llamas?"],
["¿Cómo puedes ayudarme?"],
["¿Qué puedo hacer si perdí mi contraseña?"],
["¿Qué puedo hacer si perdí mi tarjeta de cŕedito?"],
["Genera un reporte o lista de pasos de toda la info que necesito para abrir una cuenta"],
],
title="Demo IA Conversacional",
description="Pregúntame lo que quieras, te responderé con lo que sé.")
iface.launch(share=False)
|