pip install llama-index

pip install llama-index-llms-gemini

pip install -q llama-index google-generativeai

pip install llama-index-embeddings-huggingface

curl \
  -H 'Content-Type: application/json' \
  -d '{"contents":[{"parts":[{"text":"Explain how AI works"}]}]}' \
  -X POST 'https://generativelanguage.googleapis.com/v1beta/models/gemini-pro:generateContent?key=YOUR_API_KEY'

google_gemini_api = "AIzaSyD0iBVrNDJ8e2-----9p8DiuNEXCaykqjQ"

import os
os.environ["GOOGLE_API_KEY"] = google_gemini_api
os.environ["MODEL_NAME"] = "models/gemini-pro"

from llama_index.llms.gemini import Gemini
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
# Mô hình LLM
llm = Gemini(model_name="models/gemini-pro", api_key=os.environ["GOOGLE_API_KEY"])

# Mô hình embedding
embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")

from llama_index.core import Settings
Settings.llm = llm
Settings.embed_model = embed_model

from llama_index.core import SimpleDirectoryReader, VectorStoreIndex
from llama_index.core.node_parser import TokenTextSplitter

# Đọc dữ liệu từ thư mục data_animal
reader = SimpleDirectoryReader(input_dir="/home/trinh.quang.huy/LLM_project/data_animal")
documents = reader.load_data()
print("len documents: ", len(documents))

# Tách chunk bằng TokenTextSplitter
splitter = TokenTextSplitter(
    chunk_size=512,
    chunk_overlap=10,
    separator=" "
)
nodes = splitter.get_nodes_from_documents(documents)
print("len nodes: ", len(nodes))

index = VectorStoreIndex(nodes)
index.storage_context.persist(persist_dir="./animal_index_storage")

from llama_index.core import StorageContext, load_index_from_storage
PERSIST_DIR = "animal_index_storage"
storage_context = StorageContext.from_defaults(persist_dir=PERSIST_DIR)
index = load_index_from_storage(storage_context)

from llama_index.core import PromptTemplate

QA_PROMPT_TMPL = (
    "Your task is to answer all the questions that users ask based only on the context information provided below.\n"
    "Please answer the question at length and in detail, with full meaning.\n"
    "In the answer there is no sentence such as: based on the context provided.\n"
    "Context information is below.\n"
    "---------------------\n"
    "{context_str}\n"
    "---------------------\n"
    "Given the context information and not prior knowledge, "
    "answer the query.\n"
    "Query: {query_str}\n"
    "Answer: "
)

qa_prompt = PromptTemplate(QA_PROMPT_TMPL)
print(qa_prompt)

query_engine = index.as_query_engine(similarity_top_k=2)
query_engine.update_prompts(
    {"response_synthesizer:text_qa_template": qa_prompt}
)

question_1 = "What color is the polar bear?"

Answer: Polar bears have thick, double-layered fur that is white in color. The white color helps them camouflage well in snow and ice environments.

question_2 = "Average weight of Hippo?"

Answer: The provided context does not mention anything about hippos, so I cannot answer this question.