Spaces:
Sleeping
Sleeping
File size: 3,368 Bytes
d16155e 4d136b6 a189d7b d16155e ea1463f e111f52 da97e63 0af46f8 da97e63 0af46f8 da97e63 0af46f8 da97e63 0af46f8 da97e63 0af46f8 da97e63 d16155e 7846b16 d16155e 9fa8865 d16155e e111f52 d16155e b4c6230 |
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 |
import qdrant_client
from llama_index.core import VectorStoreIndex, ServiceContext, SimpleDirectoryReader
from llama_index.core import load_index_from_storage
from llama_index.llms.ollama import Ollama
from llama_index.llms.llama_cpp import LlamaCPP
from llama_index.llms.llama_cpp.llama_utils import messages_to_prompt, completion_to_prompt
from llama_index.core import StorageContext
from llama_index.vector_stores.qdrant import QdrantVectorStore
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core import Settings
from llama_index.core import set_global_service_context
import gradio as gr
DOC_PATH = './data/pdf_esg'
INDEX_PATH = './storage'
llm = LlamaCPP(
# You can pass in the URL to a GGML model to download it automatically
# model_url='https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF/resolve/main/mistral-7b-instruct-v0.1.Q4_K_M.gguf',
model_url='https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/resolve/main/mistral-7b-instruct-v0.2.Q4_K_M.gguf',
# optionally, you can set the path to a pre-downloaded model instead of model_url
model_path=None,
temperature=0.1,
max_new_tokens=256,
# llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
context_window=4096,
# kwargs to pass to __call__()
generate_kwargs={},
# kwargs to pass to __init__()
# set to at least 1 to use GPU
model_kwargs={"n_gpu_layers": -1},
# transform inputs into Llama2 format
messages_to_prompt=messages_to_prompt,
completion_to_prompt=completion_to_prompt,
verbose=True,
)
# Settings.llm = Ollama(model="mistral")
Settings.llm = llm
Settings.embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")
embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")
service_context = ServiceContext.from_defaults(llm=llm,embed_model = embed_model)
set_global_service_context(service_context)
def construct_index(doc_path=DOC_PATH, index_store=INDEX_PATH, use_cache=False):
client = qdrant_client.QdrantClient(path="./qdrant_data")
vector_store = QdrantVectorStore(client=client, collection_name="esg")
storage_context = StorageContext.from_defaults(vector_store=vector_store)
if use_cache:
# rebuild storage context
storage_context = StorageContext.from_defaults(persist_dir=index_store)
index = load_index_from_storage(storage_context) # load index
else:
reader = SimpleDirectoryReader(input_dir='./data/pdf_esg')
documents = reader.load_data()
index = VectorStoreIndex.from_documents(documents)
index.storage_context.persist(index_store)
return None
def qabot(input_text, index_store = INDEX_PATH):
storage_context = StorageContext.from_defaults(persist_dir=index_store)
# Load the data
index = load_index_from_storage(storage_context)
query_engine = index.as_query_engine()
response = query_engine.query(input_text)
return response.response
if __name__ == "__main__":
# construct_index(DOC_PATH, use_cache=False)
# create_index_retriever_query_engine()
iface = gr.Interface(fn=qabot, inputs=gr.Textbox(lines=7, label='Enter your query'),
outputs="text",
title="ESG Chatbot")
iface.launch(inline=False) |