import os import time from pathlib import Path from textwrap import dedent from types import SimpleNamespace import gradio as gr from charset_normalizer import detect from chromadb.config import Settings from epub2txt import epub2txt from langchain.chains import RetrievalQA from langchain.docstore.document import Document from langchain.document_loaders import ( CSVLoader, Docx2txtLoader, PDFMinerLoader, TextLoader, ) # from constants import CHROMA_SETTINGS, SOURCE_DIRECTORY, PERSIST_DIRECTORY from langchain.embeddings import HuggingFaceInstructEmbeddings from langchain.llms import HuggingFacePipeline from langchain.text_splitter import ( RecursiveCharacterTextSplitter, ) import torch # FAISS instead of PineCone from langchain.vectorstores import Chroma from loguru import logger from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline import argparse parser = argparse.ArgumentParser('LocalGPT falcon', add_help=False) parser.add_argument('--device_type', type=str, default="cuda", choices=["cpu", "mps", "cuda"], help='device type', ) args = parser.parse_args() ROOT_DIRECTORY = Path(__file__).parent PERSIST_DIRECTORY = f"{ROOT_DIRECTORY}/db" # Define the Chroma settings CHROMA_SETTINGS = Settings( chroma_db_impl="duckdb+parquet", persist_directory=PERSIST_DIRECTORY, anonymized_telemetry=False, ) ns = SimpleNamespace(qa=None) # INSTRUCTORS_EMBEDDINGS_MODEL = "hkunlp/instructor-xl" INSTRUCTORS_EMBEDDINGS_MODEL = "hkunlp/instructor-large" # INSTRUCTORS_EMBEDDINGS_MODEL = "hkunlp/instructor-large" # INSTRUCTORS_EMBEDDINGS_MODEL = "hkunlp/instructor-base" def load_single_document(file_path: str or Path) -> Document: """ingest.py""" # Loads a single document from a file path # encoding = detect(open(file_path, "rb").read()).get("encoding", "utf-8") encoding = detect(Path(file_path).read_bytes()).get("encoding", "utf-8") if file_path.endswith(".txt"): if encoding is None: logger.warning( f" {file_path}'s encoding is None " "Something is fishy, return empty str " ) return Document(page_content="", metadata={"source": file_path}) try: loader = TextLoader(file_path, encoding=encoding) except Exception as exc: logger.warning(f" {exc}, return dummy ") return Document(page_content="", metadata={"source": file_path}) elif file_path.endswith(".pdf"): loader = PDFMinerLoader(file_path) elif file_path.endswith(".csv"): loader = CSVLoader(file_path) elif Path(file_path).suffix in [".docx"]: try: loader = Docx2txtLoader(file_path) except Exception as exc: logger.error(f" {file_path} errors: {exc}") return Document(page_content="", metadata={"source": file_path}) elif Path(file_path).suffix in [".epub"]: # for epub? epub2txt unstructured try: _ = epub2txt(file_path) except Exception as exc: logger.error(f" {file_path} errors: {exc}") return Document(page_content="", metadata={"source": file_path}) return Document(page_content=_, metadata={"source": file_path}) else: if encoding is None: logger.warning( f" {file_path}'s encoding is None " "Likely binary files, return empty str " ) return Document(page_content="", metadata={"source": file_path}) try: loader = TextLoader(file_path) except Exception as exc: logger.error(f" {exc}, returnning empty string") return Document(page_content="", metadata={"source": file_path}) return loader.load()[0] def greet(name): """Test.""" logger.debug(f" name: [{name}] ") return "Hello " + name + "!!" def upload_files(files): """Upload files.""" try: file_paths = [file.name for file in files] except: file_paths = [files] logger.info(file_paths) res = ingest(file_paths) logger.info("Processed:\n{res}") del res ns.qa = load_qa() return file_paths def ingest( file_paths: list ): """Gen Chroma db. torch.cuda.is_available() file_paths = [] """ logger.info("Doing ingest...") documents = [] for file_path in file_paths: documents.append(load_single_document(f"{file_path}")) text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200) texts = text_splitter.split_documents(documents) logger.info(f"Loaded {len(documents)} documents ") logger.info(f"Split into {len(texts)} chunks of text") # Create embeddings logger.info(f"Load InstructEmbeddings model: {INSTRUCTORS_EMBEDDINGS_MODEL}") embeddings = HuggingFaceInstructEmbeddings( model_name=INSTRUCTORS_EMBEDDINGS_MODEL, model_kwargs={"device": args.device_type} ) db = Chroma.from_documents( texts, embeddings, persist_directory=PERSIST_DIRECTORY, client_settings=CHROMA_SETTINGS, ) db.persist() db = None logger.info("Done ingest") return [ [Path(doc.metadata.get("source")).name, len(doc.page_content)] for doc in documents ] # https://huggingface.co/tiiuae/falcon-7b-instruct def gen_local_llm(): """Gen a local llm. localgpt run_localgpt """ model = "tiiuae/falcon-7b-instruct" if args.device_type == "cuda": tokenizer = AutoTokenizer.from_pretrained(model) else: # cpu tokenizer=AutoTokenizer.from_pretrained(model) model=AutoModelForCausalLM.from_pretrained(model, trust_remote_code=True) pipe = pipeline( "text-generation", model=model, tokenizer=tokenizer, torch_dtype=torch.float32 if args.device_type =="cpu" else torch.bfloat16, trust_remote_code=True, device_map="cpu" if args.device_type =="cpu" else "auto", max_length=2048, temperature=0, top_p=0.95, top_k=10, repetition_penalty=1.15, num_return_sequences=1, pad_token_id=tokenizer.eos_token_id ) local_llm = HuggingFacePipeline(pipeline=pipe) return local_llm def load_qa(): """Gen qa.""" logger.info("Doing qa") embeddings = HuggingFaceInstructEmbeddings( model_name=INSTRUCTORS_EMBEDDINGS_MODEL, model_kwargs={"device": args.device_type} ) # xl 4.96G, large 3.5G, db = Chroma( persist_directory=PERSIST_DIRECTORY, embedding_function=embeddings, client_settings=CHROMA_SETTINGS, ) retriever = db.as_retriever() llm = gen_local_llm() # "tiiuae/falcon-7b-instruct" qa = RetrievalQA.from_chain_type( llm=llm, chain_type="stuff", retriever=retriever, return_source_documents=True ) logger.info("Done qa") return qa def main1(): """Lump codes""" with gr.Blocks() as demo: iface = gr.Interface(fn=greet, inputs="text", outputs="text") iface.launch() demo.launch() def main(): """Do blocks.""" logger.info(f"ROOT_DIRECTORY: {ROOT_DIRECTORY}") with gr.Blocks(theme=gr.themes.Soft()) as demo: with gr.Accordion("Info", open=False): _ = """ Talk to your docs (.pdf, .docx, .csv, .txt .md). It takes quite a while to ingest docs (10-30 min. depending on net, RAM, CPU etc.). """ gr.Markdown(dedent(_)) title = """

LocalGPT with Falcon

Upload your docs (.pdf, .docx, .csv, .txt .md) by clicking the "Load docs to LangChain" and wait until the upload is complete,
when everything is ready, you can start asking questions about the docs

""" gr.HTML(title) with gr.Tab("Upload files"): # Upload files and generate embeddings database file_output = gr.File() upload_button = gr.UploadButton( "Load docs to LangChain", file_count="multiple", ) upload_button.upload(upload_files, upload_button, file_output) chatbot = gr.Chatbot() msg = gr.Textbox(label="Query") clear = gr.Button("Clear") def respond(message, chat_history): if ns.qa is None: # no files processed yet bot_message = "Provide some file(s) for processsing first." chat_history.append((message, bot_message)) return "", chat_history try: res = ns.qa(message) answer, docs = res["result"], res["source_documents"] bot_message = f"{answer}" except Exception as exc: logger.error(exc) bot_message = f"bummer! {exc}" chat_history.append((message, bot_message)) return "", chat_history msg.submit(respond, [msg, chatbot], [msg, chatbot]) clear.click(lambda: None, None, chatbot, queue=False) try: from google import colab share = True # start share when in colab except Exception: share = False demo.launch(share=share) if __name__ == "__main__": main()