File size: 3,506 Bytes
62f8d39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from fastapi import FastAPI, UploadFile, File, HTTPException
from pydantic import BaseModel
from llama_index.core import StorageContext, load_index_from_storage, VectorStoreIndex, SimpleDirectoryReader, ChatPromptTemplate
from llama_index.llms.huggingface import HuggingFaceInferenceAPI
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core import Settings
import os
from dotenv import load_dotenv
import shutil

# Load environment variables
load_dotenv()

app = FastAPI()

# Configure the Llama index settings
Settings.llm = HuggingFaceInferenceAPI(
    model_name="meta-llama/Meta-Llama-3-8B-Instruct",
    tokenizer_name="meta-llama/Meta-Llama-3-8B-Instruct",
    context_window=3900,
    token=os.getenv("HF_TOKEN"),
    max_new_tokens=1000,
    generate_kwargs={"temperature": 0.5},
)
Settings.embed_model = HuggingFaceEmbedding(
    model_name="BAAI/bge-small-en-v1.5"
)

# Define the directory for persistent storage and data
PERSIST_DIR = "./db"
DATA_DIR = "data"

# Ensure data directory exists
os.makedirs(DATA_DIR, exist_ok=True)
os.makedirs(PERSIST_DIR, exist_ok=True)

class Query(BaseModel):
    question: str

def data_ingestion():
    documents = SimpleDirectoryReader(DATA_DIR).load_data()
    storage_context = StorageContext.from_defaults()
    index = VectorStoreIndex.from_documents(documents)
    index.storage_context.persist(persist_dir=PERSIST_DIR)

def handle_query(query):
    storage_context = StorageContext.from_defaults(persist_dir=PERSIST_DIR)
    index = load_index_from_storage(storage_context)
    chat_text_qa_msgs = [
        (
            "user",
            """You are Q&A assistant named CHAT-DOC. Your main goal is to provide answers as accurately as possible, based on the instructions and context you have been given. If a question does not match the provided context or is outside the scope of the document, kindly advise the user to ask questions within the context of the document.

            Context:

            {context_str}

            Question:

            {query_str}

            """
        )
    ]
    text_qa_template = ChatPromptTemplate.from_messages(chat_text_qa_msgs)
    query_engine = index.as_query_engine(text_qa_template=text_qa_template)
    answer = query_engine.query(query)
    
    if hasattr(answer, 'response'):
        return answer.response
    elif isinstance(answer, dict) and 'response' in answer:
        return answer['response']
    else:
        return "Sorry, I couldn't find an answer."

@app.post("/upload")
async def upload_file(file: UploadFile = File(...)):
    file_extension = os.path.splitext(file.filename)[1].lower()
    if file_extension not in [".pdf", ".docx", ".txt"]:
        raise HTTPException(status_code=400, detail="Invalid file type. Only PDF, DOCX, and TXT are allowed.")
    
    file_path = os.path.join(DATA_DIR, file.filename)
    with open(file_path, "wb") as buffer:
        shutil.copyfileobj(file.file, buffer)
    
    data_ingestion()
    return {"message": "File uploaded and processed successfully"}

@app.post("/query")
async def query_document(query: Query):
    if not os.listdir(DATA_DIR):
        raise HTTPException(status_code=400, detail="No document has been uploaded yet.")
    
    response = handle_query(query.question)
    return {"response": response}

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)