File size: 2,326 Bytes
7f269b9 |
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 |
import os
import json
from typing import List
from fastapi import FastAPI, UploadFile, File, HTTPException
from pydantic import BaseModel
from langchain_groq import ChatGroq
from langchain.document_loaders import PyPDFLoader
# Load API key securely from environment variable
API_KEY = os.getenv("GROQ_API_KEY")
if not API_KEY:
raise ValueError("GROQ_API_KEY environment variable not set.")
app = FastAPI(title="PDF Question Extractor", version="1.0")
# Pydantic model for response
class ExtractionResult(BaseModel):
answers: List[str]
# Initialize LLM
def get_llm():
return ChatGroq(
model="llama-3.3-70b-versatile",
temperature=0,
max_tokens=1024,
api_key=API_KEY
)
llm = get_llm()
@app.post("/extract-answers/")
async def extract_answers(file: UploadFile = File(...)):
try:
# Save the uploaded file temporarily
file_path = f"./temp_{file.filename}"
with open(file_path, "wb") as buffer:
buffer.write(file.file.read())
# Load and extract text from PDF
loader = PyPDFLoader(file_path)
pages = loader.load_and_split()
all_page_content = "\n".join(page.page_content for page in pages)
# JSON schema definition
schema_dict = ExtractionResult.model_json_schema()
schema = json.dumps(schema_dict, indent=2)
# System message
system_message = (
"You are a document analysis tool that extracts the options and correct answers from the provided document content. "
"The output must be a JSON object that strictly follows the schema: " + schema
)
# User message
user_message = (
"Please extract the correct answers and options (A, B, C, D, E) from the following document content:\n\n"
+ all_page_content
)
# Construct final prompt
prompt = system_message + "\n\n" + user_message
# Get LLM response
response = llm.invoke(prompt, response_format={"type": "json_object"})
# Parse and validate response
result = ExtractionResult.model_validate_json(response.content)
# Cleanup
os.remove(file_path)
return result.model_dump()
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
|