Spaces:
Sleeping
Sleeping
File size: 3,415 Bytes
8a1b6fa 6fb0b86 9787256 6fb0b86 cce1fbb 6fb0b86 cce1fbb 8a1b6fa 9787256 8a1b6fa 6fb0b86 8a1b6fa 9787256 8a1b6fa 9787256 6fb0b86 9787256 8a1b6fa 6fb0b86 8a1b6fa 6fb0b86 8a1b6fa 9787256 8a1b6fa 9787256 8a1b6fa 6fb0b86 8a1b6fa 9787256 8a1b6fa 9787256 8a1b6fa 9787256 8a1b6fa 6fb0b86 8cbfc80 8a1b6fa 9787256 8a1b6fa 9787256 8a1b6fa 6fb0b86 8cbfc80 8a1b6fa | 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 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 | # =========================
# IMPORTS
# =========================
import os
import tempfile
import gradio as gr
from groq import Groq
from duckduckgo_search import DDGS
from langchain_community.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import Chroma
# =========================
# CONFIG
# =========================
GROQ_API_KEY = os.getenv("GROQ_API_KEY") # HF SECRET KEY
client = Groq(api_key=GROQ_API_KEY)
vectorstore = None
retriever = None
# =========================
# PROMPT
# =========================
def build_prompt(context, question):
return f"""
You are an expert AI assistant.
Use ONLY the context below.
If answer is not present, say "Not found in document".
CONTEXT:
{context}
QUESTION:
{question}
ANSWER:
"""
# =========================
# WEB SEARCH (FALLBACK)
# =========================
def web_search(query):
results = []
with DDGS() as ddgs:
for r in ddgs.text(query, max_results=3):
results.append(r["body"])
return "\n\n".join(results)
# =========================
# PROCESS PDF (HF SAFE)
# =========================
def process_pdf(file):
global vectorstore, retriever
# safe file handling
file_path = file.name
loader = PyPDFLoader(file_path)
documents = loader.load()
splitter = RecursiveCharacterTextSplitter(
chunk_size=600,
chunk_overlap=100
)
chunks = splitter.split_documents(documents)
embedding_model = HuggingFaceEmbeddings(
model_name="sentence-transformers/all-MiniLM-L6-v2"
)
vectorstore = Chroma.from_documents(
documents=chunks,
embedding=embedding_model
)
retriever = vectorstore.as_retriever(search_kwargs={"k": 4})
return "✅ PDF successfully processed"
# =========================
# RAG ENGINE (HYBRID)
# =========================
def ask_rag(query):
global retriever
if retriever is None:
return "⚠️ Please upload a PDF first."
docs = retriever.invoke(query)
pdf_context = "\n\n".join([d.page_content for d in docs])
# hybrid fallback
if len(pdf_context.strip()) < 50:
web_context = web_search(query)
context = pdf_context + "\n\nWEB:\n" + web_context
else:
context = pdf_context
prompt = build_prompt(context, query)
response = client.chat.completions.create(
messages=[{"role": "user", "content": prompt}],
model="llama3-8b-8192"
)
return response.choices[0].message.content
# =========================
# CHAT FUNCTION (FIXED FORMAT)
# =========================
def chat(user_message, history):
response = ask_rag(user_message)
if history is None:
history = []
history.append((user_message, response))
return history, history
# =========================
# UI (HUGGING FACE SAFE)
# =========================
with gr.Blocks() as app:
gr.Markdown("# 🧠 Hybrid RAG Chatbot (PDF + Web)")
file = gr.File(label="Upload PDF")
status = gr.Textbox(label="Status")
chatbot = gr.Chatbot() # IMPORTANT FIX
msg = gr.Textbox(placeholder="Ask your question...")
state = gr.State([])
file.change(process_pdf, file, status)
msg.submit(chat, [msg, state], [chatbot, state])
app.launch() |