Spaces:
Sleeping
Sleeping
Delete APP.PY
Browse files
APP.PY
DELETED
|
@@ -1,114 +0,0 @@
|
|
| 1 |
-
import os
|
| 2 |
-
import gradio as gr
|
| 3 |
-
from pypdf import PdfReader
|
| 4 |
-
from groq import Groq
|
| 5 |
-
|
| 6 |
-
def extract_text_from_pdf(pdf_path):
|
| 7 |
-
reader = PdfReader(pdf_path)
|
| 8 |
-
pages = []
|
| 9 |
-
for i, page in enumerate(reader.pages):
|
| 10 |
-
text = page.extract_text()
|
| 11 |
-
if text and text.strip():
|
| 12 |
-
pages.append(f"[Page {i+1}]\n{text.strip()}")
|
| 13 |
-
return "\n\n".join(pages) if pages else None
|
| 14 |
-
|
| 15 |
-
def truncate_text(text, max_chars=6000):
|
| 16 |
-
if len(text) <= max_chars:
|
| 17 |
-
return text
|
| 18 |
-
return text[:4000] + "\n\n...[middle truncated]...\n\n" + text[-2000:]
|
| 19 |
-
|
| 20 |
-
def ask_groq(pdf_text, question, model, history):
|
| 21 |
-
client = Groq(api_key=os.environ["GROQ_API_KEY"])
|
| 22 |
-
safe_pdf_text = truncate_text(pdf_text)
|
| 23 |
-
|
| 24 |
-
messages = [{
|
| 25 |
-
"role": "system",
|
| 26 |
-
"content": (
|
| 27 |
-
"You are a helpful assistant. Answer questions ONLY from the provided PDF. "
|
| 28 |
-
"If the answer is not in the document, say so. Mention page numbers when possible."
|
| 29 |
-
)
|
| 30 |
-
}]
|
| 31 |
-
|
| 32 |
-
for user_msg, bot_msg in history:
|
| 33 |
-
messages.append({"role": "user", "content": user_msg})
|
| 34 |
-
messages.append({"role": "assistant", "content": bot_msg})
|
| 35 |
-
|
| 36 |
-
if not history:
|
| 37 |
-
content = f"PDF Content:\n\n{safe_pdf_text}\n\n---\nQuestion: {question}"
|
| 38 |
-
else:
|
| 39 |
-
content = question
|
| 40 |
-
|
| 41 |
-
messages.append({"role": "user", "content": content})
|
| 42 |
-
|
| 43 |
-
response = client.chat.completions.create(
|
| 44 |
-
model=model,
|
| 45 |
-
messages=messages,
|
| 46 |
-
temperature=0.2,
|
| 47 |
-
max_tokens=1024,
|
| 48 |
-
)
|
| 49 |
-
return response.choices[0].message.content
|
| 50 |
-
|
| 51 |
-
def load_pdf(pdf_file, state):
|
| 52 |
-
if pdf_file is None:
|
| 53 |
-
return state, "β οΈ Please upload a PDF.", gr.update(interactive=False)
|
| 54 |
-
pdf_text = extract_text_from_pdf(pdf_file)
|
| 55 |
-
if not pdf_text:
|
| 56 |
-
return state, "β Could not extract text. Try a non-scanned PDF.", gr.update(interactive=False)
|
| 57 |
-
state["pdf_text"] = pdf_text
|
| 58 |
-
state["history"] = []
|
| 59 |
-
pages = pdf_text.count("[Page ")
|
| 60 |
-
words = len(pdf_text.split())
|
| 61 |
-
return state, f"β
Loaded! {pages} page(s), ~{words:,} words.\nNow ask your questions!", gr.update(interactive=True)
|
| 62 |
-
|
| 63 |
-
def ask_question(question, model, state, chat_history):
|
| 64 |
-
if not question.strip():
|
| 65 |
-
return chat_history, state, ""
|
| 66 |
-
if not state.get("pdf_text"):
|
| 67 |
-
chat_history.append((question, "β οΈ Please load a PDF first."))
|
| 68 |
-
return chat_history, state, ""
|
| 69 |
-
try:
|
| 70 |
-
answer = ask_groq(state["pdf_text"], question, model, state["history"])
|
| 71 |
-
state["history"].append((question, answer))
|
| 72 |
-
chat_history.append((question, answer))
|
| 73 |
-
except Exception as e:
|
| 74 |
-
chat_history.append((question, f"β Error: {e}"))
|
| 75 |
-
return chat_history, state, ""
|
| 76 |
-
|
| 77 |
-
def clear_chat(state):
|
| 78 |
-
state["history"] = []
|
| 79 |
-
return [], state
|
| 80 |
-
|
| 81 |
-
MODELS = [
|
| 82 |
-
"llama-3.3-70b-versatile",
|
| 83 |
-
"llama-3.1-8b-instant",
|
| 84 |
-
"llama3-70b-8192",
|
| 85 |
-
"gemma2-9b-it",
|
| 86 |
-
]
|
| 87 |
-
|
| 88 |
-
with gr.Blocks(title="PDF Q&A β Groq", theme=gr.themes.Soft()) as demo:
|
| 89 |
-
|
| 90 |
-
state = gr.State({"pdf_text": "", "history": []})
|
| 91 |
-
|
| 92 |
-
gr.Markdown("# π PDF Q&A App\n### Upload a PDF β Ask questions β Get AI answers")
|
| 93 |
-
|
| 94 |
-
with gr.Row():
|
| 95 |
-
with gr.Column(scale=1):
|
| 96 |
-
pdf_input = gr.File(label="Upload PDF", file_types=[".pdf"], type="filepath")
|
| 97 |
-
upload_btn = gr.Button("π€ Load PDF", variant="primary", size="lg")
|
| 98 |
-
status = gr.Textbox(label="Status", value="Upload a PDF to begin...", lines=3, interactive=False)
|
| 99 |
-
model_drop = gr.Dropdown(choices=MODELS, value=MODELS[0], label="π€ Groq Model")
|
| 100 |
-
|
| 101 |
-
with gr.Column(scale=2):
|
| 102 |
-
chatbot = gr.Chatbot(label="Conversation", height=420)
|
| 103 |
-
with gr.Row():
|
| 104 |
-
q_input = gr.Textbox(label="Your Question", placeholder="e.g. Summarize this document", lines=2, interactive=False, scale=4)
|
| 105 |
-
with gr.Column(scale=1):
|
| 106 |
-
ask_btn = gr.Button("π Ask", variant="primary")
|
| 107 |
-
clear_btn = gr.Button("ποΈ Clear", variant="secondary")
|
| 108 |
-
|
| 109 |
-
upload_btn.click(load_pdf, [pdf_input, state], [state, status, q_input])
|
| 110 |
-
ask_btn.click(ask_question, [q_input, model_drop, state, chatbot], [chatbot, state, q_input])
|
| 111 |
-
q_input.submit(ask_question, [q_input, model_drop, state, chatbot], [chatbot, state, q_input])
|
| 112 |
-
clear_btn.click(clear_chat, [state], [chatbot, state])
|
| 113 |
-
|
| 114 |
-
demo.launch() # β No share=True needed on Hugging Face
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|