Spaces:
Sleeping
Sleeping
Create app.py
#1
by
tanya17 - opened
app.py
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import google.generativeai as genai
|
| 3 |
+
from PyPDF2 import PdfReader
|
| 4 |
+
from paddleocr import PaddleOCR
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# Step 1: Gemini API Key (must be set in Hugging Face Secrets)
|
| 8 |
+
genai.configure(api_key=os.getenv("GEMINI_API_KEY"))
|
| 9 |
+
model = genai.GenerativeModel('gemini-pro')
|
| 10 |
+
|
| 11 |
+
# Step 2: OCR Setup
|
| 12 |
+
ocr_model = PaddleOCR(use_angle_cls=True, lang='en')
|
| 13 |
+
documents = []
|
| 14 |
+
|
| 15 |
+
def extract_text(file):
|
| 16 |
+
ext = os.path.splitext(file.name)[1].lower()
|
| 17 |
+
text = ""
|
| 18 |
+
if ext == ".pdf":
|
| 19 |
+
reader = PdfReader(file)
|
| 20 |
+
for page in reader.pages:
|
| 21 |
+
text += page.extract_text() or ""
|
| 22 |
+
elif ext in [".jpg", ".jpeg", ".png"]:
|
| 23 |
+
result = ocr_model.ocr(file.name)
|
| 24 |
+
text = " ".join([line[1][0] for line in result[0]])
|
| 25 |
+
return text
|
| 26 |
+
|
| 27 |
+
def process_files(files):
|
| 28 |
+
global documents
|
| 29 |
+
documents = []
|
| 30 |
+
for f in files:
|
| 31 |
+
text = extract_text(f)
|
| 32 |
+
documents.append({"filename": f.name, "text": text})
|
| 33 |
+
return f"{len(files)} files processed and stored."
|
| 34 |
+
|
| 35 |
+
def answer_query(query):
|
| 36 |
+
if not documents:
|
| 37 |
+
return "Please upload and process files first."
|
| 38 |
+
|
| 39 |
+
prompt = "You are a research assistant. Analyze the following documents and answer the query.\n"
|
| 40 |
+
for i, doc in enumerate(documents):
|
| 41 |
+
prompt += f"\nDocument {i+1} ({doc['filename']}):\n{doc['text'][:2000]}\n"
|
| 42 |
+
prompt += f"\n\nQuestion: {query}\nAnswer with key themes and cite document numbers."
|
| 43 |
+
|
| 44 |
+
response = model.generate_content(prompt)
|
| 45 |
+
return response.text
|
| 46 |
+
|
| 47 |
+
# Step 3: Gradio Interface
|
| 48 |
+
with gr.Blocks() as demo:
|
| 49 |
+
gr.Markdown("# 📄 Gemini Document Research & Theme Identification Chatbot")
|
| 50 |
+
|
| 51 |
+
with gr.Row():
|
| 52 |
+
file_input = gr.File(file_types=[".pdf", ".jpg", ".png"], file_count="multiple", label="Upload Documents")
|
| 53 |
+
process_btn = gr.Button("Process Documents")
|
| 54 |
+
|
| 55 |
+
process_output = gr.Textbox(label="Processing Status")
|
| 56 |
+
|
| 57 |
+
with gr.Row():
|
| 58 |
+
query_input = gr.Textbox(label="Ask a question")
|
| 59 |
+
query_btn = gr.Button("Get Answer")
|
| 60 |
+
|
| 61 |
+
answer_output = gr.Textbox(label="Answer with Themes and Citations", lines=10)
|
| 62 |
+
|
| 63 |
+
process_btn.click(fn=process_files, inputs=[file_input], outputs=[process_output])
|
| 64 |
+
query_btn.click(fn=answer_query, inputs=[query_input], outputs=[answer_output])
|
| 65 |
+
|
| 66 |
+
demo.launch()
|