DocTRv2 / app.py
hantech's picture
Update app.py
0e4fd71 verified
raw
history blame contribute delete
No virus
1.27 kB
import gradio as gr
from PIL import Image
import google.generativeai as genai
api_key = 'AIzaSyB7-RzBwTAfVA-7ZGk2mEOQwOxshpwzhpM' # put your API key here
genai.configure(api_key=api_key)
geminiModel = genai.GenerativeModel(model_name='gemini-pro-vision')
title="DocTR OCR (PDL Demo)"
description="Upload an image to get the OCR results !"
def greet(img):
# generate text
prompt = "the input text in vietnamese, please add accend and take this peace of information and give all the information in point wise better format also give some recomendation related to them"
# print(prompt)
response = geminiModel.generate_content([prompt,img], stream=True)
response.resolve()
print(response)
res = response.text
_output_name = "RESULT_OCR.txt"
_output_name_pdf="RESULT_OCR.pdf"
open(_output_name, 'w').close() # clear file
with open(_output_name, "w", encoding="utf-8", errors="ignore") as f:
f.write(res)
print("Writing into file")
return res, _output_name
demo = gr.Interface(fn=greet,
inputs=gr.Image(type="pil"),
outputs=["text", "file"],
title=title,
description=description
)
demo.launch(debug=True)