Update app.py

app.py

@@ -1,42 +1,24 @@
-import os
-os.environ['USE_TORCH'] = '1'
-
-from doctr.io import DocumentFile
-from doctr.models import ocr_predictor
 import gradio as gr
 from PIL import Image
-import base64
-from utils import HocrParser
 import google.generativeai as genai
+from IPython.display import Markdown
 
 api_key = 'AIzaSyB7-RzBwTAfVA-7ZGk2mEOQwOxshpwzhpM' # put your API key here
 genai.configure(api_key=api_key)
-geminiModel = genai.GenerativeModel(model_name='gemini-pro')
-predictor = ocr_predictor(det_arch='db_mobilenet_v3_large', reco_arch='crnn_vgg16_bn',pretrained=True)
-predictor.reco_predictor.model.cfg['vocab']='0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ!”#$%&’()*+,-./:;<=>?@[]^_`{|}~°£€¥¢฿áàảạãăắằẳẵặâấầẩẫậéèẻẽẹêếềểễệóòỏõọôốồổộỗơớờởợỡúùủũụưứừửữựiíìỉĩịýỳỷỹỵÁÀẢẠÃĂẮẰẲẴẶÂẤẦẨẪẬÉÈẺẼẸÊẾỀỂỄỆÓÒỎÕỌÔỐỒỔỘỖƠỚỜỞỢỠÚÙỦŨỤƯỨỪỬỮỰIÍÌỈĨỊÝỲỶỸỴ'
+geminiModel = genai.GenerativeModel(model_name='gemini-pro-vision')
 title="DocTR OCR (PDL Demo)"
 description="Upload an image to get the OCR results !"
 
+def to_markdown(text):
+  text = text.replace('•', '  *')
+  return Markdown(textwrap.indent(text, '> ', predicate=lambda _: True))
+    
 def greet(img):
-    img.save("out.jpg")
-    doc = DocumentFile.from_images("out.jpg")
-    output=predictor(doc)
-
-    xml_outputs = output.export_as_xml()
-    parser = HocrParser()
-
-    res=""
-    for obj in output.pages:
-      for obj1 in obj.blocks:
-        for obj2 in obj1.lines:
-          for obj3 in obj2.words:
-            res=res + " " + obj3.value
-        res=res + "\n"
-      res=res + "\n"
     # 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: " + res
     # print(prompt)
-    response  = geminiModel.generate_content(prompt)
+    response  = geminiModel.generate_content([prompt,img], stream=True)
+    response.resolve()
     print(response)
     res = response.text
     _output_name = "RESULT_OCR.txt"