Update app.py
Browse files
app.py
CHANGED
|
@@ -4,6 +4,8 @@ import streamlit as st
|
|
| 4 |
import easyocr
|
| 5 |
import PIL
|
| 6 |
from PIL import Image, ImageDraw
|
|
|
|
|
|
|
| 7 |
|
| 8 |
def rectangle(image, result):
|
| 9 |
# https://www.blog.pythonlibrary.org/2021/02/23/drawing-shapes-on-images-with-python-and-pillow/
|
|
@@ -17,58 +19,126 @@ def rectangle(image, result):
|
|
| 17 |
st.image(image)
|
| 18 |
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
# main title
|
| 21 |
-
st.title("Get text from image with
|
| 22 |
|
| 23 |
# subtitle
|
| 24 |
-
st.markdown("##
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
#
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
#
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
# bar = st.progress(0)
|
| 41 |
-
|
| 42 |
-
# for i in range(100):
|
| 43 |
-
# Update the progress bar with each iteration.
|
| 44 |
-
# latest_iteration.text(f'Iteration {i+1}')
|
| 45 |
-
# bar.progress(i + 1)
|
| 46 |
-
# time.sleep(0.1)
|
| 47 |
-
|
| 48 |
-
# print all predicted text:
|
| 49 |
-
for idx in range(len(result)):
|
| 50 |
-
pred_text = result[idx][1]
|
| 51 |
-
st.write(pred_text)
|
| 52 |
-
|
| 53 |
-
# collect the results in the dictionary:
|
| 54 |
-
textdic_easyocr = {}
|
| 55 |
-
for idx in range(len(result)):
|
| 56 |
-
pred_coor = result[idx][0]
|
| 57 |
-
pred_text = result[idx][1]
|
| 58 |
-
pred_confidence = result[idx][2]
|
| 59 |
-
textdic_easyocr[pred_text] = {}
|
| 60 |
-
textdic_easyocr[pred_text]['pred_confidence'] = pred_confidence
|
| 61 |
-
|
| 62 |
-
# create a data frame which shows the predicted text and prediction confidence
|
| 63 |
-
df = pd.DataFrame.from_dict(textdic_easyocr).T
|
| 64 |
-
st.table(df)
|
| 65 |
-
|
| 66 |
-
# get boxes on the image
|
| 67 |
-
rectangle(image, result)
|
| 68 |
-
|
| 69 |
-
st.spinner(text="In progress...")
|
| 70 |
-
|
| 71 |
-
else:
|
| 72 |
-
st.write("Upload your image")
|
| 73 |
|
|
|
|
|
|
|
|
|
|
| 74 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
import easyocr
|
| 5 |
import PIL
|
| 6 |
from PIL import Image, ImageDraw
|
| 7 |
+
from captcha.image import ImageCaptcha
|
| 8 |
+
import random, string
|
| 9 |
|
| 10 |
def rectangle(image, result):
|
| 11 |
# https://www.blog.pythonlibrary.org/2021/02/23/drawing-shapes-on-images-with-python-and-pillow/
|
|
|
|
| 19 |
st.image(image)
|
| 20 |
|
| 21 |
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
# define the costant
|
| 26 |
+
length_captcha = 4
|
| 27 |
+
width = 200
|
| 28 |
+
height = 150
|
| 29 |
+
|
| 30 |
+
# define the function for the captcha control
|
| 31 |
+
def captcha_control():
|
| 32 |
+
#control if the captcha is correct
|
| 33 |
+
if 'controllo' not in st.session_state or st.session_state['controllo'] == False:
|
| 34 |
+
st.title("Captcha Control on OCR")
|
| 35 |
+
|
| 36 |
+
# define the session state for control if the captcha is correct
|
| 37 |
+
st.session_state['controllo'] = False
|
| 38 |
+
col1, col2 = st.columns(2)
|
| 39 |
+
|
| 40 |
+
# define the session state for the captcha text because it doesn't change during refreshes
|
| 41 |
+
if 'Captcha' not in st.session_state:
|
| 42 |
+
st.session_state['Captcha'] = ''.join(random.choices(string.ascii_uppercase + string.digits, k=length_captcha))
|
| 43 |
+
print("the captcha is: ", st.session_state['Captcha'])
|
| 44 |
+
|
| 45 |
+
#setup the captcha widget
|
| 46 |
+
image = ImageCaptcha(width=width, height=height)
|
| 47 |
+
data = image.generate(st.session_state['Captcha'])
|
| 48 |
+
col1.image(data)
|
| 49 |
+
capta2_text = col2.text_area('Enter captcha text', height=30)
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
if st.button("Verify the code"):
|
| 53 |
+
print(capta2_text, st.session_state['Captcha'])
|
| 54 |
+
capta2_text = capta2_text.replace(" ", "")
|
| 55 |
+
# if the captcha is correct, the controllo session state is set to True
|
| 56 |
+
if st.session_state['Captcha'].lower() == capta2_text.lower().strip():
|
| 57 |
+
del st.session_state['Captcha']
|
| 58 |
+
col1.empty()
|
| 59 |
+
col2.empty()
|
| 60 |
+
st.session_state['controllo'] = True
|
| 61 |
+
st.experimental_rerun()
|
| 62 |
+
else:
|
| 63 |
+
# if the captcha is wrong, the controllo session state is set to False and the captcha is regenerated
|
| 64 |
+
st.error("🚨 Error on Captcha...")
|
| 65 |
+
del st.session_state['Captcha']
|
| 66 |
+
del st.session_state['controllo']
|
| 67 |
+
st.experimental_rerun()
|
| 68 |
+
else:
|
| 69 |
+
#wait for the button click
|
| 70 |
+
st.stop()
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
|
| 75 |
# main title
|
| 76 |
+
st.title("Get text from image with Persian and Arabic OCR")
|
| 77 |
|
| 78 |
# subtitle
|
| 79 |
+
st.markdown("## Persian and Arabic OCR :")
|
| 80 |
+
#try_again = 0
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
def main():
|
| 84 |
+
# upload image file
|
| 85 |
+
file = st.file_uploader(label = "Upload Here", type=['png', 'jpg', 'jpeg'])
|
| 86 |
+
# global try_again
|
| 87 |
+
# if try_again == 1:
|
| 88 |
+
# del st.session_state['controllo']
|
| 89 |
+
# st.experimental_rerun()
|
| 90 |
+
# try_again = 1
|
| 91 |
+
#read the csv file and display the dataframe
|
| 92 |
+
if file is not None:
|
| 93 |
+
image = Image.open(file) # read image with PIL library
|
| 94 |
+
st.image(image) #display
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 95 |
|
| 96 |
+
# it will only detect the English and Turkish part of the image as text
|
| 97 |
+
reader = easyocr.Reader(['fa','ar'], gpu=False) #, model_storage_directory='temp/',user_network_directory='temp/net'
|
| 98 |
+
result = reader.readtext(np.array(image)) # turn image to numpy array
|
| 99 |
|
| 100 |
+
# Add a placeholder
|
| 101 |
+
# latest_iteration = st.empty()
|
| 102 |
+
# bar = st.progress(0)
|
| 103 |
+
|
| 104 |
+
# for i in range(100):
|
| 105 |
+
# Update the progress bar with each iteration.
|
| 106 |
+
# latest_iteration.text(f'Iteration {i+1}')
|
| 107 |
+
# bar.progress(i + 1)
|
| 108 |
+
# time.sleep(0.1)
|
| 109 |
+
|
| 110 |
+
# print all predicted text:
|
| 111 |
+
for idx in range(len(result)):
|
| 112 |
+
pred_text = result[idx][1]
|
| 113 |
+
st.write(pred_text)
|
| 114 |
+
|
| 115 |
+
# collect the results in the dictionary:
|
| 116 |
+
textdic_easyocr = {}
|
| 117 |
+
for idx in range(len(result)):
|
| 118 |
+
pred_coor = result[idx][0]
|
| 119 |
+
pred_text = result[idx][1]
|
| 120 |
+
pred_confidence = result[idx][2]
|
| 121 |
+
textdic_easyocr[pred_text] = {}
|
| 122 |
+
textdic_easyocr[pred_text]['pred_confidence'] = pred_confidence
|
| 123 |
+
|
| 124 |
+
# create a data frame which shows the predicted text and prediction confidence
|
| 125 |
+
df = pd.DataFrame.from_dict(textdic_easyocr).T
|
| 126 |
+
st.table(df)
|
| 127 |
+
|
| 128 |
+
# get boxes on the image
|
| 129 |
+
rectangle(image, result)
|
| 130 |
+
|
| 131 |
+
st.spinner(text="In progress...")
|
| 132 |
+
|
| 133 |
+
else:
|
| 134 |
+
st.write("Upload your image")
|
| 135 |
+
|
| 136 |
+
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
# WORK LIKE MULTIPAGE APP
|
| 141 |
+
if 'controllo' not in st.session_state or st.session_state['controllo'] == False:
|
| 142 |
+
captcha_control()
|
| 143 |
+
else:
|
| 144 |
+
main()
|