app.py

import os
import gradio as gr
import cv2
import pytesseract
from fastapi import FastAPI
import numpy as np

#pytesseract.pytesseract.tesseract_cmd = r'./Tesseract-OCR/tesseract.exe'

print(os.popen(f'cat /etc/debian_version').read())
print(os.popen(f'cat /etc/issue').read())
print(os.popen(f'apt search tesseract').read())


def PreprocessIMG(image):
    
    # Convert to grayscale
    image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    # Apply Canny edge detection
    image = cv2.GaussianBlur(image,(3,3),0)
    edges = cv2.Canny(image, 90, 120, apertureSize=3)
    # Apply Hough line transform to detect lines
    lines = cv2.HoughLines(edges, 1, np.pi/180, threshold=300)

    # Remove lines from the image
    for rho, theta in lines[:, 0]:
        a = np.cos(theta)
        b = np.sin(theta)
        x0 = a * rho
        y0 = b * rho
        x1 = int(x0 + 1000 * (-b))
        y1 = int(y0 + 1000 * (a))
        x2 = int(x0 - 1000 * (-b))
        y2 = int(y0 - 1000 * (a))
        cv2.line(image, (x1, y1), (x2, y2), (255, 255, 255), 3)

    return image




def TextLineBox(img, engine):

    class Lines:
        def __init__(self,x,y,w,h,text):
            self.x = x
            self.y = y
            self.w = w
            self.h = h
            self.text = text

    lineboxes = []

    #read image
    img = PreprocessIMG(img)

    ### Cofig
    configname = r' --oem 3 --psm ' + str(engine) + ' -l eng'

    #### Text for testing
    texttest = pytesseract.image_to_string(img ,config=configname)

    ### Box of words
    boxes = pytesseract.image_to_data(img, config=configname)
    # print(boxes)

    #slit box and concatenate into line
    skip = 0
    for b in boxes.splitlines():
        ## skip header
        if (skip == 0):
            skip = 1
            continue
        ## get box of word in 1 object
        b = b.split()
        if (len(b) < 12):       ## it is a space not a word
            continue

        #print(b)
        x,y,w,h,text = int(b[6]),int(b[7]),int(b[8]),int(b[9]), b[11]


        ### Begin New line if the word having num_word is 1
        if (int(b[5]) == 1):
            lineboxes.append(  Lines(x,y,w,h,text)  )

        ### Next word inline
        else:
            lineboxes[-1].text += ' ' +  text
            if (x > lineboxes[-1].x):
                lineboxes[-1].w = x - lineboxes[-1].x + w
            if (y < lineboxes[-1].y):
                lineboxes[-1].y = y
            if (y+h > lineboxes[-1].y + lineboxes[-1].h):
                lineboxes[-1].h = y+h - lineboxes[-1].y

        #draw the box of WORD
        cv2.rectangle(img, (x,y) , (w+x,y+h), (255,0,0), 2 )


    return texttest,img
    

def Download(text):
    with open("test.txt", "w") as file:
        file.write(text)
    return "test.txt"







with gr.Blocks (theme="ParityError/Anime"  , css="#SUBMIT {background-color: #cdb4db} #DOWNLOAD {background-color: #a2d2ff}") as demo:
    with gr.Row():
        with gr.Column():
            input = gr.Image()
            engine_input = gr.Text(label="Engine Mode Number")
            text_output = gr.Text(label="Result Text")
            file_output = gr.File()
            with gr.Row():
                submit_btn = gr.Button("SUBMIT" , elem_id="SUBMIT")
                download_btn = gr.Button("DOWNLOAD", elem_id="DOWNLOAD")
                clear_btn = gr.Button("CLEAR")

        with gr.Column():
            image_output = gr.Image()

    submit_btn.click(TextLineBox, inputs=[input,engine_input,], outputs= [text_output, image_output, ] )
    download_btn.click(Download, text_output, outputs= file_output )
    clear_btn.click(lambda: [None,None,None], inputs=None, outputs= [text_output, file_output, image_output])

    demo.load(
        None,
        None,
        _js="""
            () => {
                const params = new URLSearchParams(window.location.search);
                if (!params.has('__theme')) {
                    params.set('__theme', 'dark');
                    window.location.search = params.toString();
                }
            }""",
    )



demo.queue().launch()


# Cach 1 dung app Gradio (share = True)
# Cach 2 dung Mount_Gradio_app trong FASTAPI app
# Cach 3 

# app = FastAPI()
# app = gr.mount_gradio_app(app, demo, path="/OCR" )




#bt.click(fn=None, _js="window.open('https://google.com', '_blank')")