deploy.py

import gradio as gr
from app.NER.NER import name_entity_recognizer
from app.extract_country.country_dictionary import find_country
from app.extract_dates.date_engine import date_extractor
from app.extract_gender.gender_extractor import gender_extract
from app.extract_identity_number.doc_number_extractor import doc_number
import cv2
from app.functions import get_index, get_latest_value_from_csv, image_enhacement
import json


def ocr(card_type_input, image_input, webcam_input):
    global dictionary

    dictionary = {
        "NAME": "",
        "DOB": "",
        "COUNTRY": "",
        "GENDER": "",
        "DOCUMENT NUMBER": "",
    }

    file_path = "app/images/idcards/input.jpg"

    if image_input is not None and image_input.any():
        print("In input image")
        cv2.imwrite(file_path, image_input)

    else:
        print("In webcam")
        cv2.imwrite(file_path, cv2.flip(webcam_input, 1))

    image_enhacement(file_path)
    # print("file_path", file_path)
    dictionary, NER_check, result = name_entity_recognizer(file_path, dictionary)

    extract_text = []
    extract_text = [line[1][0] for res in result for line in res]
    extract_text = " ".join(extract_text)
    print("extract_text", extract_text)

    if "" in list(dictionary.values()):
        print("Missing value found in Dic")

        if len(dictionary["GENDER"]) == 0:
            print("Gender Missing")
            gender_found, gender, ocr_list = gender_extract(extract_text)

            dictionary["GENDER"] = gender if gender_found else None

        if len(dictionary["DOB"]) == 0:
            print("Dob Missing")
            dob_found, dob, ocr_list = date_extractor(extract_text)

            dictionary["DOB"] = dob if dob_found else None

        if len(dictionary["COUNTRY"]) == 0:
            print("Country Missing")
            country_found, country, ocr_list = find_country(extract_text)
            dictionary["country"] = country if country_found else None

        if len(dictionary["DOCUMENT NUMBER"]) == 0:
            print("document Number missing")

            document_number_found, document_number, ocr_list = doc_number(extract_text)
            dictionary["DOCUMENT NUMBER"] = (
                document_number if document_number_found else None
            )

    # bool_check , training_data = get_latest_value_from_csv(extract_text)
    # print(training_data)
    print("Updated Dict ", dictionary)
    response = {"Status": 200, "OCR": dictionary}
    return response, extract_text


def save_traning_data(
    name_input, dob_input, country_input, gender_input, doc_num_input,string_output
):
    print(name_input, dob_input, country_input, gender_input, doc_num_input)
    print("Saving data")
    # print(string_output)
    entities_list = []
    
    if name_input:
        entry = []
        start_index, end_index = get_index(string_output, name_input)
        entry.append(start_index)
        entry.append(end_index)

        entry.append("NAME")

        entities_list.append(entry)

    if dob_input:
        entry = []
        start_index, end_index = get_index(string_output, dob_input)
        entry.append(start_index)
        entry.append(end_index)
        entry.append("DOB")

        entities_list.append(entry)
    
    if country_input:
        entry = []
        start_index, end_index = get_index(string_output, country_input)
        entry.append(start_index)
        entry.append(end_index)
        entry.append("COUNTRY")

        entities_list.append(entry)
    
    if gender_input:
        entry = []
        start_index, end_index = get_index(string_output, gender_input)
        entry.append(start_index)
        entry.append(end_index)
        entry.append("GENDER")

        entities_list.append(entry)
    
    if doc_num_input:
        entry = []
        start_index, end_index = get_index(string_output, doc_num_input)
        entry.append(start_index)
        entry.append(end_index)
        entry.append("DOCUMENT NUMBER")
        
        entities_list.append(entry)
        print("entities_list" , entities_list)

    with open("app/training_data/annotated_data.json") as f:
        data = json.load(f)
        
        new_list = [
            string_output,
            {
                "entities": entities_list
            }
        ]

        annotations = data["annotations"]
        annotations.append(new_list)
    
        with open('app/training_data/annotated_data.json', 'w') as file:
            json.dump(data, file)

# text = "PAKISTAN Nationalldentity Card ISLAMIC REPUBLIC OF PAKISTAN Name Muhammad Owais Siddiqui Father Name Zia Ud Din Siddiqui Country of Stay Gender M Pakistan Identity Number Date of Birth 42201-1177064-3 11.08.2001 Date of Expiry Date of Issue 24.09.2019 24.09.2029 Holder's Signature"
# save_traning_data(
#     "Muhammad Owais Siddiqui", "Card", "REPUBLIC", "PAKISTAN", "Nationalldentity",text
# )

with gr.Blocks() as interface:
    with gr.Row():
        with gr.Column():
            card_type_input = gr.inputs.Dropdown(
                ["passport", "id_card"], label="Card Type", default="id_card"
            )
            image_input = gr.inputs.Image(label="Upload Image")
            webcam_input = gr.inputs.Image(label="Webcam", source="webcam")
        with gr.Column():
            # Define the output components
            json_output = gr.outputs.JSON(label="JSON Output")
            string_output = gr.outputs.Textbox(label="OCR Extracted Text")
            ocr_btn = gr.Button("Process OCR")

            name_input = gr.inputs.Textbox(label="NAME", optional=True, type="text")
            dob_input = gr.inputs.Textbox(label="DOB", optional=True, type="text")
            country_input = gr.inputs.Textbox(
                label="COUNTRY", optional=True, type="text"
            )
            gender_input = gr.inputs.Textbox(label="GENDER", optional=True, type="text")
            doc_num_input = gr.inputs.Textbox(
                label="DOCUMENT NUMBER", optional=True, type="text"
            )
            save_btn = gr.Button("Save Data")

    ocr_btn.click(
        fn=ocr,
        inputs=[card_type_input, image_input, webcam_input],
        outputs=[json_output, string_output],
    )

    save_btn.click(
        fn=save_traning_data,
        inputs=[name_input, dob_input, country_input, gender_input, doc_num_input,string_output],
    )

interface.launch(share=True, show_tips=True, debug=True)