app.py

# -*- encoding: utf-8 -*-
# @Author: SWHL
# @Contact: [email protected]
from enum import Enum
from pathlib import Path
from typing import List, Union

import gradio as gr
import numpy as np
from rapidocr import RapidOCR


class InferEngine(Enum):
    ort = "ONNXRuntime"
    vino = "OpenVino"
    paddle = "PaddlePaddle"
    torch = "PyTorch"


def get_ocr_engine(infer_engine: str, lang_det: str, lang_rec: str) -> RapidOCR:
    engine_mapping = {
        InferEngine.vino.value: "with_openvino",
        InferEngine.paddle.value: "with_paddle",
        InferEngine.torch.value: "with_torch",
    }
    param_key = engine_mapping.get(infer_engine, "with_onnx")

    return RapidOCR(
        params={
            f"Global.{param_key}": True,
            "Global.lang_det": lang_det,
            "Global.lang_rec": lang_rec,
        }
    )


def get_ocr_result(
    img: np.ndarray,
    text_score,
    box_thresh,
    unclip_ratio,
    lang_det,
    lang_rec,
    infer_engine,
    is_word: str,
):
    return_word_box = True if is_word == "Yes" else False

    ocr_engine = get_ocr_engine(infer_engine, lang_det=lang_det, lang_rec=lang_rec)

    ocr_result = ocr_engine(
        img,
        text_score=text_score,
        box_thresh=box_thresh,
        unclip_ratio=unclip_ratio,
        return_word_box=return_word_box,
    )
    vis_img = ocr_result.vis()
    if return_word_box:
        txts, scores, _ = list(zip(*ocr_result.word_results))
        ocr_txts = [[i, txt, score] for i, (txt, score) in enumerate(zip(txts, scores))]
        return vis_img, ocr_txts, ocr_result.elapse

    ocr_txts = [
        [i, txt, score]
        for i, (txt, score) in enumerate(zip(ocr_result.txts, ocr_result.scores))
    ]
    return vis_img, ocr_txts, ocr_result.elapse


def create_examples() -> List[List[Union[str, float]]]:
    DEFAULT_VALUES = [0.5, 0.5, 1.6, "ch_mobile", "ch_mobile", "ONNXRuntime", "No"]

    image_specs = [
        ("images/ch_en_num.jpg", {}),
        ("images/japan.jpg", {3: "multi_mobile", 4: "japan_mobile"}),
        ("images/korean.jpg", {3: "multi_mobile", 4: "korean_mobile"}),
        ("images/air_ticket.jpg", {}),
        ("images/car_plate.jpeg", {}),
        ("images/train_ticket.jpeg", {}),
    ]

    examples = []
    for image_path, overrides in image_specs:
        example = DEFAULT_VALUES.copy()
        example.insert(0, image_path)
        for index, value in overrides.items():
            example[index + 1] = value
        examples.append(example)
    return examples


infer_engine_list = [InferEngine[v].value for v in InferEngine.__members__]

lang_det_list = ["ch_mobile", "ch_server", "en_mobile", "en_server", "multi_mobile"]
lang_rec_list = [
    "ch_mobile",
    "ch_server",
    "chinese_cht",
    "en_mobile",
    "ar_mobile",
    "cyrillic_mobile",
    "devanagari_mobile",
    "japan_mobile",
    "ka_mobile",
    "korean_mobile",
    "latin_mobile",
    "ta_mobile",
    "te_mobile",
]

custom_css = """
    body {font-family: body {font-family: 'Helvetica Neue', Helvetica;}
    .gr-button {background-color: #4CAF50; color: white; border: none; padding: 10px 20px; border-radius: 5px;}
    .gr-button:hover {background-color: #45a049;}
    .gr-textbox {margin-bottom: 15px;}
    .example-button {background-color: #1E90FF; color: white; border: none; padding: 8px 15px; border-radius: 5px; margin: 5px;}
    .example-button:hover {background-color: #FF4500;}
    .tall-radio .gr-radio-item {padding: 15px 0; min-height: 50px; display: flex; align-items: center;}
    .tall-radio label {font-size: 16px;}
    .output-image, .input-image, .image-preview {height: 300px !important}
"""

with gr.Blocks(
    title="Rapid⚡OCR Demo", css="custom_css", theme=gr.themes.Soft()
) as demo:
    gr.HTML(
        """
        <h1 style='text-align: center;font-size:40px'>Rapid⚡OCR</h1>
        
        <div style="display: flex; justify-content: center; gap: 10px;">
            <a href=""><img src="https://img.shields.io/badge/Python->=3.6-aff.svg"></a>
            <a href="https://rapidai.github.io/RapidOCRDocs"><img src="https://img.shields.io/badge/Docs-link-aff.svg"></a>
            <a href=""><img src="https://img.shields.io/badge/OS-Linux%2C%20Win%2C%20Mac-pink.svg"></a>
            <a href="https://pepy.tech/project/rapidocr"><img src="https://static.pepy.tech/personalized-badge/rapidocr?period=total&units=abbreviation&left_color=grey&right_color=blue&left_text=Downloads%20rapidocr"></a>
            <a href="https://pypi.org/project/rapidocr/"><img alt="PyPI" src="https://img.shields.io/pypi/v/rapidocr"></a>
            <a href="https://github.com/RapidAI/RapidOCR"><img src="https://img.shields.io/github/stars/RapidAI/RapidOCR?color=ccf"></a>
        </div>
    """
    )
    with gr.Row():
        text_score = gr.Slider(
            label="text_score",
            minimum=0,
            maximum=1.0,
            value=0.5,
            step=0.1,
            info="文本识别结果是正确的置信度，值越大，显示出的识别结果更准确。存在漏检时，调低该值。取值范围：[0, 1.0]，默认值为0.5",
        )
        box_thresh = gr.Slider(
            label="box_thresh",
            minimum=0,
            maximum=1.0,
            value=0.5,
            step=0.1,
            info="检测到的框是文本的概率，值越大，框中是文本的概率就越大。存在漏检时，调低该值。取值范围：[0, 1.0]，默认值为0.5",
        )
        unclip_ratio = gr.Slider(
            label="unclip_ratio",
            minimum=1.5,
            maximum=2.0,
            value=1.6,
            step=0.1,
            info="控制文本检测框的大小，值越大，检测框整体越大。在出现框截断文字的情况，调大该值。取值范围：[1.5, 2.0]，默认值为1.6",
        )

    with gr.Row():
        select_infer_engine = gr.Dropdown(
            choices=infer_engine_list,
            label="Infer Engine (推理引擎)",
            value="ONNXRuntime",
            interactive=True,
        )
        lang_det = gr.Dropdown(
            choices=lang_det_list,
            label="Det model (文本检测模型)",
            value=lang_det_list[0],
            interactive=True,
        )
        lang_rec = gr.Dropdown(
            choices=lang_rec_list,
            label="Rec model (文本识别模型)",
            value=lang_rec_list[0],
            interactive=True,
        )
        is_word = gr.Radio(
            ["Yes", "No"], label="Return word box (返回单字符)", value="No"
        )

    img_input = gr.Image(label="Upload or Select Image", sources="upload")

    run_btn = gr.Button("Run")

    img_output = gr.Image(label="Output Image")
    elapse = gr.Textbox(label="Elapse(s)")
    ocr_results = gr.Dataframe(
        label="OCR Txts",
        headers=["Index", "Txt", "Score"],
        datatype=["number", "str", "number"],
        show_copy_button=True,
    )

    ocr_inputs = [
        img_input,
        text_score,
        box_thresh,
        unclip_ratio,
        lang_det,
        lang_rec,
        select_infer_engine,
        is_word,
    ]
    run_btn.click(
        get_ocr_result, inputs=ocr_inputs, outputs=[img_output, ocr_results, elapse]
    )

    examples = gr.Examples(
        examples=create_examples(),
        examples_per_page=5,
        inputs=ocr_inputs,
        fn=get_ocr_result,
        outputs=[img_output, ocr_results, elapse],
        cache_examples=False,
    )


if __name__ == "__main__":
    demo.launch(debug=True)