# -*- coding: utf-8 -*-
import requests
import uvicorn
from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()


def ocr(language, base64_data):
    if language == "en":
        # English
        language_code = "models/config_en.txt"
    elif language == "zh":
        # 简体中文
        language_code = "models/config_chinese.txt"
    elif language == "tw":
        # 繁體中文
        language_code = "models/config_chinese_cht(v2).txt"
    elif language == "kr":
        # 한국어
        language_code = "models/config_korean.txt"
    elif language == "ru":
        # Русский
        language_code = "models/config_cyrillic.txt"
    elif language == "ja":
        # 日本語
        language_code = "models/config_japan.txt"
    else:
        print("没有传入语言代码")
        return False
    payload = {
        "base64": base64_data,
        "options": {
            # 通用参数
            "tbpu.parser": "multi_para",
            "data.format": "json",
            # 引擎参数
            "ocr.angle": False,  # 是否进行图像旋转校正
            "ocr.language": language_code,
        }
    }
    # 论坛不能发url,因此,自己加前缀...   IP:1224/api/ocr
    result = requests.post('IP:1224/api/ocr', json=payload).json()
    if result['code'] == 100:
        return result
    else:
        return None


def ocr_lua_data_dict(data, box):
    return {
        "text": data['text'],
        "x": (box[0][0] + box[1][0]) / 2,
        "y": (box[0][1] + box[2][1]) / 2,
        "box": data['box'],
    }


def umi_ocr(keyword, language, base64_data, contains: bool = True):
    result = ocr(language, base64_data)
    if result is None:
        return None

    # 左上角       右上角       左下角       右下角
    # [111, 826], [330, 826], [330, 854], [111, 854]
    for data in result['data']:
        if contains:
            if keyword in data["text"]:
                print(data)
                return ocr_lua_data_dict(data, data['box'])
        else:
            if data["text"] == keyword:
                print(data)
                return ocr_lua_data_dict(data, data['box'])

    print("ocr没识别到包含文本")
    return None
class OcrItem(BaseModel):
    keyword: str
    language: str
    Base64: str
    contains: bool


@app.post("/ocr_text", summary="umi ocr识别指定文本-返回坐标")
async def ocr_img(data: OcrItem):
    # "英语":"en",
    # "日语":"ja",
    # "简体中文":"zh",
    # "繁体中文":"tw",
    # "俄语":"ru",
    # "韩语":"kr",
    # keyword: 要匹配的关键词  language:语言模型  contains:包含关键词还是等于,如果为假就是等于,全部匹配
    # payload = {
    #     "keyword": "Search",
    #     "language": "en",
    #     "contains": "true",
    #     "Base64": img_data
    # }

    try:
        body = data.model_dump()
        result = umi_ocr(body['keyword'], body['language'], data.Base64, body['contains'])
        if result is None or result == []:
            return {"status": "error", "result": None}
        else:
            return {"status": "ok", "result": result}
    except Exception as e:
        return {"status": "error", "result": None, "error": e}

if __name__ == "__main__":
    # 这python文件名字一定是  server.py 不然无法启动
    uvicorn.run("server:app", host="0.0.0.0", port=8000, reload=False)

URL也不能发,真的血压高,不发了行吧

不能上传图片,上传图片就开始报错