# python 實現圖片文字辨識(ocr)

import pytesseract
from PIL import Image

# 設置tesseract執行文件的路徑(linux)
pytesseract.pytesseract.tesseract_cmd = '/usr/bin/tesseract'

# 設置提取中文和英文


# 讀取圖片
image = Image.open('data/tt.jpeg')

# 圖片轉為灰度圖
image = image.convert('L')

# 圖片二值化
threshold = 127
table = []
for i in range(256):
    if i < threshold:
        table.append(0)
    else:
        table.append(1)
image = image.point(table, '1')

# 圖片轉為字符串
#text = pytesseract.image_to_string(image, lang='chi_sim+eng')
#print(text)
#
#data = pytesseract.image_to_boxes(image, lang='chi_sim+eng')
#print(data)

# 座標(全數據)
#data = pytesseract.image_to_data(image, lang='chi_sim+eng')
#print(data)

data = pytesseract.image_to_alto_xml(image, lang='chi_sim+eng')
#print(data)

## 打印格式化的XML
#from xml.dom.minidom import parseString
#dom = parseString(data)
#print(dom.toprettyxml())

# 打印格式化的JSON
import json
from xmljson import badgerfish as bf
from xml.etree.ElementTree import fromstring

# 過濾掉@CONTENT爲空的數據和爲數字的數據
def filter_data(data):
    if isinstance(data, dict):
        if '@CONTENT' in data:
            if data['@CONTENT'] == '' or data['@CONTENT'].isdigit():
                return None
        for k, v in data.items():
            if isinstance(v, dict):
                data[k] = filter_data(v)
            elif isinstance(v, list):
                data[k] = [filter_data(i) for i in v]
    return data

xml = fromstring(data)
json = json.dumps(bf.data(xml), indent=4, ensure_ascii=False)
print(json)