pdf_image_tools.py
3.98 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
# Copyright (c) Opendatalab. All rights reserved.
from io import BytesIO
import numpy as np
import pypdfium2 as pdfium
from loguru import logger
from PIL import Image
from mineru.data.data_reader_writer import FileBasedDataWriter
from mineru.utils.pdf_reader import image_to_b64str, image_to_bytes, page_to_image
from .enum_class import ImageType
from .hash_utils import str_sha256
def pdf_page_to_image(page: pdfium.PdfPage, dpi=200, image_type=ImageType.PIL) -> dict:
"""Convert pdfium.PdfDocument to image, Then convert the image to base64.
Args:
page (_type_): pdfium.PdfPage
dpi (int, optional): reset the dpi of dpi. Defaults to 200.
image_type (ImageType, optional): The type of image to return. Defaults to ImageType.PIL.
Returns:
dict: {'img_base64': str, 'img_pil': pil_img, 'scale': float }
"""
pil_img, scale = page_to_image(page, dpi=dpi)
image_dict = {
"scale": scale,
}
if image_type == ImageType.BASE64:
image_dict["img_base64"] = image_to_b64str(pil_img)
else:
image_dict["img_pil"] = pil_img
return image_dict
def load_images_from_pdf(
pdf_bytes: bytes,
dpi=200,
start_page_id=0,
end_page_id=None,
image_type=ImageType.PIL, # PIL or BASE64
):
images_list = []
pdf_doc = pdfium.PdfDocument(pdf_bytes)
pdf_page_num = len(pdf_doc)
end_page_id = end_page_id if end_page_id is not None and end_page_id >= 0 else pdf_page_num - 1
if end_page_id > pdf_page_num - 1:
logger.warning("end_page_id is out of range, use images length")
end_page_id = pdf_page_num - 1
for index in range(0, pdf_page_num):
if start_page_id <= index <= end_page_id:
page = pdf_doc[index]
image_dict = pdf_page_to_image(page, dpi=dpi, image_type=image_type)
images_list.append(image_dict)
return images_list, pdf_doc
def cut_image(bbox: tuple, page_num: int, page_pil_img, return_path, image_writer: FileBasedDataWriter, scale=2):
"""从第page_num页的page中，根据bbox进行裁剪出一张jpg图片，返回图片路径 save_path：需要同时支持s3和本地,
图片存放在save_path下，文件名是:
{page_num}_{bbox[0]}_{bbox[1]}_{bbox[2]}_{bbox[3]}.jpg , bbox内数字取整。"""
# 拼接文件名
filename = f"{page_num}_{int(bbox[0])}_{int(bbox[1])}_{int(bbox[2])}_{int(bbox[3])}"
# 老版本返回不带bucket的路径
img_path = f"{return_path}_{filename}" if return_path is not None else None
# 新版本生成平铺路径
img_hash256_path = f"{str_sha256(img_path)}.jpg"
# img_hash256_path = f'{img_path}.jpg'
crop_img = get_crop_img(bbox, page_pil_img, scale=scale)
img_bytes = image_to_bytes(crop_img, image_format="JPEG")
image_writer.write(img_hash256_path, img_bytes)
return img_hash256_path
def get_crop_img(bbox: tuple, pil_img, scale=2):
scale_bbox = (
int(bbox[0] * scale),
int(bbox[1] * scale),
int(bbox[2] * scale),
int(bbox[3] * scale),
)
return pil_img.crop(scale_bbox)
def get_crop_np_img(bbox: tuple, input_img, scale=2):
if isinstance(input_img, Image.Image):
np_img = np.asarray(input_img)
elif isinstance(input_img, np.ndarray):
np_img = input_img
else:
raise ValueError("Input must be a pillow object or a numpy array.")
scale_bbox = (
int(bbox[0] * scale),
int(bbox[1] * scale),
int(bbox[2] * scale),
int(bbox[3] * scale),
)
return np_img[scale_bbox[1]:scale_bbox[3], scale_bbox[0]:scale_bbox[2]]
def images_bytes_to_pdf_bytes(image_bytes):
# 内存缓冲区
pdf_buffer = BytesIO()
# 载入并转换所有图像为 RGB 模式
image = Image.open(BytesIO(image_bytes)).convert("RGB")
# 第一张图保存为 PDF，其余追加
image.save(pdf_buffer, format="PDF", save_all=True)
# 获取 PDF bytes 并重置指针（可选）
pdf_bytes = pdf_buffer.getvalue()
pdf_buffer.close()
return pdf_bytes