File size: 14,053 Bytes
2c67080
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
# coding: utf-8
import dataclasses
from copy import deepcopy
from pathlib import Path
import re
from typing import Sequence, Any, Union, Optional

from PIL import Image

from .table_ocr import visualize_cells
from .utils import merge_line_texts, smart_join, y_overlap, list2box
from .layout_parser import ElementType

FORMULA_TAG = '^[(\(]\d+(\.\d+)*[)\)]$'


@dataclasses.dataclass
class Element(object):
    id: str
    box: Sequence
    text: Optional[str]
    meta: Any
    type: ElementType
    total_img: Image.Image
    isolated: bool = False
    col_number: int = -1
    score: float = 1.0
    spellchecker = None
    kwargs: dict = dataclasses.field(default_factory=dict)

    def __init__(
        self,
        *,
        id: str,
        box: Sequence,
        isolated: bool,
        col_number: int,
        meta: Any,
        type: ElementType,
        total_img: Image.Image,
        score: float,
        text: Optional[str] = None,
        spellchecker=None,
        configs: Optional[dict] = None,
    ):
        self.total_img = total_img
        self.id = id
        self.box = box
        self.isolated = isolated
        self.col_number = col_number
        self.meta = meta
        self.type = type
        self.score = score
        self.spellchecker = spellchecker
        self.kwargs = configs or {}

        if self.meta is not None and text is None:
            self.text = self._meta_to_text()
        else:
            self.text = text

        if self.isolated:
            self.text = self.text + '\n'

    def to_dict(self):
        return dataclasses.asdict(self)

    def _meta_to_text(self) -> str:
        if self.type in (ElementType.TEXT, ElementType.TITLE):
            embed_sep = self.kwargs.get('embed_sep', (' $', '$ '))
            isolated_sep = self.kwargs.get('isolated_sep', ('$$\n', '\n$$'))
            line_sep = self.kwargs.get('line_sep', '\n')
            auto_line_break = self.kwargs.get('auto_line_break', True)
            if self.type == ElementType.TITLE:
                for box_info in self.meta:
                    if box_info.get('type', 'text') == 'isolated':
                        box_info['type'] = 'embedding'
            outs = merge_line_texts(
                self.meta,
                auto_line_break,
                line_sep,
                embed_sep,
                isolated_sep,
                self.spellchecker,
            )
            if self.type == ElementType.TITLE:
                outs = smart_join(outs.split('\n'), self.spellchecker)
        elif self.type == ElementType.FORMULA:
            if isinstance(self.meta, dict):
                outs = self.meta['text']
            elif isinstance(self.meta, list):
                outs = [one['text'] for one in self.meta]
        elif self.type == ElementType.TABLE:
            outs = '\n'.join(self.meta.get('markdown', []))
        else:
            outs = ''

        return outs

    def __repr__(self) -> str:
        return f"Element({self.to_dict()})"

    def __str__(self) -> str:
        return self.__repr__()

    def __lt__(self, other) -> bool:
        """
        Adapted from https://github.com/SVJLucas/Scanipy/blob/main/scanipy/elements/element.py.
        Less than operator for Element objects.

        Args:
            other (Element): Another Element object.

        Returns:
            bool: True if this Element is "less than" the other, False otherwise.

        Raises:
            TypeError: If 'other' is not an instance or subclass of Element.
        """

        if not isinstance(other, Element):
            raise TypeError("other must be an instance or subclass of Element")

        return self._column_before(other) or (
            self._same_column(other) and (self.box[1] < other.box[1])
        )

    def _column_before(self, other) -> bool:
        """
        Adapted from https://github.com/SVJLucas/Scanipy/blob/main/scanipy/elements/element.py.
        Check if this Element is in a column before the other Element.

        Args:
            other (Element): Another Element object.

        Returns:
            bool: True if in a column before, False otherwise.
        """

        if not isinstance(other, Element):
            raise TypeError("other must be an instance or subclass of Element")

        return self.col_number < other.col_number
        # max_width = max(box_width(self.box), box_width(other.box))
        # return self.box[0] < other.box[0] - max_width / 2

    def _same_column(self, other) -> bool:
        """
        Check if this Element is in the same column as the other Element.

        Args:
            other (Element): Another Element object.

        Returns:
            bool: True if in the same column, False otherwise.
        """

        if not isinstance(other, Element):
            raise TypeError("other must be an instance or subclass of ElementOutput")

        return self.col_number == other.col_number
        # max_width = max(box_width(self.box), box_width(other.box))
        # return abs(self.box[0] - other.box[0]) < max_width / 2


def box_width(box):
    return box[2] - box[0]


class Page(object):
    number: int
    id: str
    elements: Sequence[Element]
    config: dict
    spellchecker = None

    def __init__(
        self,
        *,
        number: int,
        elements: Sequence[Element],
        id: Optional[str] = None,
        spellchecker=None,
        config=None,
    ):
        self.number = number
        self.id = id or str(number)
        self.elements = elements
        self.spellchecker = spellchecker
        self.config = config or {}

    def __repr__(self) -> str:
        return f"Page(id={self.id}, number={self.number}, elements={self.elements})"

    def to_markdown(
        self,
        out_dir: Union[str, Path],
        root_url: Optional[str] = None,
        markdown_fn: Optional[str] = 'output.md',
    ) -> str:
        """
        Convert the Page to markdown.
        Args:
            out_dir (Union[str, Path]): The output directory.
            root_url (Optional[str]): The root url for the saved images in the markdown files.
            markdown_fn (Optional[str]): The markdown file name. Default is 'output.md'.

        Returns: The markdown string.

        """
        out_dir = Path(out_dir)
        out_dir.mkdir(exist_ok=True, parents=True)
        self.elements = self._merge_isolated_formula_and_tag(self.elements)
        self.elements.sort()
        if not self.elements:
            return ''
        md_out = self._ele_to_markdown(self.elements[0], root_url, out_dir)
        for idx, element in enumerate(self.elements[1:]):
            prev_element = self.elements[idx]
            cur_txt = self._ele_to_markdown(element, root_url, out_dir)
            if (
                prev_element.col_number + 1 == element.col_number
                and prev_element.type == element.type
                and prev_element.type in (ElementType.TEXT, ElementType.TITLE)
                and md_out
                and md_out[-1] != '\n'
                and cur_txt
                and cur_txt[0] != '\n'
            ):
                # column continuation
                md_out = smart_join([md_out, cur_txt], self.spellchecker)
            else:
                md_out += '\n\n' + cur_txt

        line_sep = '\n'
        md_out = re.sub(
            rf'{line_sep}{{2,}}', f'{line_sep}{line_sep}', md_out
        )  # 把2个以上的连续 '\n' 替换为 '\n\n'
        if markdown_fn:
            with open(out_dir / markdown_fn, 'w', encoding='utf-8') as f:
                f.write(md_out)
        return md_out

    def _ele_to_markdown(
        self, element: Element, root_url: Optional[str], out_dir: Union[str, Path]
    ):
        type = element.type
        text = element.text
        if type in (ElementType.TEXT, ElementType.TABLE):
            if type == ElementType.TABLE:
                visualize_cells(
                    element.total_img.crop(element.box),
                    element.meta['cells'][0],
                    out_dir / f'{element.id}.png',
                )
            return text
        elif type == ElementType.TITLE:
            return f'## {text}'
        elif type == ElementType.FORMULA:
            isolated_sep = self.config.get('isolated_sep', ('$$\n', '\n$$'))
            return isolated_sep[0] + text.strip() + isolated_sep[1]
        elif type == ElementType.FIGURE:
            out_figure_dir = out_dir / 'figures'
            out_figure_dir.mkdir(exist_ok=True, parents=True)
            out_path = out_figure_dir / f'{element.id}-{type.name}.jpg'
            element.total_img.crop(element.box).save(str(out_path))

            _url = self._map_path_to_url(root_url, out_path, out_dir)
            return f'![]({_url})'
        return ''

    def _map_path_to_url(self, root_url: Optional[str], path: Path, out_dir: Path):
        rel_url = path.relative_to(out_dir)
        if root_url is not None:
            return f'{root_url}/{rel_url}'
        return str(rel_url)

    @classmethod
    def _merge_isolated_formula_and_tag(cls, elements):
        # 合并孤立的公式和公式标题
        # 对于每个公式标题,找到与它在同一行且在其左侧距离最近的孤立公式,并把它们合并
        isolated_formula = [
            item
            for item in elements
            if item.type == ElementType.FORMULA and item.isolated
        ]
        formula_caption = [
            item
            for item in elements
            if item.type == ElementType.TEXT and re.match(FORMULA_TAG, item.text)
        ]
        ele_ids = set([item.id for item in isolated_formula + formula_caption])
        remaining_elements = [item for item in elements if item.id not in ele_ids]
        for caption in formula_caption:
            caption_xmin, caption_ymin, caption_xmax, caption_ymax = caption.box
            min_dist = float('inf')
            nearest_formula = None
            for formula in isolated_formula:
                formula_xmin, formula_ymin, formula_xmax, formula_ymax = formula.box
                if (
                    caption.col_number == formula.col_number
                    and y_overlap(
                        list2box(*caption.box), list2box(*formula.box), key=None
                    )
                    >= 0.8
                ):
                    dist = caption_xmin - formula_xmax
                    if 0 <= dist < min_dist:
                        min_dist = dist
                        nearest_formula = formula
            if nearest_formula is not None:
                new_formula = deepcopy(nearest_formula)
                formula_xmin, formula_ymin, formula_xmax, formula_ymax = new_formula.box
                new_formula.box = [
                    min(caption_xmin, formula_xmin),
                    min(caption_ymin, formula_ymin),
                    max(caption_xmax, formula_xmax),
                    max(caption_ymax, formula_ymax),
                ]
                new_text = new_formula.text.strip() + ' \\tag{{{}}}'.format(
                    caption.text[1:-1]
                )
                new_formula.text = new_text
                if new_formula.meta and isinstance(new_formula.meta, dict):
                    new_formula.meta['text'] = new_text
                remaining_elements.append(new_formula)
                isolated_formula.remove(nearest_formula)
            else:  # not found
                remaining_elements.append(caption)
        return remaining_elements + isolated_formula


class Document(object):
    number: int
    id: str
    pages: Sequence[Page]
    config: dict
    spellchecker = None

    def __init__(
        self,
        *,
        number: int,
        pages: Sequence[Page],
        id: Optional[str] = None,
        spellchecker=None,
        config=None,
    ):
        self.number = number
        self.id = id or str(number)
        self.pages = pages
        self.spellchecker = spellchecker
        self.config = config or {}

    def __repr__(self) -> str:
        return f"Document(id={self.id}, number={self.number}, pages={self.pages})"

    def to_markdown(
        self,
        out_dir: Union[str, Path],
        root_url: Optional[str] = None,
        markdown_fn: Optional[str] = 'output.md',
    ) -> str:
        """
        Convert the Document to markdown.
        Args:
            out_dir (Union[str, Path]): The output directory.
            root_url (Optional[str]): The root url for the saved images in the markdown files.
            markdown_fn (Optional[str]): The markdown file name. Default is 'output.md'.

        Returns: The markdown string.

        """
        out_dir = Path(out_dir)
        out_dir.mkdir(exist_ok=True, parents=True)
        self.pages.sort(key=lambda page: page.number)
        if not self.pages:
            return ''
        md_out = self.pages[0].to_markdown(out_dir, root_url=root_url, markdown_fn=None)
        for idx, page in enumerate(self.pages[1:]):
            prev_page = self.pages[idx]
            cur_txt = page.to_markdown(out_dir, root_url=root_url, markdown_fn=None)
            if (
                md_out
                and prev_page.elements
                and prev_page.elements[-1].type in (ElementType.TEXT, ElementType.TITLE)
                and page.elements
                and page.elements[0].type in (ElementType.TEXT, ElementType.TITLE)
                and md_out[-1] != '\n'
                and cur_txt
                and cur_txt[0] != '\n'
            ):
                # column continuation
                md_out = smart_join([md_out, cur_txt], self.spellchecker)
            else:
                md_out += '\n\n' + cur_txt

        line_sep = '\n'
        md_out = re.sub(
            rf'{line_sep}{{2,}}', f'{line_sep}{line_sep}', md_out
        )  # 把2个以上的连续 '\n' 替换为 '\n\n'
        if markdown_fn:
            with open(out_dir / markdown_fn, 'w', encoding='utf-8') as f:
                f.write(md_out)
        return md_out